The Research Of Measurement Accuracy For Loss Compression Using JPEG2000 Of Stereo Image

As the successful implementation of Chang'E1 lunar exploration and the further demonstration of the second lunar exploration program , the second phase of China's lunar exploration project is about a formal, and the primary goal of the research is the deep space-exploration. At present, the scheme of the Chang E project is that soft landing, patrolling and reconnaissance on moon surface by unmanned exploration are achieved in 2012. Due to the limited delivery capacity of deep-space image transmission, the loss compression of images is needed for the lunar rover to obtain sufficient information from the images with which it can recognize the surrounding environment.Environment recognition system based on binocular vision has some features such as compact structure,low energy consumption,abundant information amount and high identifying precision. This system has been paid wide attention and already been applied to the Martian rover patrolling explorer to carry out navigation and obstacle avoidance. As far as the Martian rover is concerned, ICER is used for the loss compression of the images collected, and then the images compressed are transmitted to the ground where reconstructed images are obtained by means of automatic stereo matching algorithm to evaluate the surrounding terrain.The advantages of the image loss compression using JPEG2000 which is the leading method in the image compression field are preferable compression characteristics, quality and resolution progressive transmission and so on.The Research of Stereo Image Loss Compression Using JPEG2000 to the Measurement Accuracy,which is the title of this paper. Five main parts included in this paper are as follows:(1) Construction of binocular vision calibration system; (2) Calibration of binocular vision system; (3): Stereo matching of images; (4): Relationship between the matching points and the three-dimensional reconstruction under different compression rate (5): Research the accuracy about compression rate and distance measurement, give conclusionThe paper introduces theories of active vision system and passive vision system. Although active vision system has some advantages in vision field and flexibility, the found of this system is relatively complex and requires additional measure equipments. So the passive vision system is chosen because of its advantages such as simple system, less calculation, mature technology and extensive application. The same time various instrument and parameter are decided to found this system in view of requirements. Finally, with the laser range finder assemble the binocular vision calibration system.The paper calibrated the system and determined extrinsic and intrinsic parameters of camera to set up corresponding relations between object points in world coordinate and image points in image coordinate correctly. Camera calibration generally includes traditional calibration based on referenced objects and self-calibration based on active vision. The paper studies both calibrations. Self-calibration has advantages of convenience and quickness in use, but it has great calculation and unstable value, so the traditional calibration is chosen because of its simple algorithms and stable calibration result. In order to get minimum calibration error, Least Mean Square method is used to calibrate the system. When calibrating select the laser range finder as the benchmark, in the end some experiments are carried out on this algorithm in order to objectively evaluate the accuracy of binocular vision system.Image matching is the most important and difficult step in binocular vision and there isn't a widely-used matching algorithm to this day. The paper summarizes a few of current representative algorithms including region matching, feature matching and phase matching. And then compares their environment adaptive ability, advantages and shortcomings. In allusion to the need of the research, the paper adopts the Harris corner algorithm which is based upon the feature matching. Using the algorithm acquire the features and matching from the images which are collected through binocular vision. In the end some experiments are carried out on this algorithm and has a conclusion that it has high precision when image gray has some changes and in other cases it will result in mistake matching.Firstly, a surface laser scanner is used as the benchmark to calibrate the binocular system, the experimental results prove that the accuracy of the binocular system is low because of the accuracy problems of the surface laser scanner. A polygon target is set up on the board, and JPEG2000 is used to compress the images, which are collected by the calibrated binocular vision system, with different multiples, and then the compressed images are restored. By analyzing the matching points and the position deviation of the polygon center obtained from the matching algorithm, the impact of compression rate on three-dimensional reconstruction can be derived. In order to have quantitative analysis between image compression and three-dimensional reconstruction accuracy, this paper makes use of the point laser range finder as the benchmark to calibrate the binocular system. As a result, the accuracy of the binocular system improved. Laser-point played in the stones displays light spot at the binocular image. The paper select nine brightest pixel points from the light spot region, and then calculate the average of horizontal, vertical coordinates of the nine pixels. The average is used as U, V coordinates of the target point. Then using projection matrix we can achieve the three-dimensional coordinate of the object point. At different compression rate the polygon composed of nine pixels will change, so the three-dimensional coordinate of the object point will change. This paper analyses the impact between compression rate and measurement accuracy from measurement error of Y detection and overall, gives the recommendation of a minimum rate finally.The application software is developed through Matlab 7.0.