Coded structured light based on 3D detection is an active-vision method, which projects certain coding pattern through the projector, and obtains the matching points and depth information through triangle principle. For realizing fast measurement of surface, a pseudo-random array encoded method is used. Because the M-array generated by diagonal line is unable to meet the requirement of size for projecting desire pattern, an algorithm which may generate M-array of the large suitable size is studied. It lays the foundation for improving the measurement of spatial resolution and measuring precision. Over the field of three elements, monochromatic light is used, and we choose the stripe, circle and disc as M-array symbols. The symbols are easy to be identified and distinguished, provide information for decoding image search and increase the anti-interference ability for the environment. The pattern is robust as it allows a high error rate characterized by Hamming distance higher than 3. Results of reconstruction verify feasibility and effectiveness of presented method.

Aiming at on-line measurement of micro-bubbles’ size distribution in water, a proper image proceeding method was proposed. By regarding micro-bubbles’ pictures taken by micrography in laboratory as research object, the idiographic means such as image enhancement, image segmentation, morphological transformation and particle measurements adopted in the method were introduced, and the size distribution in a micro-bubbles’ picture was given. Experimental results and analysis show that the proposed method can obtain the size distribution of micro-bubbles in water effectually, and separate conglutination bubbles. The method is suitable for on-line measuring bubbles’ parameters in water.

A real-time measurement of the polarization of faint light is brought forward, and it has many promising applications. The measure can be actualized by a partially polarizing beam splitter and two polarizing beam splitters. Two phasic-compensators are appended to optimize the system. The optimization theory is educed, and the optimized apparatus matrix is given. An optimal polarimeter is given. Its Muller matrix is approached to ideal, and the error is less 0.25%. The setup is able to reconstruct the stokes vector of any input polarization to a fidelity of at least 97%.

It is principal to detect lane robustly and rapidly for intelligent vehicle based on the information of road marking or road region. The road image is divided into tow parts called near area and far area based on pre-knowledge and human visual experience. A linear model is adopted to fit the lane marking in near area, while in far area, the lane marking with the lane model is switched between linear model and cubic curve model. Combined with the gradient value, gradient direction and gray information, discriminant function of the probability is derived. Then the improved ParticleSwarm Optimization (PSO) algorithm combined with genetic algorithm operators is used to quickly search the optimal model parameter of the discriminant function to implementation lane detection. The results of the real road image experiment show the proposed method can robustly and rapidly detect the lane markings even if there are some interference factors in the road such as shadow, non-uniform illuminance, vehicle barrier and soiled lane boundaries.

Referring to the deformed scan image of the moving target obtained by a line-scan camera, a new method was proposed to estimate the motion parameters using a line-scan camera. The method, based on one-dimension imagery model of the line-scan camera, made use of the temporal and spatial relations of the key points on the target image to establish a serial of nonlinear functions related to the motion parameters, which transforms the estimation of the motion parameters into a problem of nonlinear optimization. In the analysis of the deformed license plate images of the moving vehicles, speed and direction were estimated. The experiment results show that, the method has good practicability and accuracy.

In this paper, we experimentally tested the opto-spectrum output and the pressure amplitude of intensive acoustic wave by HR4000CG-UV-NIR spectrometer and NI-PCB sensor, analyzed the energy transformation relationship of sound and light, and observed the influence of discharge parameters on the opto-spectrum. The experimental results indicate that the spectrum of the light is of continuous spectrum and its dominant energy is around 500 nm. The light energy output and sound energy output compete with each other. If we raise the discharge voltage and reduce the gap distance of electrode tips, the spectrum amplitude and its half peak width will increase. If we increase the gap distance, the light energy will spread to the far ultraviolet zone. And if we change the steel tip with copper tip, the spectrum energy distribution and amplitude will change.

A method for capturing dense 3D shape of scene consists of spatially isolated objects with high-speed is proposed. In this method, three patterns are projected onto objects and the deformed patterns are captured at the same time to obtain their 3D shape and texture. The two patterns are π phase shifting sinusoidal fringe pattern. The phase distribution of deformed fringe pattern is obtained with Fourier transform. The texture of objects is obtained by adding these two deformed fringe patterns. Another pattern is used for identifying the order of fringe. Then, absolute phase is obtained and phase of isolated objects can be unwrapped as well. The pattern consists of several vertical slits with width the same as the period of sinusoidal fringe pattern. The slits are coded with three gray levels. A slit consists of one gray level or two gray level distributed periodically. So, six codes are obtained. The slits are arranged according to a pseudorandom sequence consisting of these codes. The order of fringe pattern is determined through subsequence matching. A high-speed capture system was constructed with a modified DLP projector, which can project patterns at 360 fps, and a high-speed camera. The 3D shape and texture capture speed of 60 fps with resolution of 640×480 points and that of 120 fps with resolution of 320×240 points was archived in experiment.

According to the demand of diameter pipe-line accuracy measurement in the major industry currently, a measure system based on theory of sub-pixel and two CCD-cameras is designed. The system horizontally shoots the steel pipe by two CCD-cameras and determines the location of the steel pipe, cameras, related equipment and the external-inner of measurement system by cameras calibration method. It determines roughly edge areas of the image in two CCD-cameras by the method of differential and determines accurate edge of pipe-line by points in the roughly edge areas in sub-pixel measure method. According to the coordinate attained, external quadrilateral is formed and least squares are used to find the center of the inscribed circle (the center of steel tube) after the diameter is attained. Experimental results indicate that the accuracy of steel pipe diameter is improved, and the accuracy of steel pipe diameter error is below 0.2 percent. According to the research of measurement of steel pipe based on theory of sub-pixel, the system has simple structure, strong resistance to vibration performance and high accuracy, which is applied to industrial high-speed on line detection.

The space resolution of stereo vision system affects the accuracy of the measurement. The space resolution differs at different points and in different directions at the same point. Based on the formula of the stereo vision system, the space resolution of all the points is calculated and analyzed in observation field. The results show the space resolution is the highest with axis angle between 80°and 100°. And the space resolution in the X direction is higher than that in the Y direction. The experiment result is similar to the analysis result. Finally, we get the best structural parameters of stereo vision system according to the space resolution.

In the production process of billet detection and recognition, how to accurately determine character string target location at the end of billet from the complex illumination scene is a key technical problem. In order to solve this problem, a method to complete character string location precisely is proposed. This method consists of Mean Shift restrain complex scene image, multistage segmentation filtering and clustering processing, and least squares trim angle adaptively for finding the character string interest area. The experimental results show that the location method cancomplete location precisely in various billet character string of complex scene and has better stability and accuracy than the traditional projector location method, and solve the location problem of billet character string on complex scene, and provide key technology in the billet character recognition of work.

Total variance is used to analyze the random noise of Fiber Optic Gyroscope (FOG), which efficiently solves the problem that the Allan variance calculation is easy to be unstable at long-term τ values. But there is algorithm error with total variance when analyzing the random noise of FOG directly, which cannot really reflect the variances of rate random walk noise, quantization noise and exponentially correlated noise. So an improved method based on total variance is presented to suit random noise analysis of FOG. The estimation performance with the simulated random noise shows that the improved total variance can efficiently identify the noise types and levels. According to the analysis of measured FOG signal, it can improve the confidence in the case of great mean index, the values of variance have a good stability, and it is better and more exact than Allan Variance in identifying noise coefficient.

In order to overcome the uncertainty of target motion in non-orbit target tracking problems, a practical filtering algorithm, which compensated the variance of process noises, was proposed in this work. Based on the principle of strong tracking filter, the prediction covariance in Kalman filtering equations was modified real-time by monitoring innovation sequence in the proposed algorithm. In this way, the unknown variance of process noises was compensated effectively. Due to innovation-monitoring procedure, the proposed algorithm was adaptive to the uncertainty of target motion model. Simulation results of a rotary target tracking problem show the validity of the proposed algorithm.

Aimed at the defect of the traditional mean shift tracking algorithm by using symmetric kernel function, which contains amounts of background pixels, an enhanced mean shift tracking algorithm is presented based on active asymmetric kernel to improve the tracking accuracy and stability. The paper firstly described the calculation method of template center which is the key issue in introducing asymmetric kernel function into mean shift algorithm framework. Then, to combine the expression and evolution of asymmetric kernel function, level set contour evolution algorithm using regional similarity is presented. Finally, the asymmetric kernel function update strategy is introduced. The above three points constitute the mean shift tracking algorithm based on evolutive asymmetric kernel function overall context. Experimental results show that compared to existing methods, the mean shift tracking algorithm based on evolutive asymmetric kernel presented has higher accuracy and reliability, as well as meets the real-time requirements of general tracking tasks.

To comprehend the mechanism of multi-sensor passive localization thoroughly and then improve its localization precision, several related issues in dual-sensor bearing-only localization are theoretically analyzed. First the factors that can affect the dual-sensor localization performance are studied, which indicate that only when the measurement error is small enough and the cut angle between the sensors is not close to 0 and π, the dual-sensor system is unbiased and may be consistent. Next we point out that the negative impact of small enough cut angle on the localization precision is greater than that of large enough cut angle when they have the same magnitudes in terms of 0 and π. Finally, the equivalence of two localization error covariance methods is analyzed. Simulation results verify the above analysis, which indicate that the corresponding conclusions can be applied to multi-sensor passive localization algorithms based on sensor management.

Most conventional tracking gate techniques only use the targets’ kinematic measurement information, which typically results in great uncertainties of measurement-to-track association for multi-target tracking in clutter. Considering that the target class information can be derived from attribute sensors, the tracking gate technique for joint target state-class probability density is proposed. Firstly, a joint probability density description of the target state and target class is given, by which the method for constructing the class-conditioned gates is developed. In order to comply with nonlinearity in practical application, evaluating of the gate threshold adopted an algorithm based on simulation. Scenario 1 shows that if the target predictive measurement density is skewed distribution, the simulation-based threshold-evaluating algorithm can achieve optimal gate volume; and scenario 2 presents a target tracking process for ground formation. Compared with the data association methods using traditional tracking gates, the class-conditioned gate technique significantly improves the probabilities of the measurement-to-track association.

Motion detection in moving background is difficult. Background compensation followed by inter-frame difference and optical flow segmentation can separate object from background. Problems are that the former requiresrobust background estimation and may induce holes, while optical flow estimation is often invalid with noise, changes in illumination and high speed object. Two methods may be invalid especially when illumination changes. A new approach is presented based on feature displacement vector field fuzzy segmentation and OTSU method for motion detection in dynamic scenes without any prior information about object or dynamic scenes. A robust feature correspondences set is obtained by an improved matching strategy of SIFT, and Fuzzy C-means clustering algorithm is used to classify the generated feature displacement vectors. OTSU algorithm and morphological operations are performed for image thresholding, which modifies the convex hull of detected features. At last, moving object region is segmented. Compared with inter-frame difference and optical flow estimation, experiments demonstrate that the proposed method can detect moving object in the condition of noise, illumination change and object moving at a high speed.

In order to improve the measurement efficiency of theodolite based on visual guidance, a method for measuring the target point in the space precisely is proposed. Applying electronic laser theodolite to do the single-axis scanning, we can obtain the parallel grid model of planar scanning trace. This model efficiently helps us establish a relationship between the scanning length and the rotating angle of the single-axis. Under the consideration of the remarkable effect that the lens distortion may be caused on the measurement results, we employ an interpolation algorithm in order to realize a more accurate measured value. The result shows that, after the distortion reclamation, the accuracy within thevolume of 10 m×10 m can achieve 3″, which not only improves the measurement efficiency but guarantees the accuracy as well.

The imaging rotation problem occurred because of the rotation between the object and the imaging space when altitude-azimuth opto-electronic imaging tracking and measuring device observed a target. It made the target extraction and measuring failure. So an imaging rotation model was structured based on the theory of spherical trigonometry and coordinates transformation, the relationship between variables of this model was analyzed, several methods of eliminating image rotation were listed, and a design of CCD camera rotation was adopted to eliminate image rotation. The performance index, structure and control process of this unit were described specifically. This design has been applied onan opto-electronic imaging tracking and measuring device. The factory test and outfield trial results proves that the design of eliminating image rotation with CCD camera rotation satisfies the demand of project, and it ensures that the tracking and measuring device can extract and track targets truly and stably.

According to the traditional AdaBoost algorithm with fast detection but low accuracy, a modified AdaBoost algorithm was presented to enhance the accuracy. First, the algorithm extracted Haar features of human face by rapid integral image. On the basis of this, it set the threshold value to modify the traditional AdaBoost algorithm and found the optimal weak classifier during each test, and then it cascaded them into strong classifier. Finally, strong classifier was developed to distinguish Haar feature and detect the part of face from images. The sample test results show that the classifier accuracy of FERET database is 96.07% and the video images is 96%. The experimental results demonstrate that the algorithm of human face detection designed can not only detect static images but also detect video images, which lay the foundation of face recognition and provide a kind of effective method for research of computer vision domain.

Aiming at a phenomenon that the acquired RMB number is colorful and noised image, a method based on HIS space and improved clustering algorithm for RMB number color image segmentation is proposed. The HSI space is a colorful segmentation space, which is adopted. The 3-D searching problem is transformed into three 1-D searching problems in the HSI space. Three gray histograms on the 1-D direction is obtained. By the gray scale value of every pixel in current neighborhood 3×3 and the gray scale of the current pixel, the gray scale value p(m) of the current pixel ofcluster algorithm is determined, and improved C-means cluster method is used to distinguish the clustering center of character from non-character. The foreground and the background of RMB number image is clustering judged through using Euclidean distance. A colorful RMB number image is segmented and the segmentation method is adaptive. Experimental results show that the proposed segmentation method is not influenced by image noise and local edge change, and the amount of data is less than that of pre-transformation. This method is effective and robust for alphabet segmentation and number segmentation.

A high-resolution volumetric display system by scanning with multi-screen in parallel motion and a corresponding principle prototype is described. Through a DMD, 2D image sequence is projected on imaging screens moving in circular translation, and the dynamic refreshed image sequence is perceived as a 3D image with physical depth because of the persistence of vision. The optical projecting module keeps the optical path between the projector and the imaging screen to be constant. Hence, a clear and stable image space is constructed. A high-speed data channel is designed on a FPGA chip, whose bandwidth achieves 18.75 Gbps. Each volumetric frame contains 512 frames and over 400 million voxels at 12 Hz refresh frequency. The principle prototype is successfully manufactured, and the 3D display effect is achieved as anticipated.

Gamma correction is used to correct the nonlinear relationship between luminance and brightness, to vision information of reproduction source objects truly. So the principle of Gamma correction in Organic Light Emitting Diode (OLED) is described. For the gray scale distortion of the SVGA050 micro-display which is caused by the difference between the optical and electrical characteristic property, a Look-Up Table (LUT) Gamma correction system of AM-OLED is proposed, which is used the PIC18LF2550 as the MCU and the master of the I2C serial bus. The development of the hardware system measurement verifies the correction of VGA and CVBS video input and the picture quality also apparently improved.

Currently FBP reconstruction algorithms assume that X-ray which is used to produce projection data is of monoenergy. This presumption is incorrect in reality, which causes the mismatch between the imaging process and the reconstruction algorithms, and leads to the beam hardening phenomenon. We analyze a variety of beam hardening correction algorithms and point out that single-material object linearization is a better method. An improved linearization method is proposed, which solves the problem of how to select the degree of the polynomial and extends the application range of polynomial fitting based on reprojection to cone-beam CT imaging. Finally, experiments are given to prove the practicality and validity of the proposed method.

Imaging detection and Charge Coupled Device (CCD) detector was adopted in fluorescence detection of DNA capillary electrophoresis in order to decrease the detection width and improve the efficiency and speed of DNA separation. The importance of small detection width for high-speed capillary electrophoresis, especially the separation performed in micro-fluid device, was illustrated according to theoretical calculation based on resolution theory of capillary electrophoresis. The performances of imaging detection and conventional intensity detection were measured and compared. The method of data processing to intensify the DNA band signal of CCD detector was proposed, and verified by simulation. Imaging detection and its data processing method can remarkably decrease the detection width of capillary electrophoresis, and then improve the efficiency and speed of DNA separation.

Based on the optical partially coherent theory, a new integral phase contrast modulation transfer function theoretical model is established in consideration of the X-ray source area parameter. Through the simulation and calculation, a comprehensive analysis of the spatial coherence of micro-focus X-ray phase contrast imaging system is displayed. A new source size selection method and parameter optimization method are put forward under the actual experimental condition. The results show that the model containing X-ray source area factor can reflect the entire performance of optical imaging system better, and avail to the imaging parameters optimization design.