In order to resolve the problem encountered in industrial applications where the object to be measured is longer than that of the laser beam used as datum, a laser relay measuring method was proposed. In this method, the object was divided into several segments and each of them was measured separately. There is a superposition part of a certain length between two adjacent segments. Based on the measured results of this superposition part in two separate measurements, the relation of the two coordinate systems was found with least square method. All data measured separately could be expressed in a unified coordinate system. The method, used for straight line type object to be measured as an example, was presented in details. Its error was analyzed and the method for selecting parameters in the laser relay measuring method was given according to measurement accuracy requirement. Factors influencing the measured results were discussed with an industrial application example. This method was used in track geometry surveying of railway for passenger dedicated line, getting satisfactory results.

It is difficult, as we know, to search for interference fringes in Linnik white light interferometry with an extremely short coherence length because of the optical path mismatch of two interference arms and out of focus of the reference mirror and the test surface. An automated method to tackle this problem is presented. The determination of best foci of the reference mirror and the test surface is implemented by the astigmatic method based on a modified commercial DVD pickup head. The astigmatic method is improved by setting a threshold value in the Sum Signal (SS) to truncate the Normalized Focus Error Signal (NFES). The Truncated NFES (TNFES) has a monotonic relationship with the displacement of the test surface and the zero crossing point identifies true focus. The developed autofocus system is confirmed experimentally with a dynamic range of 190 μm, average sensitivity of 70 mV/μm, average standard deviation of 0.041 μm, displayed resolution of 4.4 nm and accuracy of 55 nm. The minimization of Optical Path Difference (OPD) of two interference arms is carried out by finding the maximum fringe contrast of the image captured by a CCD camera with the Root Mean Square Fringe Contrast (RMSFC) function. Experimental tests show that the automated method can be effectively utilized to search for interference fringes in Linnik white light interferometry.

In order to get the precise measurement results of large-amplitude high-bandwidth non-contact vibration, a new optical model and an algorithm for vibration measurement are put forward. Far-field interference fringe position is related with the phase of two beams of coherent light. The phase difference can be obtained through the position of maximum intensity fringe, which expresses the vibration displacement. Simulation results of cosine signal with the amplitude 10 μm and sound signals with the frequency 44.1 kHz expresses that sub micron vibration and sound vibration can be obtained.

Optical filters are crucial devices for wavelength channel selecting in WDM fiber-optic communication networks. To measure the inner physical parameters of optical filters is significant for analyzing emerging new and more complex structures. A general and easy-to-operate method is proposed to extract physical parameters for complex filter structures with optical low-coherence measurement. By treating optical filters with digital filter concepts and Z-transform analysis, we can build the relation between interferogram peak values and transfer function coefficients with brief linear equations. These equations can be used to derive the exact expression of transfer functions, which can manipulate the physical parameters. For a given filter structure, according to relations between transfer functions and physical parameters, the physical parameters could be extracted. This is especially meaningful in the post–fabrication process.

An objective characterization method for lighting flicker based on frequency spectrum analysis is proposed. Time-varying light signals were Fourier transformed into frequency domain, filtered by bandpass filter that matched Contrast Sensitivity Function (CSF), and applied inverse Fourier transform into time domain again. Based on the modulated transient light waveform, Flicker Degree (FD) was defined as a measure of lighting flicker. Several kinds of typical illuminants were tested and their FDs were calculated. Results agreed well with previous experiment and showed more stable repeatability. The method combined with human visual system provides an objective evaluation for lighting flicker, and gives a hand with lighting optimization design and industrial standardization.

A new method for real-time rapid detection of arc-shape surface of polished metal based on machine vision is proposed. In the offline situation, the samples of artifacts under different light intensity are analyzed, the relation function of the background brightness component and the gray level is constructed, and the statistical characteristics of reflection component of the samples are extracted. Online testing, the image gray level and the corresponding brightness component are calculated. Then the reflection component whose gray level distribution is uniform is extracted from the image. Finally, the reflection component is processed with thresholding segmentation and decision is made. The experiments show that the proposed method can adapt to the changing light environment through one-time learning. The average time to detect an image is 40ms, and the accuracy rate is over 98%. Moreover, the system has high robustness, high real-time and high accuracy.

Because of large amount of calculation, it took a long operating time for the tensor vote algorithm to restore human broken target in infrared images. Thus it can't meet the need of real-time pedestrian detection. An improved tensor voting algorithm is researched. Firstly, most of the points beyond the vote field are filtered out by converting the strength threshold into distance threshold. Then, according to symmetry of two voters’ voting effect to each other, it’s not necessary to compute strength value again. Finally, we use the strength threshold to precisely filter out the points beyond the voting field. Simulation results show that the improved tensor voting algorithm can reduce calculation cost and shorten running time, without loss of the restoring effect.

Physical models of three rectangular nozzles with different aspect ratio were established, and the 3-D exterior flow fields of these rectangular nozzles were simulated through numerical method with the software Fluent6.3. Then the data of their flow fields such as temperature, pressure and density and so on were obtained. According to the characteristics of their plumes, the computational areas of infrared radiation were established. Through statistical narrow-band model based on Lorentzian profile, mean absorption coefficients of the plumes in narrow band l were gained. Then radiative transfer equations with gas medium were solved by Finite Volume Method (FVM). Finally, the spectral characteristics and the total intensity distributions in 3～5 μm of those exhaust plumes infrared radiant were acquired. The result improves that two peaks of plume infrared radiant merge in 2.7 μm and 4.3 μm and the infrared radiant of exhaust plume is weakened greatly with the increase of aspect ratio.

According to the detection of small targets of infrared image in the sky, in order to improve the real time of detection, a algorithm of realizing image preprocess is proposed by Field Programmable Gate Array (FPGA). The FPGA achieves the morphological filter of mathematics, the image segmentation and the run length code. The experimental results indicate that the method is excellent in real time. It is effective for detecting small targets in infrared image.

Infrared scanning imaging is necessary in remote sensing field. To improve the spatial resolution of infrared scanning image as much as possible, the applying research of Super Resolution Reconstruction (SRR) in infrared scanning imaging systems become the hot research point. Applying infrared scanning imaging system arranges dislocated multi-line array, through the procedure of SRR and the ways of getting the image sequence to be reconstructed, the key of improving the spatial resolution by SRR in the adopted experiment device are analyzed. The experiment shows that, as to the experiment device, the key factor affecting the quality of the reconstructed image is the noise caused by the large difference of the pixel value in the relative location in the image sequence. The comparative filter method based on the pixel of the image sequences can improve the quality of reconstructed image.

Sea-sky-line detection is an important component involved in infrared image Auto Target Recognition (ATR) techniques. Through analyzing the characteristics of infrared images under sea-sky background, a fuzzy recognition algorithm is proposed based on four feature score factors. Firstly, Radon transform is performed on gradient images in order to choose sea-sky-line candidates. Secondly, fuzzy synthetical judgment is made combining information of 4 specific features. Finally, sea-sky-line recognition result is given according to the judge values. The algorithm comparison and real data test results show that this method is widely applicable, powerful in resistance to interferences and noises and has a high recognition rate, which can be used to effectively extract sea-sky-lines under complex backgrounds, thus lay a good foundation for further application such as target recognition.

A novel binarization method for document images based on Curvelet transform is presented. The interference caused by local high lightness is eliminated to get a better image quality. Firstly, the Curvelet transformation is applied to the document images with local high lightness area, and the Curvelet coefficients can be got. Then, according to the feature of images represented by Curvelet coefficients, the Curvelet coefficients are enhanced nonlinearly to optimize the histogram distribution. Curvelet coefficients are transformed inversely to get the images, and then the Otsu method is applied to get the binary image. According to the binarized image, the OCR recognition results are got by the ABBYY FineReader10. Experimental results show that the highest recognition accuracy of characters could reach 94.81%. The performance of this method is better than the other four typical binarization methods.

Image registration is an important research topic in machine vision. A cross correlation algorithm based on Zernike moments is developed to address the image registration problem with an arbitrary rotation angle. According to the symmetric and anti-symmetric properties of trigonometric functions, a fast approach is proposed to compute the Zernike moments, which gains the moments merely by calculation of 1/8 unit circle. Afterwards, the matching point is calculated by the correlations of multiple moments. Thereby, the sub-pixel matching position is achieved by Least Square Estimations (LSE). Meanwhile, the rotation angle is estimated by synthesizing the Zernike moments phase information. Finally, experimental results show that the proposed algorithm substantially improves the matching efficiency for a prescribed matching accuracy.

Obtaining labeled training sets for hyperspectral image data classification is often time consuming and expensive. Therefore, classification of hyperspectral data with insufficient training samples catches the attention of researchers lately. A spectral and texture feature co-training algorithm is proposed based on a semi-supervised classification scheme. The two views of spectral and spatial information of hyperspectral imagery are combined by using Co-training mechanism. Our algorithm is well suited for the hyperspectral image data classification problem, especially when the size of training samples is small. Experimental results on real data demonstrate that the algorithm can yield good results in land-cover classification with hyperspectral image data.

High-quality and low frame-rate transmission of images has extensive applications. The inherence of uncertainty, packet-losing and bit error occurring of transmission channel often causes image degradation or un-decoding. In view of the case, we introduce a method that uses Multiple Description Coding (MDC) on Low-resolution (LR) sampled images to strengthen transmission robustness and then reconstruct the super-resolution with these multiple low-resolution images transmitted. Experiment indicates a gain of 2～3 dB in PSNR under the same conditions of channel transmission. Especially, this scheme is universal and worthy to be extended to other MDC technologies and SRR algorithms.

The human tracking is the key in the video surveillance. This text focuses on non-rigid objects, learning based on the HSI color model template. Each pixel of the template is modeled using two components with H and S. Get template from the current frame, statistic the information of every frame, and predict in the next frame using Kalman filter. Decide whether to update the template after comparing the forecast region and the template, reducing the amount of calculation. And for restraining the tracking window, information which is calculated from the color I is introduced. After that, the template and the window will be updated. The experimental results show that the proposed method achieves continuously tracking, and resolve the problem with the object disappeared.

The evolution and transmission of one-order self-Fourier soliton mixed pairs signal in optical fibers is studied numerically. In terms of split-step Fourier transform method, the numerical solution of nonlinear Schrodinger equation is gotten. It is proved that there is not any theoretical error in the inner part of algorithm. The results show that: the interaction between one-order soliton and disturbing one-order soliton and stability of the solution to the initial value problem of the one-order soliton pairs depend on the different structure forms of initial input. The interaction in one-order self-Fourier soliton mixed pairs is different from one-order standard soliton mixed pairs. The interaction in one-order self-Fourier soliton mixed pairs is similar to two-order soliton or quasi-soliton. It is not enough to recognize and classify the characteristics of interaction between solitons according to the order of soliton.

The research background, system advantages, and current status of fiber-array imaging laser radar are introduced firstly. We present a design scheme of the system and introduce its principle. Window scanning on fiber array is used to avoid fabricating electrophotonic detector arrays and to reduce the number of separate electrophotonic detectors. Every component of fiber-array imaging laser radar is analyzed, including the choice of fiber, the design of fiber array and the analysis of positive lens. The effect of the F number of positive lens on fiber-array receiver is discussed mainly and optimum value of F number is 5.47. The transmission efficiency of fiber-array imaging laser radar is deduced and its expression is presented in the end.

In order to improve performance of sampling optical signals as well as level of integration for Triplexer monitor, a wavelength-insensitive directional couple (WIDC) based on asymmetric-taper was presented. The WIDC adopted SiO2-on-Si buried type optical waveguide structure, with the taper waveguides constituting the coupling region at the center part and the S type bent waveguides used for the signal light into and output. The transfer matrix method was used to analyze the normalization of the coupling optical power at the light monitoring output port. The center region of WIDC was regarded as asymmetric rectangular waveguides and the lateral region was regarded as symmetric rectangular waveguides. Numerical simulations by both the Finite Difference Beam Propagation Method (FDBPM) and Matlab showed that when the input wavelengths varied from 1 300 nm to 1 600 nm, the normalized coupled optical power achieved by the WIDC was 4%～8%. In particular, the splitting ratios of WIDC were 92:8, 96:4 and 93:7 respectively when the input wavelengths were 1 310 nm, 1 490 nm and 1 550 nm correspondingly with the polarization variations less than 5% under the condition that maximum width and minimum width of the tapered waveguide were 5.50 μm and 3.35 μm respectively. The devised WIDC offers many advantages such as small size and high reliability and suitable for the Planar Lightwave Circuit (PLC)-based Triplexer.

Qualitative analysis methods can not completely avoid the effect from the ghost of compensator, thus a quantitative simulation analysis is proposed based on the theoretical analysis. Optical analysis software in the design stage is used to completely rule out the effect from the ghost. An example is addressed to illustrate the limitations of qualitative analysis methods, and to verify the validity of quantitative simulation analysis. This simulation provides a quantitative analysis method for the compensator design and analysis to improve the yield of compensator developed and the accuracy of testing, which is important for high-precision inspection techniques.

In order to eliminate the errors resulting from fluctuation of light source, the inner surface of integrating sphere is selected as reference in design of dual-optical path d/0 measuring condition on spectrophotometer. However, the repeatability and accuracy of measurement will decrease to a certain degree because of interference by diffuse reflected light of the test sample. This paper improves the dual-optical path d/0 measuring condition, composites LED as test light source, and configures a separated referenced white board substitute for inner surface of integrating sphere in the reference optical path. Then, the experiment is established. The results indicate that this improved dual-optical path d/0 measuring condition could effectively revise the errors caused by diffuse reflected light of test sample.

Colloidal Quantum Dots (QDs) acted as luminescent layer in light emitting device can be tuned across from visible to near-infrared spectrum by changing the size of QD, and QD-LED has a narrow bandwidth (Full-width at Half-maximum (FWHM) of the Electroluminescence (EL) peak of ～30 nm). Core–shell quantum dots (wavelength at 523nm and 608nm) with a CdSe core and a ZnS shell were used as emissive layers in the devices. Poly-TPD as Hole-transport Layer (HTL) and ZnO as Electron-transport Layer (ETL) were used in device. By fabrication of red and green emitting device based on quantum dots, we researched the feature of device which provides reference for quantum-dot-based LEDs used in the cockpit and medical apparatus. However, the brightness, efficiency and lifetime of QD-LED need to be improved to meet the requirements of commercialization in the near future.

A scheme of an angle measurement circuit for the two-speed resolver is designed, which includes the decoding and measurement of the precision. The 16-bits data of parallel communication of the two AD2S82A is read, corrected, and assembled by CPLD. According to the requirement of the main system, the CPLD completes the pretreatment of the data from the AD2S82A and outputs the parallel data of the angle information, the bits of which are up to 21-bits. With the CPLD, this angle measurement system becomes one single peripheral of the main system. A scheme of measurement test is carried out with collimator and prism in the laboratory. The results achieve the designed functions of the scheme completely and the error by Root Mean Square (RMS) of the single angle measurement circuit is 20 arcseconds, which reaches 9.46 arcseconds after correcting the systemic error with software. Lastly, the causations of the systemic error of the angle measurement circuit are indicated.