Research On Technology Of Position And Orientation Measurement For Vision Navigation Based On Coplanar Feature Points

In recent years, with low cost, large amount of information, strong anti-interference,self-contained, vision navigation technology gets more and more application on the UAV flightcontrol, autonomous landing, intelligent vehicle navigation, rendezvous and docking ofspacecraft, lunar exploration. Aiming at the existing problems in vision navigation technology,the paper takes the position and orientation measurement technology as the research object. Themain contents of study include:(1) The paper studies the application of vision navigation, summarizes the research situationof position and orientation measurement, and analyzes the existing problems in the visionnavigation.(2) Taking the technology of position and orientation measurement of vision navigationbased on coplanar feature points as the research object, the key technology in the process ofposition and orientation measurement is studied and analyzed, and image processing algorithmwhich is adjust to artificial targets is summarized. The algorithm includes image segmentationbased on mathematical morphology and self-adapting threshold value, feature points fastextraction and location based on pyramid search.(3) Through analyzing the artificial marks recognition and matching, the paper bringsRANSAC algorithm into the process of matching, and artificial marks recognition and matchingbased on inners detection is proposed which solves the matching problem of sequential images.(4) Errors which affect accuracy of the position and orientation measurement are analyzed.A single variable error model is deduced based on collinear equations and their linearization，andthe relation between image error and the accuracy of position parameters is proposed. Imageerror propagation model based on the principle of covariance propagation is deduced.(5) Position and orientation verification project is designed; verifies the reliability andaccuracy of position and orientation system. The affection between image error and the accuracyof position parameters is studied by mathematical simulation experiment. Different experimentsare designed by digital camera and industrial camera to verify the feasibility of the system inpractice of the position and orientation estimation.