Study On Data Processing Techniques In Multi-Channel Optical Imaging System

The Multi-Channel Optical Imaging System to be researched in this paper is the new generation optical imaging system. Its principle system consists of four children imaging systems with large field of view and a high-resolution center major imaging system. The four children imaging systems are used to search, capture, detect, recognize, track and position the moving point target in the large field of view. The center major imaging system is used to observe the moving target detected by the children imaging systems with a high resolution. According to the functions reqiured in the Multi-Channel Optical Imaging System, the data processing techniques of the system are deeply researched.The main researched contents and their achievements are as follows: (1) The study of the real time detection of the moving point target in the children imaging systems. All the excellent detection algorithms of the moving point target have huge calculating quantity and need large storage, so the two steps detection method is used to reduce the calculating quantity. The first step is the preprocessing of the image sequence, and the second step is the detecting of the moving point target from the small number of the candidate pixels remained by the preprocessing. After the analyzing the usual preprocessing algorithm, the preprocessing algorithm named the Infinite Norm of the Difference Vector is proposed in this paper. The experiment results and its application prove that this algorithm can get rid of the low frequency scene and noise effectively, and remain a small quantity of candidate points which maybe include point targets. The preprocessing reduces the computing quantity of the following detection and recognition of the moving point targets, which makes the real time detection and recognition be realized. The new concept named the Discontinuous Frame Difference is proposed. It is applied to the classic Optical Flow Algorithm, and the Discontinuous Frame Difference Optical Flow is established. The parallel detecting system of the Discontinuous Frame Difference Optical Flow, which can fully exert the detectingability of the algorithm, is established. The experiment results prove that the new algorithm not only can detect the moving point target effectively and reliably, but also overcome the shortcoming of the traditional optical flow that could not detect the moving point target whose moving displacement is less than one pixel between two continuous frames.(2) The study of the determining and recognizing technique of the detected moving point target. The features and shortcomings of the already existing recognition algorithms are analyzed in this study. The determinant rule of the moving continuity and consitency of the moving point targets in the mono-aperture imaging system, based on the characters of the gray intensity and moving of the moving point target and its number of pixels, is proposed. According to the feature of the view-field overlapping of the multi-channel imaging system, and the space parallel processing to reduce the determining time, a more quickly and precisely determinant rule of the moving point target is established. The experiment results prove that this method can determine the moving point target effectively.(3) The study of the positioning method of the moving point target. The mathematical model of the non-linear system consisted of the moving point target and the observing system is put forth. The positioning method of the moving point target based on the analysis of the image sequence and the non-linear Kalman recursive filter algorithm is established. The experiment results prove that the proposed algorithm can overcome the classic positioning method could not determine the position of the moving point target to implement the passive positioning of the moving point target.(4) The precondition of the detecting and recognizing of the moving point target in image sequences is the registration of the image sequence. The sub-pixel image registering techniques are summarized and analyzed in this paper. The ph...