2013 Vol. 34, No. 5
Display Method:
2013, 34(5): 723-727.
Abstract:
An infrared (IR) image air defense missile searching and tracking algorithm based on digital signal processor (DSP) was proposed. The algorithm was divided into two parts, the contour approximate curvature algorithm was used for missile tracking with tail-smoke; the small target tracking algorithm based on region growth was used for burning tracer tracking. The algorithm was implemented in C6416 DSP platform. Outfield experimental results indicate that the algorithm is capable of realizing the air defense missile optoelectronic tracking and guidance.
An infrared (IR) image air defense missile searching and tracking algorithm based on digital signal processor (DSP) was proposed. The algorithm was divided into two parts, the contour approximate curvature algorithm was used for missile tracking with tail-smoke; the small target tracking algorithm based on region growth was used for burning tracer tracking. The algorithm was implemented in C6416 DSP platform. Outfield experimental results indicate that the algorithm is capable of realizing the air defense missile optoelectronic tracking and guidance.
2013, 34(5): 728-732.
Abstract:
Aiming at the cooled 320256 detector with staring focal plane array (FPA), an optical system of middle infrared continuous zoom system was designed for airborne optoelectronic detection and tracking apparatus. The system was composed of a zoom lens system, a secondary imaging system and two reflectors. In order to avoid the fabrication and testing error of aspheric surface, the whole system was made of nine spherical glasses. The design results indicate that the system achieves the continuous zoom of 37.5 mm~150 mm, with working length of 3.7 m~4.8 m, and F number of 4, which can obtain the cold shield efficiency of 100%. At the Nyquist frequency of 16 lp/mm, the MTF values of all focal length positions and field of view are more than 0l55, which are closing to the diffraction limit. The overall lens length is 280 mm and the weight is only 110 g. The optical system has the advantages of small volume, high image quality and simple structure.
Aiming at the cooled 320256 detector with staring focal plane array (FPA), an optical system of middle infrared continuous zoom system was designed for airborne optoelectronic detection and tracking apparatus. The system was composed of a zoom lens system, a secondary imaging system and two reflectors. In order to avoid the fabrication and testing error of aspheric surface, the whole system was made of nine spherical glasses. The design results indicate that the system achieves the continuous zoom of 37.5 mm~150 mm, with working length of 3.7 m~4.8 m, and F number of 4, which can obtain the cold shield efficiency of 100%. At the Nyquist frequency of 16 lp/mm, the MTF values of all focal length positions and field of view are more than 0l55, which are closing to the diffraction limit. The overall lens length is 280 mm and the weight is only 110 g. The optical system has the advantages of small volume, high image quality and simple structure.
2013, 34(5): 733-737.
Abstract:
When the traditional up-trans photoelectric scout devices search and scout the surface targets in full field, the sky background radiation and the surface reflection yawp hold absolutely a majority of the image area, so the false alarm rate is rather high. At the same time, picking-up and scouting the surface target in complete image require exorbitant hardware and software of the video tracker. Aiming at the problems above, we proposed a control method of scout wave gate collocation strategy for fitting the up-rans photoelectric scout device. We built the math model of the virtual sea level reflected and projected by the space model of up-trans photoelectric scout device, calculated the points of intersection of the virtual sea-horizon and video image sidelines and sent to the video tracker in time, then collocated the narrowband scout wave gate along the virtual sea-horizon, the device only responded the surface target in the scouting wave gate as a result. The sea experiment result show that, the efficiency of object searching is inversely proportional to the proportion of narrowband scout in the image, and the error alarm rate reduces 80%.
When the traditional up-trans photoelectric scout devices search and scout the surface targets in full field, the sky background radiation and the surface reflection yawp hold absolutely a majority of the image area, so the false alarm rate is rather high. At the same time, picking-up and scouting the surface target in complete image require exorbitant hardware and software of the video tracker. Aiming at the problems above, we proposed a control method of scout wave gate collocation strategy for fitting the up-rans photoelectric scout device. We built the math model of the virtual sea level reflected and projected by the space model of up-trans photoelectric scout device, calculated the points of intersection of the virtual sea-horizon and video image sidelines and sent to the video tracker in time, then collocated the narrowband scout wave gate along the virtual sea-horizon, the device only responded the surface target in the scouting wave gate as a result. The sea experiment result show that, the efficiency of object searching is inversely proportional to the proportion of narrowband scout in the image, and the error alarm rate reduces 80%.
2013, 34(5): 738-741.
Abstract:
According to the optical compensation theory, a 31 infrared step zoom optical system was designed based on 320240 cooled midwave infrared (MWIR) focal plane array (FPA) detector. Its operation wavelength range was 3 m~5 m, F number was 4, cold stop efficiency was 100%, and focal length range was 430 mm~12.867 mm. The step zoom optical system was optimized by adopting three kinds of optical materials and aspheric surface. Aberrations of seven fields of view (FOVs) were strictly corrected, the MTF of each FOV was close to diffraction limit, and the image quality was good.
According to the optical compensation theory, a 31 infrared step zoom optical system was designed based on 320240 cooled midwave infrared (MWIR) focal plane array (FPA) detector. Its operation wavelength range was 3 m~5 m, F number was 4, cold stop efficiency was 100%, and focal length range was 430 mm~12.867 mm. The step zoom optical system was optimized by adopting three kinds of optical materials and aspheric surface. Aberrations of seven fields of view (FOVs) were strictly corrected, the MTF of each FOV was close to diffraction limit, and the image quality was good.
2013, 34(5): 742-747.
Abstract:
Ultraviolet (UV) LED array illumination was chosen for constructing uniform illumination ultraviolet light source required in the fingerprint fluorescence detection. The angular distribution of single LED radiation was tested by using photoelectric detector, and the angular distribution function of single LED was fitted. Eight LEDs were placed on the circle with radius of 10 mm evenly, and a LED was placed in the center shaft 5 mm above the circle. For a given observation screen under the uneven illumination error, according to the laws of the Spyro, the distance between the viewer and the circle array was determined and then the uniform illuminance distribution of LED circle array was realized. when the distance from observation screen to circle was given, the distance between the LED placed point on axis of the circle to the circle could be determined too. Experimental results show that the distance from the observation screen to the circle is 11.0 cm, the relative error of uneven illumination is less than 1.27% within a circle with radius of 10.0 mm.
Ultraviolet (UV) LED array illumination was chosen for constructing uniform illumination ultraviolet light source required in the fingerprint fluorescence detection. The angular distribution of single LED radiation was tested by using photoelectric detector, and the angular distribution function of single LED was fitted. Eight LEDs were placed on the circle with radius of 10 mm evenly, and a LED was placed in the center shaft 5 mm above the circle. For a given observation screen under the uneven illumination error, according to the laws of the Spyro, the distance between the viewer and the circle array was determined and then the uniform illuminance distribution of LED circle array was realized. when the distance from observation screen to circle was given, the distance between the LED placed point on axis of the circle to the circle could be determined too. Experimental results show that the distance from the observation screen to the circle is 11.0 cm, the relative error of uneven illumination is less than 1.27% within a circle with radius of 10.0 mm.
2013, 34(5): 748-753.
Abstract:
Modulation transfer function (MTF) is one of the important indexes to evaluate the image quality of infrared electro-optical imaging system. The testing principle of MTF test by tilting target (slit or knife-edge) was detailedly analyzed, and the contrast test about the two methods was carried out. An improved knife-edge method was proposed, multiple rows of data knife-edge were aligned and arranged in a row of data as a edge spread function, which could significantly increase the simpling points, improve simpling rate and testing resolution, and the knife-edge image was obtained. Line spread function (LSF) of each row was got by differentiation, the MTF of the new image composed of LSF multiple rows of data was calculated by canted slit method. Experimental results indicate that, the proposed method is simple and can effectively reduce the noise impact during MTF test process, the difference of the test result is not more than 7.5% compared with the canted silt method.
Modulation transfer function (MTF) is one of the important indexes to evaluate the image quality of infrared electro-optical imaging system. The testing principle of MTF test by tilting target (slit or knife-edge) was detailedly analyzed, and the contrast test about the two methods was carried out. An improved knife-edge method was proposed, multiple rows of data knife-edge were aligned and arranged in a row of data as a edge spread function, which could significantly increase the simpling points, improve simpling rate and testing resolution, and the knife-edge image was obtained. Line spread function (LSF) of each row was got by differentiation, the MTF of the new image composed of LSF multiple rows of data was calculated by canted slit method. Experimental results indicate that, the proposed method is simple and can effectively reduce the noise impact during MTF test process, the difference of the test result is not more than 7.5% compared with the canted silt method.
2013, 34(5): 754-757.
Abstract:
A varifocal method by using double-motor to control the zooming lens and the compensating lens instead of cam-curve muff was described, and a mechanical compensated zoom lens with focal length range of 342.76 mm~13.15 mm and relative aperture of 1/4 was also presented. The zooming lens was moved step-by-step and the compensating lens was moved following the zooming lens by position tracking method (CAMIMG), so that the lenses were positioned accurately in the whole zooming range. The pulse data corresponding to the positions relationship of zooming lens and compensating lens was stored in the CAM table according to optical design cam datum by digital serve control cam arithmetic, therefore the relationship of the two different movement velocity axes in each position was determined. The results show that the image is clear and steady in the whole varifocal range, the position resolution of zooming lens and compensating lens is 0.18 m, the optical axis consistency is about 1.9 in horizontal direction and 1.3 in vertical direction.
A varifocal method by using double-motor to control the zooming lens and the compensating lens instead of cam-curve muff was described, and a mechanical compensated zoom lens with focal length range of 342.76 mm~13.15 mm and relative aperture of 1/4 was also presented. The zooming lens was moved step-by-step and the compensating lens was moved following the zooming lens by position tracking method (CAMIMG), so that the lenses were positioned accurately in the whole zooming range. The pulse data corresponding to the positions relationship of zooming lens and compensating lens was stored in the CAM table according to optical design cam datum by digital serve control cam arithmetic, therefore the relationship of the two different movement velocity axes in each position was determined. The results show that the image is clear and steady in the whole varifocal range, the position resolution of zooming lens and compensating lens is 0.18 m, the optical axis consistency is about 1.9 in horizontal direction and 1.3 in vertical direction.
2013, 34(5): 758-763.
Abstract:
The relationship between the light attenuation effect and the thickness of foam/water curtain produced by foam-water dual-purpose interfere curtain of countermeasure washdown system was analyzed theoretically and studied through experiments. The curtain composed of dispersed droplets with thickness from 4 to 32 layers and different gas-liquid ratio was released to experimental deck and atmosphere, and the attenuation of parallel light was measured. Results show that attenuation of visible light by foam-water dual-purpose curtain increases significantly with thickness. The maximum of light attenuation reaches 95.8% when released to atmosphere and 98.4% when released to experimental deck. The transmission diminishes slower when the thickness increases. The research is valuable for improving interfere theory and equipment technique of foam curtain.
The relationship between the light attenuation effect and the thickness of foam/water curtain produced by foam-water dual-purpose interfere curtain of countermeasure washdown system was analyzed theoretically and studied through experiments. The curtain composed of dispersed droplets with thickness from 4 to 32 layers and different gas-liquid ratio was released to experimental deck and atmosphere, and the attenuation of parallel light was measured. Results show that attenuation of visible light by foam-water dual-purpose curtain increases significantly with thickness. The maximum of light attenuation reaches 95.8% when released to atmosphere and 98.4% when released to experimental deck. The transmission diminishes slower when the thickness increases. The research is valuable for improving interfere theory and equipment technique of foam curtain.
2013, 34(5): 764-767.
Abstract:
To study the influence of substrate temperature on the refractive index of alumina thin film and the film thickness, we accomplished Al2O3 thin films by ionassisted electron beam evaporation method under different substrate temperatures and the same Tooling factors. Based on the related theory of optical film, using the spectrophotometer to measure the spectral transmittance, we calculated the actual film thickness which were 275.611 nm~348.447 nm and the refractive index under 25℃~300℃. With the curves by numerical calculation and the experimental results by simulation, the effect of substrate temperature on refractive index of thin films and deposition efficiency was given.
To study the influence of substrate temperature on the refractive index of alumina thin film and the film thickness, we accomplished Al2O3 thin films by ionassisted electron beam evaporation method under different substrate temperatures and the same Tooling factors. Based on the related theory of optical film, using the spectrophotometer to measure the spectral transmittance, we calculated the actual film thickness which were 275.611 nm~348.447 nm and the refractive index under 25℃~300℃. With the curves by numerical calculation and the experimental results by simulation, the effect of substrate temperature on refractive index of thin films and deposition efficiency was given.
2013, 34(5): 768-771.
Abstract:
Base on the accurate expression of light intensity on the cross section and Rayleigh-Sommerfeld diffraction integrals, the characteristics of axial light intensity of cosh-Gaussian beams diffracted by a small aperture were analyzed and discussed by numerical calculation, and the influences of decentered parameter , the radius of small aperture a and the parameter w0/ on the axial light intensity of cosh-Gaussian beams were also discussed in detail. Results show that the axial light intensity of cosh-Gaussian beams depends not only on the decentered parameter , the radius of small aperture a and parameter w0/, but also on the transmission distance.
Base on the accurate expression of light intensity on the cross section and Rayleigh-Sommerfeld diffraction integrals, the characteristics of axial light intensity of cosh-Gaussian beams diffracted by a small aperture were analyzed and discussed by numerical calculation, and the influences of decentered parameter , the radius of small aperture a and the parameter w0/ on the axial light intensity of cosh-Gaussian beams were also discussed in detail. Results show that the axial light intensity of cosh-Gaussian beams depends not only on the decentered parameter , the radius of small aperture a and parameter w0/, but also on the transmission distance.
2013, 34(5): 772-777.
Abstract:
In order to solve the problems in current computer-generated hologram, such as huge computational cost, slow computational speed and disturbance from conjugate image, a new computer-generating holographic stereogram method was proposed. In this method, a series of perspective view images were captured by moving camera according to the principle of binocular stereo-vision in human visual system. The kinoform of different view angles was computed to synthesize holographic stereogram of three-dimensional object. The initial phase was set up and the feedback factor was added to iterate Fourier transform algorithm (IFTA) for improving the efficiency of iterative calculation. The holographic stereogram was loaded to liquid spatial light modulator. The experimental result shows the method we proposed can eliminate the noises from conjugate image and improve the iterative calculation efficiency over 30%, the structure similarity (SSIM) between reconstruction image and target image is above 0.85, which means the results have comfortable three-dimensional vision effect.
In order to solve the problems in current computer-generated hologram, such as huge computational cost, slow computational speed and disturbance from conjugate image, a new computer-generating holographic stereogram method was proposed. In this method, a series of perspective view images were captured by moving camera according to the principle of binocular stereo-vision in human visual system. The kinoform of different view angles was computed to synthesize holographic stereogram of three-dimensional object. The initial phase was set up and the feedback factor was added to iterate Fourier transform algorithm (IFTA) for improving the efficiency of iterative calculation. The holographic stereogram was loaded to liquid spatial light modulator. The experimental result shows the method we proposed can eliminate the noises from conjugate image and improve the iterative calculation efficiency over 30%, the structure similarity (SSIM) between reconstruction image and target image is above 0.85, which means the results have comfortable three-dimensional vision effect.
2013, 34(5): 778-783.
Abstract:
In frame-difference-based intelligent video monitoring systems, changes of image brightness can seriously influence the results of target object recognizing and tracking. After comparing some common brightness correction methods, the direct proportion transformation method was selected for further study. To meet the effect requirement, the transformation factor was calculated according to the average grayscale of a manually preselected region for avoiding influence of target object. To meet the real time requirement, the correction process was started only when the brightness difference between two continuous frames was greater than a certain threshold. When programming with Visual C++ 6.0, some further measures were adopted to speed up the process, such as transforming floating-point calculation into integer calculation. Experiment results show that, target objects that cannot be recognised before correction can be recognised after correction, the time consumption is about 1.2 ms for correction of one frame, and the comprehensive effect and speed reach the best when the threshold is 1 or 2.
In frame-difference-based intelligent video monitoring systems, changes of image brightness can seriously influence the results of target object recognizing and tracking. After comparing some common brightness correction methods, the direct proportion transformation method was selected for further study. To meet the effect requirement, the transformation factor was calculated according to the average grayscale of a manually preselected region for avoiding influence of target object. To meet the real time requirement, the correction process was started only when the brightness difference between two continuous frames was greater than a certain threshold. When programming with Visual C++ 6.0, some further measures were adopted to speed up the process, such as transforming floating-point calculation into integer calculation. Experiment results show that, target objects that cannot be recognised before correction can be recognised after correction, the time consumption is about 1.2 ms for correction of one frame, and the comprehensive effect and speed reach the best when the threshold is 1 or 2.
2013, 34(5): 784-790.
Abstract:
In order to achieve the long-distance real-time transmission of image fusion system, a high-speed network transmission image fusion system design scheme based on TI TMS320DM642 and TMS320DM365 processors was presented. Fusion module based on TMS320DM642 and net transmission module based on TMS320DM365 were designed, and the data communication problem between the image fusion module and net transmission module was solved. Image registration and real-time image fusion were achieved in fusion module. Image data was compressed by H.264 algorithm and carried out real-time transport protocol (RTP) packet processing according to RTP protocol in net transmission module, and the transmission rate was about 460 kb/s. Using thread pipe communication technology, a new establishment method based on RTP/real time streaming protocol (RTSP) embedded streaming media server was achieved. Experimental results show that the overall system image delay is about 300 ms.
In order to achieve the long-distance real-time transmission of image fusion system, a high-speed network transmission image fusion system design scheme based on TI TMS320DM642 and TMS320DM365 processors was presented. Fusion module based on TMS320DM642 and net transmission module based on TMS320DM365 were designed, and the data communication problem between the image fusion module and net transmission module was solved. Image registration and real-time image fusion were achieved in fusion module. Image data was compressed by H.264 algorithm and carried out real-time transport protocol (RTP) packet processing according to RTP protocol in net transmission module, and the transmission rate was about 460 kb/s. Using thread pipe communication technology, a new establishment method based on RTP/real time streaming protocol (RTSP) embedded streaming media server was achieved. Experimental results show that the overall system image delay is about 300 ms.
2013, 34(5): 791-795.
Abstract:
Traditional block matching algorithms have many deficiencies in image feature extracting and image matching, the complex edge extraction and the matching principle based on gray scale make the processing speed and accuracy decrease. To make the identification process more quickly and accurately, a recognition algorithm of the improved block matching was proposed. The image feature was enhanced by Laplace operator to improve the feature information. The results show that the average computing time of feature enhancement of this algorithm is about 0.005 4 s for 320240 pixels picture in Matlab.
Traditional block matching algorithms have many deficiencies in image feature extracting and image matching, the complex edge extraction and the matching principle based on gray scale make the processing speed and accuracy decrease. To make the identification process more quickly and accurately, a recognition algorithm of the improved block matching was proposed. The image feature was enhanced by Laplace operator to improve the feature information. The results show that the average computing time of feature enhancement of this algorithm is about 0.005 4 s for 320240 pixels picture in Matlab.
2013, 34(5): 796-801.
Abstract:
Two-step iterative shrinkage / Threshold (TwIST) of compressive sensing theory is introduced. Experimental works on reconstruction of single in-line hologram are developed. The goal is to eliminate the reconstructed conjugate information of single in-line hologram and overcome the shortage of digital holography on identifying the axial focal plane. Simulation analysis and experiment works are developed. First, a digital image and the standard resolution target are used as the samples for hologram, the quality of the reconstructed image based on TwIST and Fresnel approximation algorithm is estimated respectively. Then two bare fibers (diameter of 125 m) are selected as the test samples, identification capability of focal plane by TwIST is analyzed. The simulation and experimental results show that two-step iterative shrinkage algorithm can not only reconstruct the original information more than 68.73%, but also has a very good ability to identify the axial focal plane. These traits can help digital holography to detect some important parameters of functionally graded materials or the coating thickness of the optical functional coating devices.
Two-step iterative shrinkage / Threshold (TwIST) of compressive sensing theory is introduced. Experimental works on reconstruction of single in-line hologram are developed. The goal is to eliminate the reconstructed conjugate information of single in-line hologram and overcome the shortage of digital holography on identifying the axial focal plane. Simulation analysis and experiment works are developed. First, a digital image and the standard resolution target are used as the samples for hologram, the quality of the reconstructed image based on TwIST and Fresnel approximation algorithm is estimated respectively. Then two bare fibers (diameter of 125 m) are selected as the test samples, identification capability of focal plane by TwIST is analyzed. The simulation and experimental results show that two-step iterative shrinkage algorithm can not only reconstruct the original information more than 68.73%, but also has a very good ability to identify the axial focal plane. These traits can help digital holography to detect some important parameters of functionally graded materials or the coating thickness of the optical functional coating devices.
2013, 34(5): 802-808.
Abstract:
For the interferometric testing of optical elements with large aperture, due to large testing device and interferometric cavity length, the process of phaseshifting interferometry was affected by airflow disturbance and environmental vibration. In order to overcome the above factors, a single interferogram processing method based on two-dimensional fast Fourier transform(FFT) was proposed. This method only required one single spatial carrier fringe pattern to obtain the phase, which had the advantage of anti-vibration testing. The basic principle and the process of algorithm were analyzed, and the optical element with 600 mm large aperture in the near-infrared phase-shifting Fizeau interferometer was tested. Experimental results show that the peaktovalley (PV) value and root mean square (RMS) value of the wavefront obtained by FFT method are 0.112 and 0.014 respectively, compared with the results obtained by phase-shifting method, there are a difference less tham (1/500) in PV and almost zero difference in RMS.
For the interferometric testing of optical elements with large aperture, due to large testing device and interferometric cavity length, the process of phaseshifting interferometry was affected by airflow disturbance and environmental vibration. In order to overcome the above factors, a single interferogram processing method based on two-dimensional fast Fourier transform(FFT) was proposed. This method only required one single spatial carrier fringe pattern to obtain the phase, which had the advantage of anti-vibration testing. The basic principle and the process of algorithm were analyzed, and the optical element with 600 mm large aperture in the near-infrared phase-shifting Fizeau interferometer was tested. Experimental results show that the peaktovalley (PV) value and root mean square (RMS) value of the wavefront obtained by FFT method are 0.112 and 0.014 respectively, compared with the results obtained by phase-shifting method, there are a difference less tham (1/500) in PV and almost zero difference in RMS.
2013, 34(5): 809-814.
Abstract:
Due to the prominent difference of the imaging mechanism between visible and synthetic aperture radar (SAR) images, it is very difficult to extract common features and register, but in some cases, the edges of these two kinds of images have some certain correlation. According to the above problem, an image registration algorithm based on edge and speed-up robust feature (SURF) is proposed. Firstly, the similarity of these two images is enhanced through appropriately preprocessing, the common edge features are extracted by Canny operator which has good performance, and the SURFs are extracted from the edges of the images; then the features matching is done by the ratio purification method, the random sample consensus (RANSAC) algorithm is applied to remove the false matching points, and the affine transformation model is calculated to realize image automatic registration of SAR and visible image. Experimental results demonstrate that the correct matching probability of the proposed algorithm is 100% and the root mean square error is 0.852 pixel, moreover, the registration accuracy can achieve sub-pixel level,which proves the validity of the algorithm.
Due to the prominent difference of the imaging mechanism between visible and synthetic aperture radar (SAR) images, it is very difficult to extract common features and register, but in some cases, the edges of these two kinds of images have some certain correlation. According to the above problem, an image registration algorithm based on edge and speed-up robust feature (SURF) is proposed. Firstly, the similarity of these two images is enhanced through appropriately preprocessing, the common edge features are extracted by Canny operator which has good performance, and the SURFs are extracted from the edges of the images; then the features matching is done by the ratio purification method, the random sample consensus (RANSAC) algorithm is applied to remove the false matching points, and the affine transformation model is calculated to realize image automatic registration of SAR and visible image. Experimental results demonstrate that the correct matching probability of the proposed algorithm is 100% and the root mean square error is 0.852 pixel, moreover, the registration accuracy can achieve sub-pixel level,which proves the validity of the algorithm.
2013, 34(5): 815-819.
Abstract:
A bionic spherical 9 compound eyes visual system was designed through simulating visual mechanism of insect compound eye, and a panoramic image mosaic algorithm was proposed. Firstly, feature points were extracted and matched by using the scale invariant feature transform (SIFT) image-matching algorithm. Secondly, the transformation matrix H was computed by random sample consensus (RANSAC) algorithm and Levenberg-Marquardt (LM) algorithm. Finally, the panoramic image was completed with improved weighted smoothing algorithm. The experimental results show that the whole seamless panorama mosaic process of multi-view image sequence is completed automatically, and the field of view is enlarged in multi-directions, which makes up the defect of the traditional wide screen panorama.
A bionic spherical 9 compound eyes visual system was designed through simulating visual mechanism of insect compound eye, and a panoramic image mosaic algorithm was proposed. Firstly, feature points were extracted and matched by using the scale invariant feature transform (SIFT) image-matching algorithm. Secondly, the transformation matrix H was computed by random sample consensus (RANSAC) algorithm and Levenberg-Marquardt (LM) algorithm. Finally, the panoramic image was completed with improved weighted smoothing algorithm. The experimental results show that the whole seamless panorama mosaic process of multi-view image sequence is completed automatically, and the field of view is enlarged in multi-directions, which makes up the defect of the traditional wide screen panorama.
2013, 34(5): 820-824.
Abstract:
According to the problems of numerous noises and incomplete detection of moving object detection in complex natural environment, an improved solution method based on fuzzy set theory is put forward. First, pyramid-style multi-resolution model is used for background difference, and the preliminary foreground mask is obtained. Then the model of color, time, space and locality for the current image are represented by fuzzy sets to form a cluster of fuzzy vector set. At last, the features of these four vectors are fused together by fuzzy mathematic theory to get the membership of foreground for each pixel and detect the object. The method not only can detect relative complete moving object effectively, but overcome the influence of small moving background in natural environment at the same time. The experiments are carried out and the results show that the foreground recognition rate by the proposed method is 0.7174 and the error rate is 0.011 8. It can adapt to the influence of the dynamic background in natural environment.
According to the problems of numerous noises and incomplete detection of moving object detection in complex natural environment, an improved solution method based on fuzzy set theory is put forward. First, pyramid-style multi-resolution model is used for background difference, and the preliminary foreground mask is obtained. Then the model of color, time, space and locality for the current image are represented by fuzzy sets to form a cluster of fuzzy vector set. At last, the features of these four vectors are fused together by fuzzy mathematic theory to get the membership of foreground for each pixel and detect the object. The method not only can detect relative complete moving object effectively, but overcome the influence of small moving background in natural environment at the same time. The experiments are carried out and the results show that the foreground recognition rate by the proposed method is 0.7174 and the error rate is 0.011 8. It can adapt to the influence of the dynamic background in natural environment.
2013, 34(5): 825-830.
Abstract:
Due to the relative posture of each subaperture measurement is not known exactly, there is some ambiguity when the individual annular subapertures are combined into a full aperture map. To reduce this ambiguity resulted from mechanical errors, a global optimization algorithm separating mechanical errors was proposed. This algorithm established a mathematical model for separating mechanical errors based on wavefront aberration theory, and used ray tracing to eliminate the difference between ideal aspheric wavefront and reference spherical wavefront in every annular subaperture. Based on the global optimization algorithm which could abstain error transfer and accumulation, and each measured subaperture phase subtracted corresponding calculated difference firstly, and then these subaperture phases were stitched into a full aperture map by this algorithm containning mechanical error separation. A paraboloid with aperture of 75 mm and vertex radius of 100 mm was tested by annular subaperture stitching interferometry, the error of peak-valley was 0.05 , and the root-mean-square error was 0.003 . Results indicate that the global optimization algorithm can effectively separate and compensate mechanical errors.
Due to the relative posture of each subaperture measurement is not known exactly, there is some ambiguity when the individual annular subapertures are combined into a full aperture map. To reduce this ambiguity resulted from mechanical errors, a global optimization algorithm separating mechanical errors was proposed. This algorithm established a mathematical model for separating mechanical errors based on wavefront aberration theory, and used ray tracing to eliminate the difference between ideal aspheric wavefront and reference spherical wavefront in every annular subaperture. Based on the global optimization algorithm which could abstain error transfer and accumulation, and each measured subaperture phase subtracted corresponding calculated difference firstly, and then these subaperture phases were stitched into a full aperture map by this algorithm containning mechanical error separation. A paraboloid with aperture of 75 mm and vertex radius of 100 mm was tested by annular subaperture stitching interferometry, the error of peak-valley was 0.05 , and the root-mean-square error was 0.003 . Results indicate that the global optimization algorithm can effectively separate and compensate mechanical errors.
2013, 34(5): 831-836.
Abstract:
Aiming at the drawback of large time need for traditional phase unwrapping process, a novel method, absolute phase calculation for Fourier transform profilometry(FTP) with colour coding grating is proposed. Fourier is adopted to calculate the relative phase (wrapped phase) because of the least need of images. Color strips are used to mark each 2 phase-change period. With the two parts, relative phase and fringe orders, the absolute phase can be obtained directly. The fringe orders and the relative phase are calculated simultaneously, and this algorithm can circumvent phase unwrapping process. Moreover, errors can not diffuse, because no unwrapping process is needed in our algorithm, and also because the relative phase and fringe orders are uncorrelated. Besides, no need of unwrapping processing and calculating the absolute phase directly in our algorithm can lead to high speed in 3D profile measurement. Experimental results show the validity of our algorithm both for complex object and separate objects.
Aiming at the drawback of large time need for traditional phase unwrapping process, a novel method, absolute phase calculation for Fourier transform profilometry(FTP) with colour coding grating is proposed. Fourier is adopted to calculate the relative phase (wrapped phase) because of the least need of images. Color strips are used to mark each 2 phase-change period. With the two parts, relative phase and fringe orders, the absolute phase can be obtained directly. The fringe orders and the relative phase are calculated simultaneously, and this algorithm can circumvent phase unwrapping process. Moreover, errors can not diffuse, because no unwrapping process is needed in our algorithm, and also because the relative phase and fringe orders are uncorrelated. Besides, no need of unwrapping processing and calculating the absolute phase directly in our algorithm can lead to high speed in 3D profile measurement. Experimental results show the validity of our algorithm both for complex object and separate objects.
2013, 34(5): 837-840.
Abstract:
Aiming at the problems of laser micro energy measurement, a new method and device were advanced. This device used a set of stable power continuous wave laser, a pulse laser generator based on chopping, a special data acquisition system that included a pulse width measuring module and a pulse power measuring module, and a control software that could calculate the value of micro energy and give the correction coefficient of the calibrate energy meter by means of contrast method. The experiment results show that the laser pulse generated by this device has stable waveform, pulse power and pulse width. The pulse width measurement uncertainty is 0.08%, the laser power measurement uncertainty is 0.25%. It implements accurate replication and transfer of the laser energy of 0.1 pJ~1 mJ.
Aiming at the problems of laser micro energy measurement, a new method and device were advanced. This device used a set of stable power continuous wave laser, a pulse laser generator based on chopping, a special data acquisition system that included a pulse width measuring module and a pulse power measuring module, and a control software that could calculate the value of micro energy and give the correction coefficient of the calibrate energy meter by means of contrast method. The experiment results show that the laser pulse generated by this device has stable waveform, pulse power and pulse width. The pulse width measurement uncertainty is 0.08%, the laser power measurement uncertainty is 0.25%. It implements accurate replication and transfer of the laser energy of 0.1 pJ~1 mJ.
2013, 34(5): 841-844.
Abstract:
As it cannot integrally divide the defects under the complex background by the traditional measurement methods of the warhead scratch defect size, the method of adaptive threshold segmentation is used to solve this problem. Using this method for image segmentation, each pixel of the image corresponds to the different threshold, it can avoid the error division between defects and background, and the noise interference can also be avoid when the defects are completely divided. Defect segmentation process are given, the defect size is calculated, and the precision of the size measurement is analyzed. Experimental results shows that the defects can be completely divided by this methods when the gray level are different between the defects and the background, standard error is 0.122 8, and the uncertainty is 0.368 4.
As it cannot integrally divide the defects under the complex background by the traditional measurement methods of the warhead scratch defect size, the method of adaptive threshold segmentation is used to solve this problem. Using this method for image segmentation, each pixel of the image corresponds to the different threshold, it can avoid the error division between defects and background, and the noise interference can also be avoid when the defects are completely divided. Defect segmentation process are given, the defect size is calculated, and the precision of the size measurement is analyzed. Experimental results shows that the defects can be completely divided by this methods when the gray level are different between the defects and the background, standard error is 0.122 8, and the uncertainty is 0.368 4.
2013, 34(5): 845-848.
Abstract:
During the actual measurement, the accuracy of Fourier transform profilometry (FTP) is vulnerable to the hardware and calculation methods, the fringe image preprocessing technique is proposed based on the least square method. The filter design, the phase unwrapping algorithm and influence of system calibration model on the measurement results are discussed. The result shows that, this improved FTP can reduce the errors, which are caused by the sampling shortage, spectrum leakage, spectrum aliasing, phase wrapping and system calibration. The measurement error of the improved method can be controlled within 2%.
During the actual measurement, the accuracy of Fourier transform profilometry (FTP) is vulnerable to the hardware and calculation methods, the fringe image preprocessing technique is proposed based on the least square method. The filter design, the phase unwrapping algorithm and influence of system calibration model on the measurement results are discussed. The result shows that, this improved FTP can reduce the errors, which are caused by the sampling shortage, spectrum leakage, spectrum aliasing, phase wrapping and system calibration. The measurement error of the improved method can be controlled within 2%.
2013, 34(5): 849-853.
Abstract:
In order to improve the accuracy of measurement result of optical lenses with large depth of field in industrial visual inspection, image restoration must be carried out on the intermediate image in order to obtain a more accurate measurement result, and the point spread function (PSF) is the key to the image restoration. Non-diffracting imaging system was designed according to the formula of maximum non-diffracting distance of axicon, based on the scalar diffraction theory, PSF of the system was theoretically deduced by the stationary phase method within frequency range. Using the theory of quasi monochromatic light, the relationship between the PSF and the axicon angle or defocus was analyzed under incoherent light illumination. Simulations and experimental results show that the intensity of central spot of PSF is stronger and the density of diffraction fringes is larger when the refracting angle increases, while the influence of defocus on PSF has an opposite effect.
In order to improve the accuracy of measurement result of optical lenses with large depth of field in industrial visual inspection, image restoration must be carried out on the intermediate image in order to obtain a more accurate measurement result, and the point spread function (PSF) is the key to the image restoration. Non-diffracting imaging system was designed according to the formula of maximum non-diffracting distance of axicon, based on the scalar diffraction theory, PSF of the system was theoretically deduced by the stationary phase method within frequency range. Using the theory of quasi monochromatic light, the relationship between the PSF and the axicon angle or defocus was analyzed under incoherent light illumination. Simulations and experimental results show that the intensity of central spot of PSF is stronger and the density of diffraction fringes is larger when the refracting angle increases, while the influence of defocus on PSF has an opposite effect.
2013, 34(5): 854-859.
Abstract:
Experiments on polishing and surface-modification of an aspheric surface silicon carbide (SiC) mirror with 520 mm diameter and some SiC plane mirrors with different diameters were carried out, the technology of surface-modification was analyzed and discussed in detail. The effect of surface-modification (ion beam assisted deposition) on optical properties of SiC mirrors was researched. Several types of association scheme were constituted by combining the physical and chemical behavior of physical vapor deposition (PVD) coatings with the mirror polishing technique, the PVD coatings- requirements for the surface properties before surface-modification and the optimum combining site between them were mainly studied. Results show that the polish of modified surface with suitable plan can both improve the surface error and roughness. The surface roughness of 520 mm aspheric surface SiC mirror is 0.756 nm (Sq). Compared with the surface before modified, the roughness is improved.
Experiments on polishing and surface-modification of an aspheric surface silicon carbide (SiC) mirror with 520 mm diameter and some SiC plane mirrors with different diameters were carried out, the technology of surface-modification was analyzed and discussed in detail. The effect of surface-modification (ion beam assisted deposition) on optical properties of SiC mirrors was researched. Several types of association scheme were constituted by combining the physical and chemical behavior of physical vapor deposition (PVD) coatings with the mirror polishing technique, the PVD coatings- requirements for the surface properties before surface-modification and the optimum combining site between them were mainly studied. Results show that the polish of modified surface with suitable plan can both improve the surface error and roughness. The surface roughness of 520 mm aspheric surface SiC mirror is 0.756 nm (Sq). Compared with the surface before modified, the roughness is improved.
2013, 34(5): 860-863.
Abstract:
A method of infrared camouflage target recognition based on polarization-gated imaging was proposed, in order to prevent the infrared camouflage from disturbing the traditional thermal imaging system and recognize the IR camouflage targets on the battlefield. The acquisition of infrared polarization images was implemented through 4 fixed polarizers gated by scanner. Experiment result shows that the method can successfully acquire the IR polarization image at the rate of 4 f/s, meaning this method is real-time to a certain extent. Comparing with the original thermal intensity image, the polarization image has an increase of average gray-scale value up to 14%, gray-scale standard deviation up to 42% and average gradient up to 98%, meaning an ideal camouflage target recognition effect.
A method of infrared camouflage target recognition based on polarization-gated imaging was proposed, in order to prevent the infrared camouflage from disturbing the traditional thermal imaging system and recognize the IR camouflage targets on the battlefield. The acquisition of infrared polarization images was implemented through 4 fixed polarizers gated by scanner. Experiment result shows that the method can successfully acquire the IR polarization image at the rate of 4 f/s, meaning this method is real-time to a certain extent. Comparing with the original thermal intensity image, the polarization image has an increase of average gray-scale value up to 14%, gray-scale standard deviation up to 42% and average gradient up to 98%, meaning an ideal camouflage target recognition effect.
2013, 34(5): 864-869.
Abstract:
In order to meet the needs of high power balanced laser beam, taking laser diode bar and stack as models, parallel glass plates were used in beam segmentation and rearrangement based on their beam displacement property. The system was simulated in Zemax. The initial divergence angels in diode laser bar were 40and 10. Before shaping, the beam parameter products of the collimated beam in fast axis and slow axis were 0.455 2 mmmrad and 20.484 mmmrad respectively, which had a great difference in beam quality. After shaping system, the beam parameter products in fast axis and slow axis were 2.731 2 mmmrad and 3.414 0 mmmrad respectively, with a better balanced performance. Similarly, parallel glass plates were used to eliminate the emitting dead zone of laser diode (LD) stack , increasing the beam energy density. After shaping system, the beam parameter products in fast axis and slow axis were 7.002 mmmrad and 10.242 mmmrad respectively, the coupling efficiency was 90.13%.
In order to meet the needs of high power balanced laser beam, taking laser diode bar and stack as models, parallel glass plates were used in beam segmentation and rearrangement based on their beam displacement property. The system was simulated in Zemax. The initial divergence angels in diode laser bar were 40and 10. Before shaping, the beam parameter products of the collimated beam in fast axis and slow axis were 0.455 2 mmmrad and 20.484 mmmrad respectively, which had a great difference in beam quality. After shaping system, the beam parameter products in fast axis and slow axis were 2.731 2 mmmrad and 3.414 0 mmmrad respectively, with a better balanced performance. Similarly, parallel glass plates were used to eliminate the emitting dead zone of laser diode (LD) stack , increasing the beam energy density. After shaping system, the beam parameter products in fast axis and slow axis were 7.002 mmmrad and 10.242 mmmrad respectively, the coupling efficiency was 90.13%.
2013, 34(5): 870-873.
Abstract:
We reported a continuously and passively mode-locked Nd∶YVO4 crystal picosecond oscillator based on 888 nm semiconductor pumped. We used the way of thermal excitation to improve the thermal performance of the laser, as well as increase the average output power and the single pulse energy of the modelocked oscillator. We achieved the starting and maintaining of the laser mode-locking by semiconductor saturable absorber mirror(SESAM). Under the absorbed pump power of 60 W, the maximum output power of 15 W at 53 MHz repetition rate was obtained, and the optical-to-optical efficiency of 24% was achieved. The pulse width was measured at 45 ps when the output power was 15 W.
We reported a continuously and passively mode-locked Nd∶YVO4 crystal picosecond oscillator based on 888 nm semiconductor pumped. We used the way of thermal excitation to improve the thermal performance of the laser, as well as increase the average output power and the single pulse energy of the modelocked oscillator. We achieved the starting and maintaining of the laser mode-locking by semiconductor saturable absorber mirror(SESAM). Under the absorbed pump power of 60 W, the maximum output power of 15 W at 53 MHz repetition rate was obtained, and the optical-to-optical efficiency of 24% was achieved. The pulse width was measured at 45 ps when the output power was 15 W.
2013, 34(5): 874-877.
Abstract:
By using the side-pumped Nd∶GdVO4 crystal and type-I critically phase-matched LiB3O5 (LBO) crystal in cavity, the red light at 671nm was obtained. The cavity type of plane-concave was used to avoid decreasing the efficiency of doubling frequency caused by astigmatism. The parameters such as the resonant cavity curvature radius, length and laying position of crystal were optimized by software. The 671 nm red light of 9.7 W was obtained with injected electric power of 750 W. The repetition rate was 10 kHz, the angle of divergence was 7.9 mrad, and the electric-optical conversion efficiency was about 1.3%.
By using the side-pumped Nd∶GdVO4 crystal and type-I critically phase-matched LiB3O5 (LBO) crystal in cavity, the red light at 671nm was obtained. The cavity type of plane-concave was used to avoid decreasing the efficiency of doubling frequency caused by astigmatism. The parameters such as the resonant cavity curvature radius, length and laying position of crystal were optimized by software. The 671 nm red light of 9.7 W was obtained with injected electric power of 750 W. The repetition rate was 10 kHz, the angle of divergence was 7.9 mrad, and the electric-optical conversion efficiency was about 1.3%.
2013, 34(5): 878-881.
Abstract:
To acquire high efficiency multi-band laser,we used the acousto-optic Q-switched technology driven by high repetition rate and the laser-diode(LD) sidepumped technology to obtain the 1.06 m laser output with high repetition rate ,high power and narrow width. Then we got two 1.06 m linearly polarized light beams in vertical and horizontal directions by external cavity polarimeter. The vertical beam was split into one 1.46 m laser beam and one 3.9 m mid-IR beam by pumping periodically poled lithium tantalate (PPLT), and another 1.06 m horizontal laser beam was mixed with the 1.46 m and 3.9 m beams to realize triple-waveband coaxial laser output finally. As the input current was 35A and the Q-switched frequency was 10 kHz, we could get 1.06 m laser with 140 W output. After laser splitting and pumping PPLT, we could get 3.9 m and 1.46 m laser output with top power of 6.3 W and 8.6 W, respectively, and the difference frequency conversion efficiency was 21.3%. The experiment results demonstrate that the high-repetition narrow-width 1.06 m laser can be acquired by using high-repetition-rate electro-optical Q-switched technology and LD-side-pumped technology, and pumping PPLT can gain 3.9 m and 1.46 m laser output.
To acquire high efficiency multi-band laser,we used the acousto-optic Q-switched technology driven by high repetition rate and the laser-diode(LD) sidepumped technology to obtain the 1.06 m laser output with high repetition rate ,high power and narrow width. Then we got two 1.06 m linearly polarized light beams in vertical and horizontal directions by external cavity polarimeter. The vertical beam was split into one 1.46 m laser beam and one 3.9 m mid-IR beam by pumping periodically poled lithium tantalate (PPLT), and another 1.06 m horizontal laser beam was mixed with the 1.46 m and 3.9 m beams to realize triple-waveband coaxial laser output finally. As the input current was 35A and the Q-switched frequency was 10 kHz, we could get 1.06 m laser with 140 W output. After laser splitting and pumping PPLT, we could get 3.9 m and 1.46 m laser output with top power of 6.3 W and 8.6 W, respectively, and the difference frequency conversion efficiency was 21.3%. The experiment results demonstrate that the high-repetition narrow-width 1.06 m laser can be acquired by using high-repetition-rate electro-optical Q-switched technology and LD-side-pumped technology, and pumping PPLT can gain 3.9 m and 1.46 m laser output.
2013, 34(5): 882-886.
Abstract:
A novel all-optical wavelength conversion scheme based on stimulated Raman scattering (SRS) in optical fiber and the corresponding model were theoretically proposed and experimentally demonstrated. The amplified signal light and continuous probe light were injected into fiber simultaneously and the wavelength conversion were achieved between the two lights. The results show that SRS can be used to achieve wavelength conversion and realize the conversion between the wavelength across several THz. The maximum conversion efficiency and optimum extinction ratio are -17.3 dB and 15.7 dB respectivly. By changing the wavelength of probe light,the all-optical wavelength conversion and tunable wavelength conversion seperated several THz can be realized.
A novel all-optical wavelength conversion scheme based on stimulated Raman scattering (SRS) in optical fiber and the corresponding model were theoretically proposed and experimentally demonstrated. The amplified signal light and continuous probe light were injected into fiber simultaneously and the wavelength conversion were achieved between the two lights. The results show that SRS can be used to achieve wavelength conversion and realize the conversion between the wavelength across several THz. The maximum conversion efficiency and optimum extinction ratio are -17.3 dB and 15.7 dB respectivly. By changing the wavelength of probe light,the all-optical wavelength conversion and tunable wavelength conversion seperated several THz can be realized.