2020 Vol. 41, No. 1
Display Method:
2020, 41(1): 1-9.
doi: 10.5768/JAO202041.0101001
Abstract:
Based on the theory of coded exposure, an applied embedded coded exposure camera system was designed. The exposure shutter of camera was controlled according to the preset coding sequence when collecting the moving target image, and the image of photo-generated charge multiple superimposed with a transfer was obtained. The acquired image was a motion blurred image with coded exposure information, which had more details of the target. The data was temporarily stored in the data memory by the coded exposure image after the AD conversion, and the restored image was decoded by the core device. The experimental results show that the coded exposure camera can effectively solve the problem of linear motion blur. In the same conditions, the index average value of restored image quality under coded exposure condition is increased nearly twice than that of restored image quality under single exposure condition by using the no-reference image evaluation index test.
Based on the theory of coded exposure, an applied embedded coded exposure camera system was designed. The exposure shutter of camera was controlled according to the preset coding sequence when collecting the moving target image, and the image of photo-generated charge multiple superimposed with a transfer was obtained. The acquired image was a motion blurred image with coded exposure information, which had more details of the target. The data was temporarily stored in the data memory by the coded exposure image after the AD conversion, and the restored image was decoded by the core device. The experimental results show that the coded exposure camera can effectively solve the problem of linear motion blur. In the same conditions, the index average value of restored image quality under coded exposure condition is increased nearly twice than that of restored image quality under single exposure condition by using the no-reference image evaluation index test.
2020, 41(1): 10-17.
doi: 10.5768/JAO202041.0101002
Abstract:
The photoelectric positioning of moving targets cannot do the mean filtering as simple as the stationary targets. In view of this, the particle filter algorithm was introduced, which could be applied in both linear and nonlinear systems. Firstly, based on the requirements of photoelectric positioning, the calculation formulas as well as the initial value and parameter selection formulas were deduced in detail. Secondly, the photoelectric passive positioning algorithm of sea moving targets with only measurement noise and with the measurement and motion noises was simulated, which verified the validity of algorithm. Finally, the influence of noise intensity on filtering effect, the influence of filtering parameters selection on filtering effect, the influence of target motion mode on filtering following performance and the resampling algorithms on filtering effect were discussed. The conclusions are as follows: ① the particle filter can be used in the photoelectric positioning process of moving targets to effectively reduce the positioning error; ② the particle filter algorithm has strong robustness, which is suitable for the situations such as large noise, big changes of target motion morphology, etc.
The photoelectric positioning of moving targets cannot do the mean filtering as simple as the stationary targets. In view of this, the particle filter algorithm was introduced, which could be applied in both linear and nonlinear systems. Firstly, based on the requirements of photoelectric positioning, the calculation formulas as well as the initial value and parameter selection formulas were deduced in detail. Secondly, the photoelectric passive positioning algorithm of sea moving targets with only measurement noise and with the measurement and motion noises was simulated, which verified the validity of algorithm. Finally, the influence of noise intensity on filtering effect, the influence of filtering parameters selection on filtering effect, the influence of target motion mode on filtering following performance and the resampling algorithms on filtering effect were discussed. The conclusions are as follows: ① the particle filter can be used in the photoelectric positioning process of moving targets to effectively reduce the positioning error; ② the particle filter algorithm has strong robustness, which is suitable for the situations such as large noise, big changes of target motion morphology, etc.
2020, 41(1): 18-23.
doi: 10.5768/JAO202041.0101003
Abstract:
In order to solve the serious problem of energy loss in long distance diffuse reflection laser ranging system and improve the response speed of ranging system, a receiving lens with small detecting target surface and large clear aperture was designed. The three groups of 4-slices Petzval mode were used as the initial prototype of receiving lens, the aperture diaphragm was placed forward in the first side and its aperture was increased to improve the receiving efficiency. The concave of the last lens was designed as aspheric shape, which aimed to reduce the spherical aberration and miniaturize the system, and an avalanche photodiode detector with the diameter of 75 µm was used for the target surface to ensure the response speed was less than 45 ns. The receiving lens with the clear aperture of 120 mm is obtained by the optimized design, and its F-number is 0.6, which can be used in the field of long distance diffuse reflection laser ranging.
In order to solve the serious problem of energy loss in long distance diffuse reflection laser ranging system and improve the response speed of ranging system, a receiving lens with small detecting target surface and large clear aperture was designed. The three groups of 4-slices Petzval mode were used as the initial prototype of receiving lens, the aperture diaphragm was placed forward in the first side and its aperture was increased to improve the receiving efficiency. The concave of the last lens was designed as aspheric shape, which aimed to reduce the spherical aberration and miniaturize the system, and an avalanche photodiode detector with the diameter of 75 µm was used for the target surface to ensure the response speed was less than 45 ns. The receiving lens with the clear aperture of 120 mm is obtained by the optimized design, and its F-number is 0.6, which can be used in the field of long distance diffuse reflection laser ranging.
2020, 41(1): 24-32.
doi: 10.5768/JAO202041.0101004
Abstract:
Recovering the depth of scenes from images is a key issue in the field of computer vision. Considering that the single type images were limited by different illumination of scenes in depth estimation, a method of depth estimation based on the progressively fusion convolution neural network of infrared and visible light images (PF-CNN) was proposed. This method includes the two-stream filter partially coupled network (TFCNet), the adaptive multi-model feature fusion network (AMFNet) and adaptive progressively feature fusion network (APFNet). The filters of infrared and visible light images are partially coupled in the TFCNet to enhance the features of them. The AMFNet learns the residual features of infrared and visible light images and fused them adaptively to fully utilize the complementary information. The APFNet learns the combination of multi-layer fusion features to make full use of the different characteristics of different convolutional layers. The experimental results show that the proposed PF-CNN obtains best performances on the data set and increases the threshold accuracy by 5%, which is better than other methods significantly.
Recovering the depth of scenes from images is a key issue in the field of computer vision. Considering that the single type images were limited by different illumination of scenes in depth estimation, a method of depth estimation based on the progressively fusion convolution neural network of infrared and visible light images (PF-CNN) was proposed. This method includes the two-stream filter partially coupled network (TFCNet), the adaptive multi-model feature fusion network (AMFNet) and adaptive progressively feature fusion network (APFNet). The filters of infrared and visible light images are partially coupled in the TFCNet to enhance the features of them. The AMFNet learns the residual features of infrared and visible light images and fused them adaptively to fully utilize the complementary information. The APFNet learns the combination of multi-layer fusion features to make full use of the different characteristics of different convolutional layers. The experimental results show that the proposed PF-CNN obtains best performances on the data set and increases the threshold accuracy by 5%, which is better than other methods significantly.
2020, 41(1): 33-36.
doi: 10.5768/JAO202041.0101005
Abstract:
Optical image sensor is an important part of satellite guidance, navigation, and control (GNC) subsystem, the sensor can get the moon’s image, and the image information can help the moon lander to recognize obstacles, plan paths and select safety areas. In order to realize the function and performance detection of the optical image sensor, a kind of equipment that can afford infinite optical target in the limited space is needed, which can give the moon’s image information to the optical image sensor. A static optical target simulator using films was designed and developed. The simulator used films as the image sources, the films were illuminated by an integrating sphere and the films images were projected to infinity by optical lenses, so as the images at infinity were received by optical image sensor. Results of the designed static optical target simulator show that the field view is more than 30°×30°, the entrance pupil is Φ5 mm, the distance of the entrance pupil is 41 mm, the focal length is 44.78 mm, and the image resolution is more than 1 024×1 024. All the indicators of the static optical target simulator can reach the requirements of the optical image sensor light closed-loop test.
Optical image sensor is an important part of satellite guidance, navigation, and control (GNC) subsystem, the sensor can get the moon’s image, and the image information can help the moon lander to recognize obstacles, plan paths and select safety areas. In order to realize the function and performance detection of the optical image sensor, a kind of equipment that can afford infinite optical target in the limited space is needed, which can give the moon’s image information to the optical image sensor. A static optical target simulator using films was designed and developed. The simulator used films as the image sources, the films were illuminated by an integrating sphere and the films images were projected to infinity by optical lenses, so as the images at infinity were received by optical image sensor. Results of the designed static optical target simulator show that the field view is more than 30°×30°, the entrance pupil is Φ5 mm, the distance of the entrance pupil is 41 mm, the focal length is 44.78 mm, and the image resolution is more than 1 024×1 024. All the indicators of the static optical target simulator can reach the requirements of the optical image sensor light closed-loop test.
2020, 41(1): 37-42.
doi: 10.5768/JAO202041.0101006
Abstract:
With the rapid commercialization of image sensors and the development of biometric recognition algorithms in recent years, iris recognition function has been applied to mobile terminal devices. The iris image acquisition is the first and the key step in iris recognition. Here one group of iris recognition lens was designed for mobile phones by the ZEMAX optical design software. This design is composed of three chip aspherical lenses of optical plastic, with OmniVision_OV2281 sensor of Howe technologies. With designed F parameter of 2.3 and a viewing angle of 34°, the MTF values can all be greater than 0.39 at 1/2 Nyquist frequency 220 lp/mm. And the length of the whole system is only 3 mm. According to the ZEMAX image quality evaluation method and tolerance analysis, the optimized results show that the designed lens can match well with all the optical requirements, and has the characters of good image quality, small volume, light weight, low cost and easy processing.
With the rapid commercialization of image sensors and the development of biometric recognition algorithms in recent years, iris recognition function has been applied to mobile terminal devices. The iris image acquisition is the first and the key step in iris recognition. Here one group of iris recognition lens was designed for mobile phones by the ZEMAX optical design software. This design is composed of three chip aspherical lenses of optical plastic, with OmniVision_OV2281 sensor of Howe technologies. With designed F parameter of 2.3 and a viewing angle of 34°, the MTF values can all be greater than 0.39 at 1/2 Nyquist frequency 220 lp/mm. And the length of the whole system is only 3 mm. According to the ZEMAX image quality evaluation method and tolerance analysis, the optimized results show that the designed lens can match well with all the optical requirements, and has the characters of good image quality, small volume, light weight, low cost and easy processing.
2020, 41(1): 43-54.
doi: 10.5768/JAO202041.0101007
Abstract:
Aiming at the image blurring problem caused by defocusing of sensor target in the use of large-aperture optical equipment, from the image restoration method, the defocus diffusion function and defocus transfer function of optical equipment were deduced, the judgment method of image defocus degree was proposed, the mountain climbing and the inverse filtering method were used to find the best defocus value and direction and restore the blurred image into the clear image. The real-time video focusing method of large-aperture optical equipment was given and verified by experiments. The results show that this method has important promotion and application value in automatic focusing of the outfield optical equipment.
Aiming at the image blurring problem caused by defocusing of sensor target in the use of large-aperture optical equipment, from the image restoration method, the defocus diffusion function and defocus transfer function of optical equipment were deduced, the judgment method of image defocus degree was proposed, the mountain climbing and the inverse filtering method were used to find the best defocus value and direction and restore the blurred image into the clear image. The real-time video focusing method of large-aperture optical equipment was given and verified by experiments. The results show that this method has important promotion and application value in automatic focusing of the outfield optical equipment.
2020, 41(1): 55-59.
doi: 10.5768/JAO202041.0101008
Abstract:
In order to analyze the electromagnetic shielding performance of optical windows of different substrate materials, based on Kohin's equivalent thin film model, the electromagnetic wave interface reflection coefficient was obtained considering the multiple reflections and refractions of electromagnetic waves between two interfaces in material. The matlab was used to calculate the shielding efficiency curves of the same mesh, different thicknesses and different materials, and the influence of thickness and material on the shielding efficiency of the light windows was analyzed. To verify the correctness of the simulation data, the metallic mesh with line width of 15 um, period of 500 um and resistance of ≤20 Ω was fabricated on ZnS, and the shielding effectiveness at 8 GHz～18 GHz was tested. Comparison results show that the testing results are in good agreement with the theoretical calculation data, and the error of the curves is about 2 dB to 4 dB. The result can estimate the shielding effectiveness of optical windows with metallic mesh coating, which canprovide references for the shielding effectiveness of optical window design subsequently.
In order to analyze the electromagnetic shielding performance of optical windows of different substrate materials, based on Kohin's equivalent thin film model, the electromagnetic wave interface reflection coefficient was obtained considering the multiple reflections and refractions of electromagnetic waves between two interfaces in material. The matlab was used to calculate the shielding efficiency curves of the same mesh, different thicknesses and different materials, and the influence of thickness and material on the shielding efficiency of the light windows was analyzed. To verify the correctness of the simulation data, the metallic mesh with line width of 15 um, period of 500 um and resistance of ≤20 Ω was fabricated on ZnS, and the shielding effectiveness at 8 GHz～18 GHz was tested. Comparison results show that the testing results are in good agreement with the theoretical calculation data, and the error of the curves is about 2 dB to 4 dB. The result can estimate the shielding effectiveness of optical windows with metallic mesh coating, which canprovide references for the shielding effectiveness of optical window design subsequently.
2020, 41(1): 60-66.
doi: 10.5768/JAO202041.0101009
Abstract:
In order to expound the temperature characteristics of homogeneous material reflective system in principle, and provide the basis for the design of optical system, the temperature characteristics of single mirror, coaxial multi-mirror reflective system and off-axis reflective system were introduced successively, which were based on the thermal deformation characteristics and its calculation method, and the optical system focus was taken as the research object. Meanwhile, the corresponding thermal aberration model of reflective system was established. The simulation results show that the change of focal length with temperature in the off-axis three-mirror reflective system is completely consistent with the thermal deformation amount of corresponding lens barrel, and it has good imaging quality in the range of large temperature difference. The thermal defocus caused by the temperature change can be compensated by the lens barrel, which verifies the correctness of the theoretical model and draw a conclusion that there is no thermal aberration in the homogeneous material reflective system in the condition of uniform temperature field. Therefore, the optical system, especially the infrared system with obvious thermal aberration effect has good temperature adaptability when the homogeneous material reflective system is adopted.
In order to expound the temperature characteristics of homogeneous material reflective system in principle, and provide the basis for the design of optical system, the temperature characteristics of single mirror, coaxial multi-mirror reflective system and off-axis reflective system were introduced successively, which were based on the thermal deformation characteristics and its calculation method, and the optical system focus was taken as the research object. Meanwhile, the corresponding thermal aberration model of reflective system was established. The simulation results show that the change of focal length with temperature in the off-axis three-mirror reflective system is completely consistent with the thermal deformation amount of corresponding lens barrel, and it has good imaging quality in the range of large temperature difference. The thermal defocus caused by the temperature change can be compensated by the lens barrel, which verifies the correctness of the theoretical model and draw a conclusion that there is no thermal aberration in the homogeneous material reflective system in the condition of uniform temperature field. Therefore, the optical system, especially the infrared system with obvious thermal aberration effect has good temperature adaptability when the homogeneous material reflective system is adopted.
2020, 41(1): 67-73.
doi: 10.5768/JAO202041.0102001
Abstract:
The lateral shearing interferometry based on wavelength phase-shifting is proposed to achieve the collimated wavefront reconstruction of an interferometer. The emergent wavefront of an interferometer is reflected by front and rear surfaces of a wedge respectively, and returns through a corner cube, forming shearing interferogram on the CCD of the interferometer. The wavelength phase-shifting method is used to extract the phase information of the shearing interference fringes to achieve the collimated wavefront reconstruction. The impact of the relative shear ratio on the accuracy of the wavefront reconstruction is analyzed and the relationship between the relative shear ratio and its influencing f actors is derived. The estimation method of the constant component of the OPD in wavelength phase-shifting is given. Three wavefronts have been measured and their PV (peak to valley) values are 3.22λ, 2.10λ and 0.83λ. This method simplifies traditional lateral shearing measurement of collimated wavefront and improves the measurement accuracy. It is particularly suitable for measuring emergent wavefront of wavelength phase-shifting interferometers.
The lateral shearing interferometry based on wavelength phase-shifting is proposed to achieve the collimated wavefront reconstruction of an interferometer. The emergent wavefront of an interferometer is reflected by front and rear surfaces of a wedge respectively, and returns through a corner cube, forming shearing interferogram on the CCD of the interferometer. The wavelength phase-shifting method is used to extract the phase information of the shearing interference fringes to achieve the collimated wavefront reconstruction. The impact of the relative shear ratio on the accuracy of the wavefront reconstruction is analyzed and the relationship between the relative shear ratio and its influencing f actors is derived. The estimation method of the constant component of the OPD in wavelength phase-shifting is given. Three wavefronts have been measured and their PV (peak to valley) values are 3.22λ, 2.10λ and 0.83λ. This method simplifies traditional lateral shearing measurement of collimated wavefront and improves the measurement accuracy. It is particularly suitable for measuring emergent wavefront of wavelength phase-shifting interferometers.
2020, 41(1): 74-78.
doi: 10.5768/JAO202041.0102002
Abstract:
The image needs to be binarized during the identification process of the high-voltage meter. However, the phenomenon of uneven illumination and digital ghosting often appears in the instrument images, so that it is difficult to binarize the instrument images with traditional method. Therefore, a binarization method based on convolutional neural network was proposed to binarize the instrument images with digital ghosting under complex illumination. The data sets used in the network were the instrument images in real environment. Firstly, the dimensionality reduction was used to extract features of the input images, and then the foreground of images was reconstructed by deconvolution. Finally, the binary images were output by the network. Comparing the designed network with the traditional binarization method, the experimental results show that the binary images of the proposed network are clear and have no ghosting. The average IoU is 95.12, which is most similar to the sample label images. Therefore, the method can effectively solve the problem of binarization of instrument images with ghosting under complex environment.
The image needs to be binarized during the identification process of the high-voltage meter. However, the phenomenon of uneven illumination and digital ghosting often appears in the instrument images, so that it is difficult to binarize the instrument images with traditional method. Therefore, a binarization method based on convolutional neural network was proposed to binarize the instrument images with digital ghosting under complex illumination. The data sets used in the network were the instrument images in real environment. Firstly, the dimensionality reduction was used to extract features of the input images, and then the foreground of images was reconstructed by deconvolution. Finally, the binary images were output by the network. Comparing the designed network with the traditional binarization method, the experimental results show that the binary images of the proposed network are clear and have no ghosting. The average IoU is 95.12, which is most similar to the sample label images. Therefore, the method can effectively solve the problem of binarization of instrument images with ghosting under complex environment.
2020, 41(1): 79-85.
doi: 10.5768/JAO202041.0102003
Abstract:
Stereo matching is one of the key steps in binocular stereo vision system, and its accuracy has a significant impact on the subsequent research. The Census algorithm is widely used because of its advantages of simplicity and clarity, good running effect and strong real-time performance. However, the Census stereo matching algorithm also has some disadvantages, such as the center point of transform window is easily disturbed by the external conditions and the accuracy of the depth discontinuous region is low. Therefore, a new stereo matching algorithm based on Census transformation and guided filter was proposed. In the stage of Census transformation, the impact of the external interference was reduced by calculating the pixel average around the transform window. And at the same time, in the stage of cost aggregation, the guided filter with tipping characteristics and its computational complexity was independent of the size for filter nuclear was introduced as the adaptive weight. The experimental results show that the average mistake matching error of the proposed algorithm is 6.03% on the Middlebury test platform. Compared with the current average mistake matching rate of 16.2% for Census stereo matching algorithm, the matching effect is obviously improved, and the algorithm has high efficiency and good radiation invariance.
Stereo matching is one of the key steps in binocular stereo vision system, and its accuracy has a significant impact on the subsequent research. The Census algorithm is widely used because of its advantages of simplicity and clarity, good running effect and strong real-time performance. However, the Census stereo matching algorithm also has some disadvantages, such as the center point of transform window is easily disturbed by the external conditions and the accuracy of the depth discontinuous region is low. Therefore, a new stereo matching algorithm based on Census transformation and guided filter was proposed. In the stage of Census transformation, the impact of the external interference was reduced by calculating the pixel average around the transform window. And at the same time, in the stage of cost aggregation, the guided filter with tipping characteristics and its computational complexity was independent of the size for filter nuclear was introduced as the adaptive weight. The experimental results show that the average mistake matching error of the proposed algorithm is 6.03% on the Middlebury test platform. Compared with the current average mistake matching rate of 16.2% for Census stereo matching algorithm, the matching effect is obviously improved, and the algorithm has high efficiency and good radiation invariance.
2020, 41(1): 86-93.
doi: 10.5768/JAO202041.0102004
Abstract:
In machine vision, the collected images with distortion need to be corrected. In order to improve the real-time performance of distorted images correction, aiming at the problem of complex on-line calculation of reverse mapping coordinates and insufficient ROM capacity on chip when using FPGA(Field-Programmable Gate Array) to realize the algorithm of image distortion correction, the compressed reverse mapping table was reconstructed online by interpolation method on FPGA. The reverse mapping coordinates were obtained by seeking the reconstructed reverse mapping table, which reduced the online computation of FPGA and the capacity requirement of ROM on chip. The simulation results of MATLAB show that when the compressed parameters are 4, 8 and 16, respectively,the distortion images can be corrected better and the image information cannot be lost. The simulation results verify the effectiveness of algorithm, which can be used for image distortion correction based on FPGA.
In machine vision, the collected images with distortion need to be corrected. In order to improve the real-time performance of distorted images correction, aiming at the problem of complex on-line calculation of reverse mapping coordinates and insufficient ROM capacity on chip when using FPGA(Field-Programmable Gate Array) to realize the algorithm of image distortion correction, the compressed reverse mapping table was reconstructed online by interpolation method on FPGA. The reverse mapping coordinates were obtained by seeking the reconstructed reverse mapping table, which reduced the online computation of FPGA and the capacity requirement of ROM on chip. The simulation results of MATLAB show that when the compressed parameters are 4, 8 and 16, respectively,the distortion images can be corrected better and the image information cannot be lost. The simulation results verify the effectiveness of algorithm, which can be used for image distortion correction based on FPGA.
2020, 41(1): 94-99.
doi: 10.5768/JAO202041.0102005
Abstract:
Aiming at the degradation problem of the images collected by the optical imaging equipment in hazy days, a Dual Dehazing Network based on dual learning from the source domain to the target domain is proposed to realize the image dehazing. First, the network learns the bilateral mapping relationship between the hazy image and the haze-free image by using the Dual Generative Adversarial Network, and obtains the preliminary dehazing result. Then the pre-training model is used to calculate the feature vector of the dehazed image and the real haze-free image in the feature space. The Euclidean distance is used as the loss function to minimize the distance between the feature vectors to ensure that the dehazed image is close to the real haze-free image at the feature level. The experimental results show that the dehazing results obtained by the Dual Dehazing network have higher peak signal-to-noise ratio and lower color difference, and can effectively preserve the structural information of the image.
Aiming at the degradation problem of the images collected by the optical imaging equipment in hazy days, a Dual Dehazing Network based on dual learning from the source domain to the target domain is proposed to realize the image dehazing. First, the network learns the bilateral mapping relationship between the hazy image and the haze-free image by using the Dual Generative Adversarial Network, and obtains the preliminary dehazing result. Then the pre-training model is used to calculate the feature vector of the dehazed image and the real haze-free image in the feature space. The Euclidean distance is used as the loss function to minimize the distance between the feature vectors to ensure that the dehazed image is close to the real haze-free image at the feature level. The experimental results show that the dehazing results obtained by the Dual Dehazing network have higher peak signal-to-noise ratio and lower color difference, and can effectively preserve the structural information of the image.
2020, 41(1): 100-106.
doi: 10.5768/JAO202041.0102006
Abstract:
Aiming at the small transmittance estimate and the Halo effect caused by the sudden change in the depth of field of the dark channel prior algorithm in the large sky area, an image defogging algorithm combined with lab space and single-scale retinex was proposed. Firstly, the RGB image was transferred to the lab space to extract the luminance component, and the edge information of this was extracted by the Canny operator to enrich the details of the restored image. Secondly, the single-scale retinex was used to perform a Gaussian adaptive filtering of the non-edge region to estimate the optimized luminance component, and the “false” defogging image as well as the rough transmittance was obtained. Then, the cross-bilateral filtering was used to eliminate the Halo effect by optimizing the transmittance. Finally, the defogging image was recovered on the basis of the atmospheric scattering model. The experimental results show that the image recovered by this method is clear and smooth overall, and also has a good recovery effect on images with large sky areas.
Aiming at the small transmittance estimate and the Halo effect caused by the sudden change in the depth of field of the dark channel prior algorithm in the large sky area, an image defogging algorithm combined with lab space and single-scale retinex was proposed. Firstly, the RGB image was transferred to the lab space to extract the luminance component, and the edge information of this was extracted by the Canny operator to enrich the details of the restored image. Secondly, the single-scale retinex was used to perform a Gaussian adaptive filtering of the non-edge region to estimate the optimized luminance component, and the “false” defogging image as well as the rough transmittance was obtained. Then, the cross-bilateral filtering was used to eliminate the Halo effect by optimizing the transmittance. Finally, the defogging image was recovered on the basis of the atmospheric scattering model. The experimental results show that the image recovered by this method is clear and smooth overall, and also has a good recovery effect on images with large sky areas.
2020, 41(1): 107-113, 194.
doi: 10.5768/JAO202041.0102007
Abstract:
In order to improve the effect of backscattering light on imaging in the conditions of underwater active light illumination, an underwater active polarization defogging algorithm for global parameter estimation based on the fitting function was proposed. First, this algorithm combined with the image enhancement as the image preprocessing. Then, the binary polynomial function was set and the least square method was used to solve the fitting of backscattering light polarization degree variable. Finally, the underwater restored image with higher contrast and richer information was obtained. The experimental results show that the proposed algorithm can effectively improve the underwater image quality in the conditions of active light illumination, improve the image contrast and restore the submerged details. The image enhancement measured value of the restored image is 70% higher than that of the previous algorithms, and it can be applied in medias of different concentrations.
In order to improve the effect of backscattering light on imaging in the conditions of underwater active light illumination, an underwater active polarization defogging algorithm for global parameter estimation based on the fitting function was proposed. First, this algorithm combined with the image enhancement as the image preprocessing. Then, the binary polynomial function was set and the least square method was used to solve the fitting of backscattering light polarization degree variable. Finally, the underwater restored image with higher contrast and richer information was obtained. The experimental results show that the proposed algorithm can effectively improve the underwater image quality in the conditions of active light illumination, improve the image contrast and restore the submerged details. The image enhancement measured value of the restored image is 70% higher than that of the previous algorithms, and it can be applied in medias of different concentrations.
2020, 41(1): 114-119.
doi: 10.5768/JAO202041.0102008
Abstract:
Aiming at the problem of insufficient estimate of transmittance and atmospheric light in linear transformation algorithm, an adaptive optimization defogging algorithm based on linear model was proposed. First, the edge information model was used to enhance the detailed information of the initial transmittance image, so that the edge region details of the restored image were richer. Then, an adaptive optimization transmittance was obtained to better process the image including the depth of field region by the dark channel prior. Finally, the local atmospheric light estimate method was used instead of the quadtree method to avoid the inaccuracy of atmospheric light estimate, and the image was restored by combining with the physical model. The simulation experiment was carried out in matlab2014, and the experimental results show that the proposed algorithm has good validity and timeliness.
Aiming at the problem of insufficient estimate of transmittance and atmospheric light in linear transformation algorithm, an adaptive optimization defogging algorithm based on linear model was proposed. First, the edge information model was used to enhance the detailed information of the initial transmittance image, so that the edge region details of the restored image were richer. Then, an adaptive optimization transmittance was obtained to better process the image including the depth of field region by the dark channel prior. Finally, the local atmospheric light estimate method was used instead of the quadtree method to avoid the inaccuracy of atmospheric light estimate, and the image was restored by combining with the physical model. The simulation experiment was carried out in matlab2014, and the experimental results show that the proposed algorithm has good validity and timeliness.
2020, 41(1): 120-126.
doi: 10.5768/JAO202041.0102009
Abstract:
The method of sample generation and automatic annotation based on the generative countermeasure network (GAN) and cross-domain adaptive transfer learning was studied. In this method, the adaptive transfer learning network is used to explore the intrinsic correlation of target features in infrared and visible images based on the small number of existing visible image samples, and the adaptive transfer learning network model is constructed to generate tagged target images. The problem of small number of infrared image samples and time-consuming labeling can be solved by proposed method, which provides enough sample data for subsequent multi-band cooperative target detection and recognition. Moreover, automatic standard tests were carries out on the 1 000 pieces of actual acquired and 1 000 pieces of generated armored target images ,respectively,by using the automatic standard algorithm The experimental results show that the accuracy of the actual armored target image labeling is more than 95%, and that of the generated armored target image labeling is more than 83%. The performance of classifiers trained with the mixed dataset of real images and generate images is basically the same as when using the pure real images.
The method of sample generation and automatic annotation based on the generative countermeasure network (GAN) and cross-domain adaptive transfer learning was studied. In this method, the adaptive transfer learning network is used to explore the intrinsic correlation of target features in infrared and visible images based on the small number of existing visible image samples, and the adaptive transfer learning network model is constructed to generate tagged target images. The problem of small number of infrared image samples and time-consuming labeling can be solved by proposed method, which provides enough sample data for subsequent multi-band cooperative target detection and recognition. Moreover, automatic standard tests were carries out on the 1 000 pieces of actual acquired and 1 000 pieces of generated armored target images ,respectively,by using the automatic standard algorithm The experimental results show that the accuracy of the actual armored target image labeling is more than 95%, and that of the generated armored target image labeling is more than 83%. The performance of classifiers trained with the mixed dataset of real images and generate images is basically the same as when using the pure real images.
2020, 41(1): 127-133.
doi: 10.5768/JAO202041.0102010
Abstract:
The adjacent meshes of the urban digital surface model (UDSM) are often subject to curvature upheavals, and these locations are details of the UDSM that should be preserved as much as possible during simplification. In order to solve this problem, the centroidal Voronoi diagram was introduced to repartition the mesh, and the point cloud density of the surface with small curvature could be greatly reduced. The size of the redrawn mesh surface was greatly different from that of the surrounding smooth triangular mesh. On this basis, the quadric error matrix edge collapse was used for levels of details (LOD) construction, and the area of obvious changes in the mesh could be greatly reduced. On the premise that the time performance and mesh error are similar to that of the direct edge simplification, the algorithm can save more simplified mesh details.
The adjacent meshes of the urban digital surface model (UDSM) are often subject to curvature upheavals, and these locations are details of the UDSM that should be preserved as much as possible during simplification. In order to solve this problem, the centroidal Voronoi diagram was introduced to repartition the mesh, and the point cloud density of the surface with small curvature could be greatly reduced. The size of the redrawn mesh surface was greatly different from that of the surrounding smooth triangular mesh. On this basis, the quadric error matrix edge collapse was used for levels of details (LOD) construction, and the area of obvious changes in the mesh could be greatly reduced. On the premise that the time performance and mesh error are similar to that of the direct edge simplification, the algorithm can save more simplified mesh details.
2020, 41(1): 134-139.
doi: 10.5768/JAO202041.0103001
Abstract:
Modulation transfer function (MTF) is an important parameter to evaluate the imaging performance of TV cameras. A measuring device and method for MTF of visible TV camera were introduced. A three-integral spherical target generator system was designed, the target and background brightness can be independently controlled by electric attenuators. The outlet non-uniformity is 4.1%. An automatic sub-pixel interpolation algorithm based on sliting slit method and a least square method of line spread functions were wtilized to realize the angle calculation automaticly and MTF measurement accurately. The repeatability of the measuring device is 0.007, and the measurement uncertainty is U=0.05 (k=2). Finally, the comparison experiment with foreign advanced instruments shows that, the measurement device established can effectively solve the problem of evaluating the image quality of TV cameras.
Modulation transfer function (MTF) is an important parameter to evaluate the imaging performance of TV cameras. A measuring device and method for MTF of visible TV camera were introduced. A three-integral spherical target generator system was designed, the target and background brightness can be independently controlled by electric attenuators. The outlet non-uniformity is 4.1%. An automatic sub-pixel interpolation algorithm based on sliting slit method and a least square method of line spread functions were wtilized to realize the angle calculation automaticly and MTF measurement accurately. The repeatability of the measuring device is 0.007, and the measurement uncertainty is U=0.05 (k=2). Finally, the comparison experiment with foreign advanced instruments shows that, the measurement device established can effectively solve the problem of evaluating the image quality of TV cameras.
2020, 41(1): 140-144.
doi: 10.5768/JAO202041.0103002
Abstract:
Aiming at on-orbit radiation calibration of spaceborne low-light-level remote sensor, an out-site radiation (LED) light source based on the theory of radiation flux equivalence is designed. The designed calibration light source uses twelve 670 nm pick-wavelength LED modules, with each LED module max power dissipation of 300 W, and the max output radiance over 5.2×10−2 W·cm−2·sr−1. Based on the calibrating area of equivalent radiation flux, it can be applied to the on-orbit calibration of low-light-level channels in the order of 10−9 W·cm−2·sr−1. The test results show that the planar uniformity of the source is better than 99.7% at the exit port, the angular uniformity is better than 99.6% within ±30° to the exit normal, and the stability of source is better than 99.9% within 1 h. In-orbit test of the low-light-level channel of VIIRS was carried out at the radiation correction field of Dunhuang Chinese remote sensing satellite, on-orbit response results was 8.27×10−9 W·cm−2·sr−1. (including moonlight contribution and atmospheric impact). The rationality of the design of the radiation value of the out-site low-light-level calibration light source proposed in this paper is proved.
Aiming at on-orbit radiation calibration of spaceborne low-light-level remote sensor, an out-site radiation (LED) light source based on the theory of radiation flux equivalence is designed. The designed calibration light source uses twelve 670 nm pick-wavelength LED modules, with each LED module max power dissipation of 300 W, and the max output radiance over 5.2×10−2 W·cm−2·sr−1. Based on the calibrating area of equivalent radiation flux, it can be applied to the on-orbit calibration of low-light-level channels in the order of 10−9 W·cm−2·sr−1. The test results show that the planar uniformity of the source is better than 99.7% at the exit port, the angular uniformity is better than 99.6% within ±30° to the exit normal, and the stability of source is better than 99.9% within 1 h. In-orbit test of the low-light-level channel of VIIRS was carried out at the radiation correction field of Dunhuang Chinese remote sensing satellite, on-orbit response results was 8.27×10−9 W·cm−2·sr−1. (including moonlight contribution and atmospheric impact). The rationality of the design of the radiation value of the out-site low-light-level calibration light source proposed in this paper is proved.
2020, 41(1): 145-149.
doi: 10.5768/JAO202041.0103003
Abstract:
During the flight training of one machine, the two-dimensional relative attitude data of the approach segment plays a crucial role in the safe landing of the aircraft. The two-way data transmission system based on the wireless network radio station can obtain the horizontal and vertical relative position data of the approaching aircraft relative to the ideal landing point in real time. The attitude data of the landing point is uploaded through the uplink, the differential data after the dynamic difference is driven in the downlink, and finally the secondary data processing is performed at the central control station. The two-dimensional relative attitude accuracy of the approach segment obtained by the technology reaches the centimeter level, which satisfies the requirements of flight training. At the same time, combined with the aircraft’s display control data and the visual map, the real-time acquired attitude data is presented to the commander in a multi-angled dynamic association manner, and the assistant commander is better assisted in the landing command work.
During the flight training of one machine, the two-dimensional relative attitude data of the approach segment plays a crucial role in the safe landing of the aircraft. The two-way data transmission system based on the wireless network radio station can obtain the horizontal and vertical relative position data of the approaching aircraft relative to the ideal landing point in real time. The attitude data of the landing point is uploaded through the uplink, the differential data after the dynamic difference is driven in the downlink, and finally the secondary data processing is performed at the central control station. The two-dimensional relative attitude accuracy of the approach segment obtained by the technology reaches the centimeter level, which satisfies the requirements of flight training. At the same time, combined with the aircraft’s display control data and the visual map, the real-time acquired attitude data is presented to the commander in a multi-angled dynamic association manner, and the assistant commander is better assisted in the landing command work.
2020, 41(1): 150-155.
doi: 10.5768/JAO202041.0103004
Abstract:
For the task of detecting objects on the ground such as vehicles, it’s difficult to obtain good detection results for the reason of small size and little appearance information of objects and the interference of complex background. Aiming at the problem of accurate localization of small size targets on the ground, a sub-network with feature fusion was proposed from the perspective of target feature extraction. The fusion network introduced important context information, and effectively improved the precision of small target detection.To solve the problem of transformation of sizes and angles, a extensional feature pyramid network was designed as prediction module based on fused feature. Prediction boxes were generated on different scales of extensional feature layers to match the specific objects.Experiments were conducted on the small vehicle target data set VEDAI (vehicle detection in aerial imagery).Results indicate that while the algorithm retains the advantages of detection speed of traditional SSD (single-shot multibox detector), it can significantly improve the accuracy and greatly improve the practicability of the algorithm.
For the task of detecting objects on the ground such as vehicles, it’s difficult to obtain good detection results for the reason of small size and little appearance information of objects and the interference of complex background. Aiming at the problem of accurate localization of small size targets on the ground, a sub-network with feature fusion was proposed from the perspective of target feature extraction. The fusion network introduced important context information, and effectively improved the precision of small target detection.To solve the problem of transformation of sizes and angles, a extensional feature pyramid network was designed as prediction module based on fused feature. Prediction boxes were generated on different scales of extensional feature layers to match the specific objects.Experiments were conducted on the small vehicle target data set VEDAI (vehicle detection in aerial imagery).Results indicate that while the algorithm retains the advantages of detection speed of traditional SSD (single-shot multibox detector), it can significantly improve the accuracy and greatly improve the practicability of the algorithm.
2020, 41(1): 156-162.
doi: 10.5768/JAO202041.0105001
Abstract:
The laser scanning micro electromechanical device projection system is realized by the two-dimensional deflection characteristics of micro electromechanical device and the laser diode as the light source. Compared with the traditional micro-projection technology, the micro-projection display system based on the micro electromechanical device has the advantages of small size, low energy consumption and high image contrast. According to the characteristics of the components, the laser beam was collimated, concentrated and homogenized. The post-scanning system of the objective lens was designed by using the reflection deflection characteristics of two-dimensional micro electromechanical device, and the scanning methods as well as the pixel arrangement were optimally analyzed. The design results show that the size of the image plane is 104mm×59mm, and the single-pixel point geometric radius is within 186 μm when 80% energy is included. The resolution of the system can reach 558×314 pixel, and the refresh rate is 30Hz. The design results can be used in the small projection systems such as vehicle head-up display system, and can also provide reference for similar designs in the future.
The laser scanning micro electromechanical device projection system is realized by the two-dimensional deflection characteristics of micro electromechanical device and the laser diode as the light source. Compared with the traditional micro-projection technology, the micro-projection display system based on the micro electromechanical device has the advantages of small size, low energy consumption and high image contrast. According to the characteristics of the components, the laser beam was collimated, concentrated and homogenized. The post-scanning system of the objective lens was designed by using the reflection deflection characteristics of two-dimensional micro electromechanical device, and the scanning methods as well as the pixel arrangement were optimally analyzed. The design results show that the size of the image plane is 104mm×59mm, and the single-pixel point geometric radius is within 186 μm when 80% energy is included. The resolution of the system can reach 558×314 pixel, and the refresh rate is 30Hz. The design results can be used in the small projection systems such as vehicle head-up display system, and can also provide reference for similar designs in the future.
2020, 41(1): 163-169.
doi: 10.5768/JAO202041.0105002
Abstract:
In order to optimize the phase reconstruction algorithm, the distribution characteristics and spectral response of different filtering windows were analyzed for the Fourier spectrum of the wavefront interferogram. The optimal filtering window type of the fast Fourier transform (FFT) dynamic phase reconstruction algorithm was determined by computer simulation and experimental test. The residual error of the wavefront peak-valley value between the reconstructed wavefront by the simulated interferogramand the original wavefront is 0.008 5λ, as well as the residual of the wavefront root-mean-square is 0.000 1λ,while the residual of the wavefront peak-valley value between the obtained wavefront by the experimental interferogram and by the phase-shifting interferometry is 0.009 3λ, as well as the residual of the wavefront root-mean-square is 0.000 5λ. The results show that, after fitting the phase filtered and reconstructed by Hamming window, the residual surface error of the wavefront obtained is smaller that the reference wavefront, and the phase reconstruction accuracy is better than 0.01λ, which can be further applied to the measurement of large-caliber optical elements.
In order to optimize the phase reconstruction algorithm, the distribution characteristics and spectral response of different filtering windows were analyzed for the Fourier spectrum of the wavefront interferogram. The optimal filtering window type of the fast Fourier transform (FFT) dynamic phase reconstruction algorithm was determined by computer simulation and experimental test. The residual error of the wavefront peak-valley value between the reconstructed wavefront by the simulated interferogramand the original wavefront is 0.008 5λ, as well as the residual of the wavefront root-mean-square is 0.000 1λ,while the residual of the wavefront peak-valley value between the obtained wavefront by the experimental interferogram and by the phase-shifting interferometry is 0.009 3λ, as well as the residual of the wavefront root-mean-square is 0.000 5λ. The results show that, after fitting the phase filtered and reconstructed by Hamming window, the residual surface error of the wavefront obtained is smaller that the reference wavefront, and the phase reconstruction accuracy is better than 0.01λ, which can be further applied to the measurement of large-caliber optical elements.
2020, 41(1): 170-175.
doi: 10.5768/JAO202041.0105003
Abstract:
In order to solve the difficult problems of collimation, irradiance uniformity and spectral matching of existing LED solar simulators, a simple, efficient and high-power optical system design method of LED solar simulator was proposed based on photoelectric integrated secondary optical design. The small angle collimating lens and the parabolic mirror were utilized to integrate the light source, and the mixing light rod and microstructure were used to homogenize the light, finally the collimating output of light was realized by the parabolic mirror. Based on the design idea of coaxial collimation, the design of the LED solar simulator was completed.In addition, the optical software LightTools was used to simulate and optimize the optical system of LED solar simulator. The experimental results show that the irradiation inhomogeneity is 2.5% and the collimation angle is 1.5° on the effective irradiation surface with a diameter of 260 mm.
In order to solve the difficult problems of collimation, irradiance uniformity and spectral matching of existing LED solar simulators, a simple, efficient and high-power optical system design method of LED solar simulator was proposed based on photoelectric integrated secondary optical design. The small angle collimating lens and the parabolic mirror were utilized to integrate the light source, and the mixing light rod and microstructure were used to homogenize the light, finally the collimating output of light was realized by the parabolic mirror. Based on the design idea of coaxial collimation, the design of the LED solar simulator was completed.In addition, the optical software LightTools was used to simulate and optimize the optical system of LED solar simulator. The experimental results show that the irradiation inhomogeneity is 2.5% and the collimation angle is 1.5° on the effective irradiation surface with a diameter of 260 mm.
2020, 41(1): 176-179.
doi: 10.5768/JAO202041.0106001
Abstract:
A planar array structure of multi-channel narrowband filter was proposed, which consisted of bilayer metallic grating and its bilateral distributed Bragg reflectors (DBRs). The central wavelength range of the designed multi-channel filter exceeded 200 nm, and the central wavelength of each channel could be adjusted by grating slit width to realize the multi-channel narrowband filter array. The results provides a new approach for designing the multi-channel mid-infrared narrowband filter.
A planar array structure of multi-channel narrowband filter was proposed, which consisted of bilayer metallic grating and its bilateral distributed Bragg reflectors (DBRs). The central wavelength range of the designed multi-channel filter exceeded 200 nm, and the central wavelength of each channel could be adjusted by grating slit width to realize the multi-channel narrowband filter array. The results provides a new approach for designing the multi-channel mid-infrared narrowband filter.
Semantic segmentation of nocturnal infrared images of unmanned vehicles based on improved DeepLabv3+
2020, 41(1): 180-185.
doi: 10.5768/JAO202041.0106002
Abstract:
In order to enhance the understanding ability of unmanned vehicle to night scene, a semantic segmentation algorithm based on improved DeepLabv3+ network is proposed for infrared images acquired by unmanned vehicles at night. Because the objects in the autopilot scene often show very large scale changes, the algorithm based on DeepLabv3+ network can cover a larger scale range by introducing densely connected atrous spatial pyramid pooling module. In addition, the algorithm splices the multi-layer results of the encoder module into the decoder module to recover more spatial information and low-level features lost in the downsampling process. Through end-to-end learning and training, it can be directly used for semantics segmentation of night vision infrared images. The experimental results show that the segmentation accuracy of the algorithm is better than that of the original DeepLabv3+ algorithm, and the mean intersection over union reaches 80.42, which has good real-time performance and accuracy.
In order to enhance the understanding ability of unmanned vehicle to night scene, a semantic segmentation algorithm based on improved DeepLabv3+ network is proposed for infrared images acquired by unmanned vehicles at night. Because the objects in the autopilot scene often show very large scale changes, the algorithm based on DeepLabv3+ network can cover a larger scale range by introducing densely connected atrous spatial pyramid pooling module. In addition, the algorithm splices the multi-layer results of the encoder module into the decoder module to recover more spatial information and low-level features lost in the downsampling process. Through end-to-end learning and training, it can be directly used for semantics segmentation of night vision infrared images. The experimental results show that the segmentation accuracy of the algorithm is better than that of the original DeepLabv3+ algorithm, and the mean intersection over union reaches 80.42, which has good real-time performance and accuracy.
2020, 41(1): 186-193.
doi: 10.5768/JAO202041.0107001
Abstract:
Tunable lasers are widely used in optical fiber sensing and optical fiber communications, which have attracted extensive attention. A tunable laser system based on FFP-TF2 filter was designed and built. The driving voltage generated by the DDS(direct digital frequency synthesis) function signal generator and the analog amplifying circuit designed by the FPGA was adjusted to change the cavity length of the FFP-TF2, thereby changing the output of the laser wavelength. This system could drive the single point voltage and scanning voltage of FFP-TF2, so that the single-wavelength laser and the multi-wavelength scanning laser could be output by laser. In the laser output wavelength range of 1 532 nm to 1 568 nm, the single-wavelength output and the multi-wavelength scanning range could be adjusted, and the two modes could be switched. Experimental results show that the 3 dB linewidth of the laser output is kept at about 0.01 nm, the linear fitting degree of the laser output center wavelength and the driving voltage is 99.934%, the sensitivity is 3.915 nm/V, while the linear fitting degree of the FFP-TF2 output center wavelength and the driving voltage is 99.986%, the sensitivity is 4.021 nm/V. And the error of the laser output center wavelength relative to FFP-TF2 is 2.6%, which indicates that the laser system output wavelength has the advantages of narrow line width, good stability and high flexibility.
Tunable lasers are widely used in optical fiber sensing and optical fiber communications, which have attracted extensive attention. A tunable laser system based on FFP-TF2 filter was designed and built. The driving voltage generated by the DDS(direct digital frequency synthesis) function signal generator and the analog amplifying circuit designed by the FPGA was adjusted to change the cavity length of the FFP-TF2, thereby changing the output of the laser wavelength. This system could drive the single point voltage and scanning voltage of FFP-TF2, so that the single-wavelength laser and the multi-wavelength scanning laser could be output by laser. In the laser output wavelength range of 1 532 nm to 1 568 nm, the single-wavelength output and the multi-wavelength scanning range could be adjusted, and the two modes could be switched. Experimental results show that the 3 dB linewidth of the laser output is kept at about 0.01 nm, the linear fitting degree of the laser output center wavelength and the driving voltage is 99.934%, the sensitivity is 3.915 nm/V, while the linear fitting degree of the FFP-TF2 output center wavelength and the driving voltage is 99.986%, the sensitivity is 4.021 nm/V. And the error of the laser output center wavelength relative to FFP-TF2 is 2.6%, which indicates that the laser system output wavelength has the advantages of narrow line width, good stability and high flexibility.
2020, 41(1): 194-201.
doi: 10.5768/JAO202041.0107002
Abstract:
With the development of space industry and new energy technology, small electric unmanned aerial vehicles (UAVs) have higher application value in modern warfare and scientific research,etc. The proposal of laser wireless energy transmission technology can effectively solve the problem of short life of small electric UAV, and greatly improve the work efficiency of UAV. Based on the structure principle of UAV laser energy supply system, a maximum power point optimization tracking method for the characteristics of laser wireless power supply to small electric UAV was proposed, which referred to the optimal control algorithm combined with the constant voltage (CV) method and the firefly algorithm (FA). After the laser was projected onto the photovoltaic panel on the drone, by tracking the maximum power point during the drone laser wireless charging process, the laser utilization got higher and the charging was more stable. Finally, numerical simulation was conducted and results can verified the feasibility and efficiency of the proposed algorithm.
With the development of space industry and new energy technology, small electric unmanned aerial vehicles (UAVs) have higher application value in modern warfare and scientific research,etc. The proposal of laser wireless energy transmission technology can effectively solve the problem of short life of small electric UAV, and greatly improve the work efficiency of UAV. Based on the structure principle of UAV laser energy supply system, a maximum power point optimization tracking method for the characteristics of laser wireless power supply to small electric UAV was proposed, which referred to the optimal control algorithm combined with the constant voltage (CV) method and the firefly algorithm (FA). After the laser was projected onto the photovoltaic panel on the drone, by tracking the maximum power point during the drone laser wireless charging process, the laser utilization got higher and the charging was more stable. Finally, numerical simulation was conducted and results can verified the feasibility and efficiency of the proposed algorithm.
2020, 41(1): 202-208.
doi: 10.5768/JAO202041.0107003
Abstract:
In order to ensure the scanning quality of F-θ lens for semiconductor laser marking machine and realize the linear change of the image height and the scanning angle, it is necessary to give a certain distortion to the F-θ lens and satisfy the isoplanatic condition. Firstly, the working principle and aberration requirements of the F-θ lens were analyzed, and the appropriate glass material was selected by the light source imaging requirements of 1 064 nm semiconductor laser marking machine to reasonably distribute the focal power of each lens to ensure the isoplanatic imaging. Secondly, according to the linear imaging requirements of F-θ lens, the total distortion of the system was 1.6%, which was the sum of the actual barrel distortion and the relative distortion. Finally, in the optimization design of the optical system, these two optimal parameters were introduced to observe the system imaging changes during the optimization process. The design results show that the system modulation transfer function (MTF) curve is closed to the diffraction limit, the relative distortion of the F-θ lens is less than 0.36%, the root-mean-square (RMS) radius of each field of view (FOV) is smaller than the Airy disk diameter, and 70% energy of the whole system is concentrated in a circle with a diameter of 16 μm, the total distortion of the system is 1.58%, which meets the requirements of design index.
In order to ensure the scanning quality of F-θ lens for semiconductor laser marking machine and realize the linear change of the image height and the scanning angle, it is necessary to give a certain distortion to the F-θ lens and satisfy the isoplanatic condition. Firstly, the working principle and aberration requirements of the F-θ lens were analyzed, and the appropriate glass material was selected by the light source imaging requirements of 1 064 nm semiconductor laser marking machine to reasonably distribute the focal power of each lens to ensure the isoplanatic imaging. Secondly, according to the linear imaging requirements of F-θ lens, the total distortion of the system was 1.6%, which was the sum of the actual barrel distortion and the relative distortion. Finally, in the optimization design of the optical system, these two optimal parameters were introduced to observe the system imaging changes during the optimization process. The design results show that the system modulation transfer function (MTF) curve is closed to the diffraction limit, the relative distortion of the F-θ lens is less than 0.36%, the root-mean-square (RMS) radius of each field of view (FOV) is smaller than the Airy disk diameter, and 70% energy of the whole system is concentrated in a circle with a diameter of 16 μm, the total distortion of the system is 1.58%, which meets the requirements of design index.
2020, 41(1): 209-213.
doi: 10.5768/JAO202041.0107004
Abstract:
For the demand of auxiliary driving, a vehicle-mounted LIDAR system was built to scan road scene. A set of method for calculating road parameters was designed based on 3D LIDAR point cloud, including road width, bridge or tunnel height limitation, turnaround radius and so on, which provided decision basis for auxiliary judgment of road passability. Experiments and comparison tests show the stable performance and satisfactory accuracy of algorithms, which have advantages of efficiency and automation, and also can provide useful decision parameters for auxiliary driving.
For the demand of auxiliary driving, a vehicle-mounted LIDAR system was built to scan road scene. A set of method for calculating road parameters was designed based on 3D LIDAR point cloud, including road width, bridge or tunnel height limitation, turnaround radius and so on, which provided decision basis for auxiliary judgment of road passability. Experiments and comparison tests show the stable performance and satisfactory accuracy of algorithms, which have advantages of efficiency and automation, and also can provide useful decision parameters for auxiliary driving.
2020, 41(1): 214-219.
doi: 10.5768/JAO202041.0107005
Abstract:
In order to study the propagation law of laser-excited ultrasonic waves on metal plate with transitional fillet and the method of detecting surface defects, the finite element method was used to simulate the laser ultrasonic phenomenon in the plate, and the propagation process of surface wave in the rounded area and its interaction with defects were analyzed.The numerical result shows that, the laser excites the longitudinal wave, the transverse wave and the Rayleigh wave, and the Rayleigh wave mainly exists on the surface of millimeter magnitude. At the transition fillet, these surface waves are converted in mode, and various surface waves such as direct wave R′ and infrasound source generated wave RR are generated;the waves passing through the transition area are reflected and transmitted at the surface defects, and the position of the defect can be detected by the B-scan. As the depth of the defect increases, the transmission coefficient constantly decreases,and there is a difference in arrival time of about 0.5 μs between the transmitted waves Rt and Rst, which is linearly positively correlated with the depth of the defect. The numerical results provide valuable references for laser ultrasonic detection of plate surface defects with transition fillet.
In order to study the propagation law of laser-excited ultrasonic waves on metal plate with transitional fillet and the method of detecting surface defects, the finite element method was used to simulate the laser ultrasonic phenomenon in the plate, and the propagation process of surface wave in the rounded area and its interaction with defects were analyzed.The numerical result shows that, the laser excites the longitudinal wave, the transverse wave and the Rayleigh wave, and the Rayleigh wave mainly exists on the surface of millimeter magnitude. At the transition fillet, these surface waves are converted in mode, and various surface waves such as direct wave R′ and infrasound source generated wave RR are generated;the waves passing through the transition area are reflected and transmitted at the surface defects, and the position of the defect can be detected by the B-scan. As the depth of the defect increases, the transmission coefficient constantly decreases,and there is a difference in arrival time of about 0.5 μs between the transmitted waves Rt and Rst, which is linearly positively correlated with the depth of the defect. The numerical results provide valuable references for laser ultrasonic detection of plate surface defects with transition fillet.
2020, 41(1): 220-227.
doi: 10.5768/JAO202041.0108001
Abstract:
The formation principle of zero error on high precision fiber-optic gyro was introduced, and the impact of temperature and magnetic field on the zero error of high precision fiber-optic gyro was analyzed in detail. The simulation results show that the zero-bias peak value drift caused by temperature is 0.06°/h. Finally, the theoretical model and experimental research were carried out under the cross-linking condition of alternating temperature field and radial static magnetic field. The experimental results show that under the action of radial magnetic field and temperature, the maximum drift value of the fiber-optic gyro generated by the cross-linking coupling effect is about 1% order of the sensitivity of the radial magnetic field, which is consistent with the theoretical model.
The formation principle of zero error on high precision fiber-optic gyro was introduced, and the impact of temperature and magnetic field on the zero error of high precision fiber-optic gyro was analyzed in detail. The simulation results show that the zero-bias peak value drift caused by temperature is 0.06°/h. Finally, the theoretical model and experimental research were carried out under the cross-linking condition of alternating temperature field and radial static magnetic field. The experimental results show that under the action of radial magnetic field and temperature, the maximum drift value of the fiber-optic gyro generated by the cross-linking coupling effect is about 1% order of the sensitivity of the radial magnetic field, which is consistent with the theoretical model.
2020, 41(1): 228-234.
doi: 10.5768/JAO202041.0108002
Abstract:
Based on the principle of Raman scattering temperature measurement, a distributed optical fiber system was built to carry out the simulated leakage points location experiment of UPVC (unplasticized polyvinyl chloride) and cast-iron water pipes. Firstly, the original Anti-Stokes optical signal was filtered and smoothed. Then, the temperature measurement signal was classified and analyzed by using correlation coefficient method to identify whether the leakage occured in water pipes. Finally, the leakage points location in UVPC and cast-iron water pipes was identified by selective threshold method. The experimental results show that the distributed fiber optic system works stably and can accurately identify the leakage points of water pipes. The selective average threshold method can accurately locate the leakage points of 200 m underground water pipes and the location error ranges from 0.25 m to 0.65 m.
Based on the principle of Raman scattering temperature measurement, a distributed optical fiber system was built to carry out the simulated leakage points location experiment of UPVC (unplasticized polyvinyl chloride) and cast-iron water pipes. Firstly, the original Anti-Stokes optical signal was filtered and smoothed. Then, the temperature measurement signal was classified and analyzed by using correlation coefficient method to identify whether the leakage occured in water pipes. Finally, the leakage points location in UVPC and cast-iron water pipes was identified by selective threshold method. The experimental results show that the distributed fiber optic system works stably and can accurately identify the leakage points of water pipes. The selective average threshold method can accurately locate the leakage points of 200 m underground water pipes and the location error ranges from 0.25 m to 0.65 m.
New website launched!
陕公网安备 61011302001501号