In cutting and welding processes of the H-shaped steel whose section size is about 300(W) mm×300(H) mm, 400(W) mm×300(H) mm or larger, a vision inspection method is adopted and set up in work field actually. By CCD image processing and programming, this system can measure geometry size and deformation of the steel with the resolution of 0.2 mm, and provide coordinate parameters for orientation and position of cutting/welding torch. For the problem that needs 0.7‰ resolution and large depth of field to measure the large section size of H-shaped steel, a vision imaging system is presented. From the actual measurement results, it can be seen that this system has a good measurement repeatability and stability. The theoretical analysis and experimental results show that the system achieves a resolution of 0.2 mm, and its processing time, from image capturing to final data uploading, meets the engineering requirements.

A Narrow Bandpass Filter (NBF) with finite impulse response for denoising the original signal is designed based on the characteristics of the frequency spectrum and the autocorrelation algorithm. Frequency spectrum correction technologies were used in beam expanded differential Laser Doppler Velocimeter (LDV) to process the signal simultaneously. Experimental results indicated that the random and high-frequency noise of the signal were mainly removed by the NBF. The frequency spectrum of the Doppler signal after the NBF was acquired by windowed Fast Fourier Transform (FFT). The measurement accuracy of frequency was enhanced by the frequency spectrum correction algorithm. Therefore, the signal processing system, including the NBF and spectrum correction, is suitable for the beam expanded differential LDV.

At present, the method we often use to measure the parallax of Large-aperture collimator is pentaprism method. Since artificial reading caused error, a kind of automatic inspection system for parallax of large-aperture collimator was put forward based on circular matching and data fusion. A high-definition camera was used to collect the image for large-aperture collimator. The center of the circle was found through OTSU, filtering, edge detection, circular catching and data fusion. A calculation result for parallax was given finally. In the process of calculating parallax, it can do more testing, so the curve of the parallax can been given out, which will make the collimator calibration convenient. Test results of automatic inspection system for parallax are proved to be accurate, and the automatic inspection system has good repeatability.

According to the requirements and characteristics of the hard resin lens on-line detection, an on-line defect detection method based on machine vision is proposed. An on-line detection system with two image acquisition structures is designed to detect all of the defects with different sizes. The first image acquisition system is used to detect the characteristics of the defects such as the point-like impurities and bubbles, and to obtain the information, including the length, amount and location of the scratches, and the location of sensitive areas contain the scratches and feathery impurities, and to control the working status of the second image acquisition system. The second image acquisition system is used to detect the diameter of the scratches and feathery impurities, after locating the position of sensitive areas with the movement of the conveyor belt and the one-dimensional linear guiding of the second image acquisition structure. This proposed method is able to solve the technical problems such as high cost, low-detection-speed and mass-data, which meets the needs of online detection. Experiments show that detecting speed of the system is 1.5 s / piece. Compared with subaperture stitching, the least amount of data is reduced by about 88% and a decrease of approximately 96% at most.

A kind of diffractive pyramid wavefront sensor for the measurement of wavefront aberration is designed based on the refractive pyramid wavefront sensor and the binary optical principle. The theoretical derivation of diffractive pyramid wavefront detection is made and the intensity distribution of the four sub-pupil image for the single term Zernike aberration through the two wavefront sensor is calculated. The average slope of the two devices on different aberration and energy utilization quantitative is quantitatively calculated and compared. The calculation results show that the diffractive pyramid wavefront sensor is capable of detecting the wavefront aberrations effectively and it can build miniature measurement system of human-eye wavefront aberration.

A triangular lattice is proposed as the basis of the 1 550 nm-band photonic crystal structure. We use a duty cycle r/a=0.3 in order to achieve the greatest band gap. Using the combination of point defect and line defect, we propose lateral coupling structure. With the simulation calculation based on the FDTD Rsoft software, the defect mode, transmission spectra and steady-state time domain response are analyzed. At last, the insertion loss, extinction ratio and response time of the optical switch are calculated. The results show that the insertion loss, extinction ratio and other performance of the structure of the optical switch are better, and it can fully meet the requirements of modern applications.

In order to minimize volume, lighten the weight of a satellite mounting effectively and satisfy the requirement of cold copy, accomplishing electric motor’s turn precision and cylinder memory at the same time, the design for satellite mounting dual numerical minitype multi-turn absolute photoelectric encoder is put forward. Based on turn precision and cylinder memory of electric motor, the two sets of absolute grating measuring system are applied to the mechanical structure. Based on the requirement of reliability, a dual numerical reading slit disc and a numerical reading system are developed. Finally, the precision test, the high temperature or low temperature test and the vibration test are implemented on the encoder. The experimental result shows that the satellite mounting dual numerical minitype multi-turn absolute photoelectric encoder can offer its resolution 80″, precision<±100″, and cylinder number 16. The encoder has the function of cold copy in the reading system and electric motor cylinder memory, light weight, high precision and high reliability, which can meet the requirement of satellite mounting.

A design method of LED uniform luminous intensity lens based on virtual target surface was proposed. Hemispherical virtual target surface was set to establish the energy relationship between chip and target surface, and get the shape of uniform luminous intensity lens. By optical simulation to discuss the influence of radius of virtual target surface to luminous intensity distribution, combining with uniformity of luminous intensity platform and ratio of luminous intensity platform width were proposed to describe lens overall uniformity index. A reasonable radius of virtual target surface was ultimately determined to get a uniform luminous intensity lens that meets the design indexes. Simulation and analysis show that the larger radius of virtual target surface, the smaller calculation error and higher luminous intensity uniformity of lens when the thickness of the lens and angle of lens design goal are constant, the larger angle of lens design goal, the higher luminous intensity uniformity of lens when the thickness of the lens and the radius of virtual target surface are reasonable. Using this method, a half-angle 60 degrees lens of uniform luminous intensity was designed. Its overall uniform luminous intensity is better, and uniformity of luminous intensity platform is 0.983 and ratio of luminous intensity platform width is 0.859, which meets the design requirement.

Laser resistor trimming machine, whose precision depends on the optical performance of the scanning lens, is widely used in tuning the chip resistors. According to the practical requirements, two kinds of F-Theta lenses working in different wavelength (1 064 nm and 532 nm) are designed. The Fields of View (FOV) are ±28°and ±25°. The focusing spot diameter are about 30μm and 10μm. The final design result shows that the calibrated F-Theta distortions are both less than 0.1％ and the energy utilization rate are nearly 90％ which means that the focusing performances are close to the diffraction-limited. Both lenses can be exchanged and located conveniently on account of the simple and compact structures. The requirements of resistor trimming are satisfied well.

For calculation models with symmetric characteristic, symmetry boundary conditions truncated by Perfect Electric Conductor (PEC) boundary and Perfect Magnetic Conductor (PMC) boundary are systematically analyzed, and a symmetric truncation method is proposed. Therefore the field values beyond truncation boundaries can be obtained so as to achieve the FDTD calculation on truncation boundaries. By analysis and calculation for dispersion characteristic of metallic-pole-planar SWS with floor structures, numerical results exactly consistent with global FDTD are obtained, thus feasibility and correctness of symmetric truncation method are verified. The results demonstrated by numerical simulation of dispersion characteristic of metallic-pole-planar SWS with floor structures are entirely equivalent to the results of global FDTD method. Calculation accuracy could be fully considered by the symmetric truncated FDTD algorithm. Numerical results show obvious advantages in both memory usage and computing time compared with the global FDTD method.

Three-dimensional video (3DV) will be the next step in the evolution of video services. Multi-view video using Hierarchical B Pictures (HBP) coding structure may be the main data format of 3DV. To address the problem of data packet loss during transmission of compressed multi-view video over wireless channel, we propose an intra-frame Error Concealment (EC) method for multi-view video with HBP structure. The concealment process is performed in three steps. First, we divide an intra-frame into background and foreground regions by utilizing the characteristic of short length of Group of Pictures (GOP) in multi-view video coding. Second, direct copy is applied to lost background blocks. Finally, adaptive orientation interpolation is used to conceal lost foreground blocks. The experimental results demonstrate that the proposed method outperforms the previous state-of-the-art intra-frame EC methods based on both subjective and objective measures, and decreases the effect of error propagation across view.

Ground filtering algorithms operate currently on either raw LiDAR point clouds or values that are derived by interpolation of raw data. In order to overcome interpolation error of grid elevation and time-consuming problem of raw LiDAR data, a new method of filtering based on 1-D progressive morphological method is proposed. This method can detect LiDAR non-ground objects from point cloud data by gradually increasing the size of the 1-D filter and using different elevation thresholds. Finally, we conducted a comprehensive test of the performance on fifteen study sites and compared our results to eight other publicized methods reported by ISPRS. The result shows that the filter can remove most of the non-ground points effectively. Overall, the 1-D progressive morphological filtering algorithm produces a promising performance in both urban and forest areas.

Mixture resolution format is an important representation method in multi-view video coding to reduce the bit rate. To obtain the same spatial and temporal image/video resolution scale in the low resolution view, a novel super resolution method is presented. The cracks of the warped reference high resolution frame are restored by non-uniform sampling principle, and the holes of that are filled with the corresponding regions of the low resolution frame. Then, based on the reverse hole filling method and 2D Discrete Cosine Transform (DCT) image fusion, the high frequency components, which are fetched from the warped reference high resolution frame, are merged into the low resolution view to construct the output high resolution frame. Experimental results show that the proposed method outperforms the traditional up-sampling and DCT-based up-sampling methods.

Referring to both situations when the color of clothes people dressed is similar to the background color and when lower limbs of human body have the shadow in front-view gait recognition are hard to recognize, an improved background subtraction by using the different threshold at above and lower part of human body divided by the hip is proposed, which is used for extracting the binary contour of the human body. For cycle detection, first of all, we use hip ideas to find the optimal point P, and then obtain the lowest points of the left foot and right foot. The cycle is calculated by the angle based on the distance difference of the two feet, while the angle as a dynamic feature is used to describe the gait. The next step is to extract the key frames and be normalized, and the unified Hu moment features are extracted, being integrated with gait cycle and body proportion as the static features. The method of the fusion of static and dynamic feature solves the low recognition rate which is caused by the single feature. Finally, the Support Vector Machine (SVM) is used for classification. The recognition experiments of this paper are all trained in the Chinese Academy of Sciences gait database (CASIA), and the recognition rate is greater than 97%. The results show that the proposed approach has an encouraging recognition performance.

The Shearlet multi-orientation features fusion and weighted histogram are proposed to overcome the disadvantage of Shearlet transform, which has data redundance in extracting features and cannot sparsely represent the global characters. First, Shearlet transform is used to extract the multi-orientation facial features. Then two coding methods are proposed to fuse the features from different directions of the same scale into a single feature, and the fused image is divided into a number of equal-sized nonoverlapping rectangular blocks, weighted fusion of each model using the Shannon entropy theory. Many experiments have been done on the ORL, FERET and YALE face database, which fully proved that this method has more advantages in terms of recognition than the traditional Shearlet.