3D Pose Measurement Of The Complex And Rigid Flying Target Based On Matching Perspective Projection

3D pose parameters of the target space is a key parameter which reflects its state of motion, these parameters have an important role in target motion analysis, performance evaluation of objectives and locating the failure. In the military field, the motion state of artillery, missiles and aircraft flight is an important parameter to evaluate its performance. In the process of range’s test and evaluation, the test method of posture parameters outside objectives is often needed to measure its flight attitude data. The pose measurement method outside objectives which is existing in the range can only estimate the pose parameters of the flying targets that is swivel body. The pose measurement method of the rigid and flying target whose geometries is complex currently is not perfect. In this paper, flight target images of single or multiple photography theodolite are simulated, then the simulation image and the real image will be compared, thus the pose parameters of flight target can be got.In this paper, the research work carries out in three areas:First, perspective projection simulation platform is built. The relationship between perspective projection imaging for Open GL and the ideal camera imaging model is analyzed, this paper proves that perspective projection imaging for Open GL and the ideal camera imaging model are the same without regard for Open GL frustum cut. View transformation matrix and model transformation matrix are derived, when photography theodolite imaging is simulated in Open GL.In this paper, when the simulation image and the real image are comparing, the target area feature and the outer contour feature of the image are used. According to the characteristics of image of target in the sky, this paper designs a image segmentation method which joints dynamic threshold of gray and global threshold of gradient to get the target area of an image; According to the requirements of iterative optimization algorithm based on matching the outer contour point set for missed detection and error detection, this paper designs the outer contour point of subpixel precision extraction algorithm of target in the real image and the simulation image.Finally, the paper studies a pose measurement algorithm of the flying target, and builds a pose measurement framework of the flying target. This framework utilizes the respective advantages of the pose initial algorithm(which is based on the pyramid region matching search algorithm) and the pose increase precision algorithm(which is iterative optimization algorithm based on matching the outer contour point set) and avoids each other’s flaws. It enhances the robustness and accuracy of pose measurement of target, but this is at the expense of longer time-consuming. The algorithm discusses monocular and binocular cases. Experiment shows: if the accuracy of the given initial pose value is not high, this algorithm can also pull the final pose result into the nearby of true value, and algorithm precision of binocular case is higher than precision of monocular case.