Eyes Ccd Imaging Target Recognition Algorithm

This paper is arranged under the application of spacecraft rendezvous and docking(RVD) technologies. Our purpose is to solve the real-time measurement of relative positionand relative attitude between objective and tracking devices, when they are nearby.In the dissertation, we firstly elaborate the optical measurement principium based onbinocular stereo vision system, then prove the feasibility of this method in theory. Secondly,the theoretical derivation of relative position and relative attitude will be presented.When discussing the solve of projection matrix in position and attitude measurement,we define five coordinate system and describe their relationship, then obtain the projectionmatrix from camera's internal and external parameters. Before that procedure, we analyzeand compare several camera calibration tenologies, and choose a most appropriate one foruse. Its algorithm core and calibration procedure is presented as following. With the help ofImage Grabber SDK, we programme the utilities to complete the function of image captureand its pretreatment. Using the method of cluster in pattern recognition, the feature area ofcamera calibration moulding board image and objective image can be tagged, and theireffective information can be calculated and recorded, which is crucial for the followingcomputation. At the end, we emulate the algorithm for the measurement of relative positionand relative attitude by using C/C++ language and matlab in this experiment.Finally, we build the whole experiment system of objective measurement in laboratory,and verify the feasibility of this thought. The result proves that solution is correct andfulfill the application of binocular stereo vision measurement.