2016 Vol. 37, No. 2
Aiming at the problem that the extraction accuracy of laser stripe center can be easily disturbed by a large amount of noise in line structured light 3D shape measurement, a method for extracting the stripe subpixel center was proposed. First, the noise in the stripe image was analyzed and preprocessed by the average method and median filtering. Then, the stripe was obtained by iterative threshold segmentation and morphological method, and the pixel center was extracted by distance transform. Finally, on the basis of the pixel center, binary information and light intensity, the precise subpixel center of the stripe was extracted by combining curve fitting with barycenter method. After simulation analysis and experimental verification, the maximum value and average value of column coordinate deviation between adjacent rows of stripe center are about 2 pixels and 0.3 pixels, respectively, smaller than the traditional method. Experimental results show that, the algorithm based on the effective use of the stripe gray distribution, can effectively reduce the influence of noise, approach to the true center of stripe more cosely, with extremely strong noise proof ability.
In order to solve the problems of low energy utilization and low efficiency in femtosecond laser micromachining, a new method of femtosecond parallel processing based on spatial light modulator (SLM) was proposed. Based on the discussion of the relationship between incident angle and diffraction angle, the principle of superposition of multiple diffraction gratings was analyzed. Multiple beams were generated by loading the phase holograms calculated using gerchberg and saxton(GS) algorithm on the SLM. The parallel femtosecond laser beam with the pulse width of 100 fs and the wavelength of 800 nm was focused inside a polymethyl methacrylate (PMMA) sample, uniform multipoint patterns of 一 and 工 were observed. Experimental results show that it is an effective method for controlling multibeam parallel processing.
Linestructured light vision measurement system is an important part of laser scanning probe, and its structural parameters affect the measurement accuracy and reliability of the probe directly. In this study, the linestructured light vision system used in the field of laser scanning probe was researched. The relations between the structural parameters and overall accuracy were analyzed, and an optimization design method was proposed based on the simulation system, and then verified experimentally. The experiment results show that the relative error of two spatial points distance is 0.019 8 mm in the optimized system, and the approach in this study can meet demands of accuracy measurement in the field of majority detection applications.
Aiming at the deficiency of the existing laser absorption rate measurement，a laser absorbing rate measuring device based on integrating sphere method was designed on the basis of the existing laser absorption rate measurement device.The light tube structure was added, the light source was monitored, and the light beam was modulated. The method of correlation detection and synchronous acquisition was adopted, which could effectively remove the influence on the measurement of background noise, detector noise and detection circuit noise. Furthermore, the influence of the light tube on the sample absorption rate was compared in experiment, and the experiments on multiple samples were carried out. The results show that the measurement error of the device can reach ±2% or less for the sample reflectance in the range of 10×10-6~10 000×10-6ppm,which indicates that highaccuracy measurement of laser absorption rate is achieved.
The in situ infrared thermometry methods of metal organic chemical vapor deposition (MOCVD) mainly include monochromatic radiation thermometry and doublewavelength colorimetric thermometry. We calculated the 940 nm and 1 550 nm spectral emissivity of 10 m GaN on silicon (111) substrate by film thickness interference model and Kirchhoffs law,compared the relative error and relative sensitivity among 940 nm monochromatic radiation thermometry, 1 550 nm monochromatic radiation thermometry, 940 nm /1 550 nm doublewavelength colorimetric thermometry within the scope of 500 ℃ to 1 300 ℃ by taking the Thomas Swan CSS MOCVD as an example,as well as the calibration betweenmonochromatic radiation thermometry and doublewavelength colorimetric thermometry.Moreover,by using the 940 nm /1 550 nm doublewavelength colorimetric thermometry,the temperature of LED epitaxial wafer with InGaN/GaN MQW structure growing on silicon(111) substrate was monitored online.The result shows that the 940 nm/1 550 nm doublewavelength colorimetric thermometry offers advantages over the 940 nm and 1 550 nm monochromatic radiation thermometry in terms of relative error and correction. The conclusion can provide reference for the in situ infrared thermometry design of MOCVD.
To solve the detection distance of infrared thermal imager, using MODTRAN program, the atmospheric spectral transmittance in different cloud and rain conditions was calculated. For detector anglefixed and heightfixed operating, using the spectrumpath bisection method and spectrumangle bisection method respectively, the thermal imaging system detection distances of point source target were evaluated. Under cirrus weather conditions, the effects of the detection probability, rainfall rate, cloud thickness and cloud base height on the atmospheric spectral transmittance and the detection distance were analyzed. The results show that the selection of detection probability of infrared thermal imaging system has a great influence on the detection distance. In the case of tank target, when the detection probability increases from 10% to 90%, the detection distance reduces by 21.7%. The rainfall rate also has a great influence on atmospheric transmittance and the detection distance. When the rainfall rate increases from 0 to 4.0 mm/h, the detection distance decreases by 76.8%. The effects of the cloud thickness and cloud base height are associated with the geometric position relationship between the detector and the cloud. The upper bound of the detection distance of human target is 5.60 km. The detection distance is also influenced by the detector operating model. The biggest difference is about 3.8%.
An optical fiber displacementvibration sensor using the incoherent light as a light source and the mirror as a reflective surface was analyzed. The analytic expressions of the optical fiber output power and the measurement distance were derived.The relationship between power ratio and measurement distance was detailedly investigated based on different optical fiber radii, different numerical apertures, different spacings and included angles between transmitting and receiving optical fibers.The results of numerical simulation show that the measurement precision and scope are affected by the optical fiber radius,the optical fiber numerical aperture, the distance and axis angle of both transmitting and receiving optical fibers. By analyzing the variation curves of distance and output power ratio, it is helpful to select the appropriate sensor according to the measurement target in practical engineering applications. Moreover, the measurement error was discussed. The results shows that the measuring range of distance is increased with the increasing of optic fiber included angle, while the error rate is reduced.The error rate is 1% for the included angle of 0.25.
In order to meet the demands of the calibration for ground equipment, a new type of highprecision star magnitude simulator with limited space and full spectrum was presented. The optical system structure of the star magnitude simulator was designed to be the compact structure space. The control system based on the pulse width modulation (PWM) technology can realize the digital adjustment of the LED light intensity from 640 lx to 1.0 lx. The system which was drived by PT4115 used the white light XLamp LED as the light source. Compared with the traditional analog adjustment star magnitude simulator, the new star magnitude simulator has the characteristics of small volume, low heat and high stability, etc. Through the application in dark room environment, the analysis and test results show that the 0m~7m magnitude stars can be simulated, the spectral range can cover the whole visible light, and the control precision of the star magnitude is up to 0.1m.
In order to obtain the infrared/visible/lowlevel light images of targets, a multichannel photoelectric system optical axis parallelism calibration device was proposed. This device could easily load multichannel photoelectric detection system, it realized the adjustment for the optical axis parallism in the azimuth and pitch directions mainly through the waistshaped hole of the backplane clearance and the pad under repair backplane, to meet the optical axis parallelism not more than 15. The instrument is characterized by simply adjusting the optical system of optical axis parallelism for the first time, it is possible to achieve different detectors repeating interchangeable lenses,and the repeatable positioning accuracy is higher than 0.01 mm. In addtional,the instrument can be easily loaded on other light equipments, such as photoelectric rotary table,convenient for field experiments, which can satify the requirement for using.
To realize the correction of the optical axis and satisfy the requirement for miniaturization and lightweight, a laser/visual sighting telescope system was designed, which had the common path with laser receiving and visual collimator. By means of shared telephoto objective and coating beamsplitting film on inclined plane of the block prism, the system was applied to opticelectronic observatory, with visual correction of optic axis and laser receiving. The magnification of the visual sighting telescope is 7×, the field of view (FOV) is 5°, the exit pupil diameter is 3.7 mm, and the exit pupil distance is not less than 20mm. The laser receiving system has the FOV of 1mrad and the entrance pupil diameter of 50 mm. The designed results meet the index requirments.
The temperature change can affect the change of the shape and interval of the surfaces of the ultraviolet(UV) lens. In order to study the effect of the temperature on the imaging quality of the UV lens, the method of thermal/structural/optical(TSO) integrated analysis was used. The basic processes and key technologies of the TSO integrated analysis were discussed and the Zernike polynomials was used as the interface between structural analysis and optical analysis. Based on the design of the optical system and mechanical structure of the UV lens, the thermal-structural model of the lens was established and the changes of the surface shape and lens spacing were obtained. Finally, the results were coupled in the optical design software, then the image quality was further analyzed. The analysis results show that, in the working range of the lens, the design value of the modulation transfer function (MTF) at 12lp/mm is about 0.7, which can meet the use requirements of hyperspectral imaging spectrometer. At the same time, it also provides a reference for the analysis of the final machine of the spectrometer.
The thermal deformation of silicone on glass (SOG) Fresnel lens made up of silicone and glass can affect the performance of concentrating photovoltaic (CPV) system. In order to investigate the influence, we designed and optimized two kinds of CPV concentrator systems. Then, we calculated the thermal deformation of Fresnel lens with finite element analysis method, as well as the concentration efficiency of the system through Monte Carlo method. The calculation indicates for the single stage concentrator system, the Fresnel lens designed with monochromatic light uniformity method is obviously affected by the warpage deformation of concentrator and free deformation of prisms. The difference of optical efficiency is about 10% between the deformed and the undeformed concentrator system. However, the Fresnel lens designed with improved method that compensates for dispersion is little influenced by thermal deformation. Moreover, the twostage concentrator system can be greatly resisted to the thermal deformation.
For the solution of the robotic handeye calibration, on the basis of existing research methods, a new method which combines the advantage of the highprecision of traditional calibration method with the active visual method was proposed.This method lets the robot perform orthogonal translational motion only twice and photograph the calibration plate in each position separately, so it can solve the handeye rotation matrix R, then it guides the robot move to the origin of the calibration plate, and can solve the handeye translation matrix T. Using the Pingfang 6 degree of freedom( DOF) serial robot as an experiment example, we applied the proposed method to calculate the handeye relation matrix, then used the traditional equation to analyze the accuracy of the calculation result, finally proved that the error of this method was within 1 mm. This method can achieve high accuracy and easy implement without the consideration of solving complex traditional equation.
External screw parameters such as top diameter, effective diameter, bottom diameter and lead of the downhole explosive device body are needed to measure during producing. In order to improve the measurement automation, the external screw machine vision measuring method was studied on the basis of the device body structural features. A doubletelecetric optical measuring system was designed for the purpose of improving the measurement accuracy. The image region extraction method, quick feature points extraction method, and others parameters measuring methods were proposed. In addition, experiments were performed on the built platform. According to the results, the thread lead is 3 980.8 m, the large diameter is 65 435.6 m, the small diameter is 60 669.4 m and the toothform height is 2 383.1 m, meeting the demand of 4class threaded precision, which can prove the correctness and feasibility of the method.
Aiming at the problems encountered in the actual inspection work and the requirements proposed in the GB/T 108992009 for test method of stray light in an optical system, we discussed the necessity for testing the stray light coefficient in the full field of an optical system and the corresponding methods of measurement. Through the transformation of the existing stray light testing device, we achieved the testing of stray light coefficient in the full field of an optical system, and gave the test results and distribution curves of stray light coefficient in the full field of 2 kinds of zoom lenses. It is concluded that there is a great difference in stray light coefficient and distribution between the lenses which have similar parameters but different models,and the different form of optical and mechanical structure is the main reason for this difference.The measurement data can offer a reference for improving the suppression ability of stray light in optic systems.
To study the influence of modulation error of electrooptic material on the quality of the interference fringes, an imaging model for interference fringes in electrooptic crystal was established. The relationship among the refractive index, the flatness and the wavefront modulation was analyzed, as well as the fringes imaging characteristics under the nonideal conditions. The experimental results show that the distortion of the refractive index of the crystal can causes the distortion of the interference fringes, the high harmonic components of the projection fringes are directly caused by the discrete order of refractive index, even severe distortion of projected fringes; In the case of the refractive index of linear growth, the crystal plane in a range of 0.1 μm can also cause interference fringes distortion.
Traditional Hough transform has large amount of calculation, is poor realtime, and easily influenced by image noise, so the road detection using it cannot achieve good effect. Consequently, an improved algorithm for road detection based on Hough transform was presented. Based on the analysis of the actual road shape modeling, the algorithm can effectively remove the error message of actual road with / without signs or any other interfering factors, and can achieve better effect of road detection within 200 ms.
In order to improve the accuracy of multifocus image fusion, combining with good localization and shift invariance of finite discrete shearlet transform(FDST), a new image fusion algorithm based on FDST and improved contrast was proposed. Firstly, the registration multifocus images were decomposed by FDST, and the low frequency subband coefficients and high frequency subband coefficients of different scales and directions were obtained. The fusion principle of low frequency subband coefficients was based on the method of regional average energy matching degree. As for high frequency subband coefficients,the sum of contrast was adopted as the fusion rule, which combined the lowfrequency coefficients with the high frequency coefficients. Finally，low frequency information and high frequency information were reconstructed to image by finite discrete shearlet inverse transform, and both subjective visual evaluation and objective performance assessments of the fusion results were implemented. Simulation results indicate that the proposed algorithm is obviously superior on subjective visual effects. The values of entropy, mutual information quantity and edge similarity increase averagely by 1.4%,3.4% and 0.8%, respectively, compared with other fusion algorithms. It is superior to other fusion algorithms on objective evaluation.
Singleframe phase demodulation algorithms for fringe pattern are widely employed in realtime and dynamic measurement techniques. Subfringe integration(SI) algorithm has an advantage of strong antinoise ability. However, the period estimation of the deformed pattern using the carrier frequency would result in frequency mismatch. Against the inherent defect, an improved algorithm was proposed. The proposed algorithm estimated the local frequency, and then computed the intensity integration for phase extraction. Results show that the phase error of the new algorithm is one third of the original algorithm in simulations. The finge projection profilometry experiment was also conducted to demonstrate the validity of the proposed algorithm. This papar provides a candidate for singleframe fringe phase analysis technique.
Aiming at the matching and intersection of multiblasting point data tested with video theodolite in conventional ordnance test, it offers an entire set of analyzing data techniques. Firstly we made the allcombination intersection calculation with the angle data from each testing station to obtain the collection of coordinate data of many possible blasting points. Then we set the distance threshold value according to the near neighbor rule to get rid of the majority of fake objections by clustering method, and retained the possible blasting point collections closed to the target. Thirdly we processed the preliminary blasting points collections with matrix analysis based on the evidence theory. Eventually we got the collection of the coordinate data of blasting points. Simulation and experiment results show that the method can effectively process the angle data of no more than 10 ground exploding point targets with the same frame, and get the coordinates.
There are many quality problems in depth image，for example, images may be combined with blurring, empty holes and noise. In order to solve these problems, from the view point of software,an algorithm guided by color images based on structural characteristics was studied without changing any physical structure of the sensor imaging system,which could realize depth image enhancement and empty holes repairing,as well as improve the quality of the depth image. Through the extraction for the structure of the color image and depth image, the common global features were obtained. After joint bilateral filtering for the structural features,the depth image was enhanced based on Markov random filed (MRF), and the depth enhanced image was finally obtained with low cost. Experimental results show that the algorithm has a better effect in keeping the detail, smoothness and integrity of image edge; moreover, the rootmeansquare error (RMSE) of this algorithm is smaller than other algorithms,which is 0.50693 and 1.16930 for image Teddy and Art,respectively.
Liquid lens is a new optical element, which implements zooming without mechanical moving. A new way to design zoom lens system was proposed.A rare kind of liquid lens with large aperture was used to design a zoom system. The principle of the liquid lens was introduced first, then the possibility of designing the zoom liquid lens system was analyzed and various design plans were provided, and finally the Zemax software was utilized to simulate zoom system. A frontal lens group scheme was given aiming at the large field of view(FOV),and a zoom system was achieved with the FOV of 60 and the zoom ratio of 6.2.
In order to obtain distinct images for the different distances of the Cassegrain system, the secondary reflector must be adjusted slightly by revolving with a screw thread so as to change the focus of the Cassegrain system. However，the revolving of the secondary reflector can lead to the cyclical shaking image, and this kind of shaking is forbidden in many highprecision optical instruments. Aiming at this problem,the factor causing the periodic shaking of the image in the system was analyzed, and the specific mathematical formula was given distinctly.Additional, through a particular instance based on the precision of manufactory, the maximum deviation under this precision was analyzed,and the fine adjustment mechanism of secondary mirror without revolving the screw thread should be utilized if the shaking radius is beyond the range of allowable error. Results show that the fine adjustment mechanism of secondary mirror with revolving lead can lead to 0.65 mm radius shaking,while the other method without revolving lead can lead to 0.13 mm, implying the precision of the system can be greatly improved.
To observe the light spectrum of the gas puff imaging (GPI) on experimental advanced superconducting tokamak (EAST), a GPI diagnostic optical window was designed. Based on the initial structure of the Petzwal lens,the optimization of the system was conducted. The aberration of the designed system was controlled within 1/4 ,and the modulation transfer function (MTF) was 0.76 at 16 lp/mm with the wavelength of 656.2/587.6nm. Then the multiconfiguration at the temperature of 20~250 was analyzed，the lens material which had the similar coefficient of thermal expansion was selected, and the MTF at 16 lp/mm was almost invariable in T direction, but it was changed from 0.82 to 0.69 in S direction, which was still in the design range. In addition, according to the physical environment of the optical window, the statics and the thermodynamic simulation was done to the window, the deformation of the window was 1/5 , the tolerance analysis based on the value of the deformation showed that the optical systems MTF was still bigger than 0.5. Finally, the spectral images captured by rapid camera showed that the distribution and the change of the intensity could be clearly observed by the optical system.
Resolution, imaging time and distance are crucial parameters for terahertz imaging technology. In order to improve the terahertz imaging capability for practical use, especially for quick identification and alarm for hidden objects in a long distance, a passive rapid terahertz scanning imaging system for long distance was presented. In terms of core scanning system,a hexahedral mirror for twodimensional fast scanning through high speed rotating and reciprocating swing combination was employed. The system was placed in the parallel light path of the offaxis threemirror system. The scan area focused on image space instead of the object space, which could effectively narrow the scan area, so as to shorten the imaging time. The main parameters of the passive scanning imaging system are as follows: the detector frequency is 0.34 THz, the imaging distance is 10 m, the field of view is 1.51.5, the resolution is 3 cm, and the imaging time is 1 s.
To resolve the problem of lineofsight (LOS) drift caused by the earth’s rotation in gyrostabilized platform, a LOS drift compensation method based on inertial attitude parameters was proposed. According to the geographic coordinate and the platform’s attitude information computed by the inertial system, the method calculated the earth rotation angular rate’s projection on the optoelectronic stabilized platform coordinate system. Using this projection, the sightstabilizing control unit compensated the speed feedback signals, thereby the LOS drift caused by earth’s rotation was reduced. Test results show that using the compensation method, the azimuth drift amount is 1.69 mrad/h and the pitch drift amount is 1.84 mrad/h, far less than the drift amount of 6.89 mrad/h in azimuth orientation and 7.32 mrad/h in pitch orientation without using compensation method,which proves that this method can effectively control the LOS drift.
In order to satisfy the need of digital ultra highspeed imaging and improve the time resolution of the camera, an optical system of 8frame gated camera applied to 350 nm~800 nm waveband was designed. The structure of the system was based on splitting the focused imaging light beam with the aperture located externally. It could capture 8 of the same object images. The influence of some significant optical parameters, such as image distance and image side numerical aperture on illumination difference between multiframe framing system images, the correction of chromatic aberration in wide spectral imaging system and other problems were analyzed. Moreover,the relay lens was designed and optimized with Code V. The effective image size is Φ26mm,the modulation transfer funtion (MTF) of image space is better than 0.5 at 40 lp/mm, the distortion is less than 1%, and the illumination difference of 8frame images is within ±10%. The results show that the practical system has good imaging quality and each split image has high consistency.
A close circuit television(CCTV) lens was designed with a complicated reverse telephoto configuration to meet the requirements of wide field of view (FOV)and high resolution which was applied to monitor system. The FOV of the designed lens is 80,the F number is 3,the focal length is 5 mm,and the spectral range is from 486 nm to 656 nm.A 1/2 inch (1.27 cm) CCD was adopted with 7.5 m7.5 m pixel size. The modulation transfer function(MTF) of full FOV is more than 0.65 at 67 lp/mm Nyquist frequency and more than 0.85 at 1/2 Nyquist frequency, and the MTF value of FOV is more than 0.3 at 220 lp/mm, approaching to the diffraction limit. The wavefront map shows that the peakvalley (PV) value is 0.077 9 and rootmeansuare (RMS) value is 0.015 9, which can meet the Rayleigh criterion. The evaluation of optical system shows that the lens has low aberration and high image quality, satisfying the requirement of CCTV monitor system with wide FOV and high resolution.
In order to improve the autonomous navigation capability of the maritime operations ships, an autonomous integrated navigation model based on polarization neutral points of the entire skylight was put forward in view of the limitation of inertial navigation system/global positioning system (INS/GPS) integrated navigation method sensitive to interference. Based on the analysis of movement law of the polarization neutral points of the entire skylight, the polarization neutral points was used to determine the heading angle to aid INS for navigation combing with the speed of information measured by marine electromagnetic log. The Kalman filter was used in integrated navigation system for the information fusion. The Matlab simulation results show that the integrated navigation system can correct the INS navigation parameters errors effectively, which makes the heading angle, horizontal velocity and horizontal position stabilized in 0.6 , 0.7 m/s and 10 m，respectively， and can enhance the autonomous navigation performance of ships,moreover,it has a high degree of concealment and antijamming capability which has certain military engineering application value.
Aiming at the increasing demands for the search scope and oprerating distance of current military photoelectric searching and tracking sighting device TV camera system, through analysis and comparison of the mature TV camera system, a research direction using the structure with miniaturized fixed focal length, multifield of view (FOV) was clearly determined. Based on the research direction, a design idea using embedded modular front lens group optomechanical srtucture was proposed, which can realize the fixedfocal length, multiFOV TV camera system miniaturization design plan. After optical design and optomechanical integration design, a 4FOV TV camera with large optical zoom range(20), stable optical axis and 4 FOVs sharing 2 common imaging devices was finally completed, it can meet the needs of existing military searching and tracking photoelectric sighting device TV camera system.