The Ground Simulation System Of Vision-based Space Rendezvous Radar

With the deepening of our human space exploartion, space rendezvous increasesfrequently, which leads to space rendezvous measurement technology becomes crucial.This paper systematically studied vision-based space rendezvous measurement methodsand techniques, in which the monocular vision measurement system based oncooperation target was developed, and research was extended to binocular and multiplecameras and non-cooperative target measurement method.First, the feasibility of space rendezvous radar terrestrial simulation was demonstratedand the vision measurement system performance was set. Then the basic principles ofvisual measurement and closely related PnP problem was studied and the visionmeasurement framework and pose estimation algorithms was selected. According to thevision measurement system analysis and design, the based on cooperation targetmonocular vision measurement system was developed, and the spacecraft rendezvoussimulation system was used for the visual measurement system testing, which showedthat the system performance met the expected standards.The research was expanded to binocular and multiple cameras vision measurement. Inthe particular configuration, one measurement method without matching was deduced,and the advantage of method, compared to traditional binocular method, was analysed.The algorithm is extended to the universal binocular and multiple cameras, and theinfluence of feature extraction error and calibration error were simulated and analysed.Finally, the uniaxial and triaxial tests were carried out on high-precision three-axisturntable, the experimental results showed that the algorithm had higher solutionaccuracy, especially at close range the algorithm has significant robustness.According to project requirements, research was expanded to the non-cooperativetarget binocular vision measurement furtherly. The characteristics of non-cooperativetarget were analyzed, and four kinds of selected primitives were tested, and ultimately ellipse contour was determined as the non-cooperative feature. For ellipse contourfeatures, one non-cooperative binocular vision measurement method was proposed. Thealgorithm simulation and data analysis showed that the algorithm had the goodrobustness and higher solution accuracy.