Research On The Key Techniques Of Target Position Based On Image Matching

When unmanned aerial vehicle(UAV)is performing reconnaissance and positioning task,it is necessary to use the UAV airborne photoelectric equipment to locate the remote target.Because the task has high requirements on the accuracy and speed of target positioning,its success or failure almost depends on the quality of positioning system.It is of great significance to study the remote target location technology based on UAV.The target location technology on traditional uav photoelectricity equipment acquisition drones,photoelectric pod,the position and posture information of the target,through computer calculation to determine the location of the object,due to the photoelectric equipment and other related detection equipment installation error,measuring error and the precision of the device itself,the influence of the traditional target positioning precision of the positioning technology is not high.With the further research of image matching technology,it is of great significance for the development of airborne avionics technology to develop the algorithm of target location based on image matching technology.This paper focuses on the image matching algorithm of large viewing Angle and studies the key technology of target location in the environment of large viewing Angle.Thesis first designed the overall scheme of the target positioning system based on image matching,by airborne reference images with the target measured image matching your goal orientation,design the overall scheme of the target localization algorithm proposed in this paper,under the indoor environment,verify the feasibility of target localization algorithm,in a real outdoor environment,verify the target localization algorithm is practical,and the positioning of the target positioning error is analyzed.Firstly,the overall scheme of fast and accurate positioning system based on image matching is proposed,where the fast and accurate positioning of the target is achieved by matching the target image with the reference image under the large view.Then design the theoretical framework of this paper and the simulation platform for the validation of the theoretical part.From UAV to test the image will be extracted from the target area effectively,lays the foundation for this article later target positioning,this paper first puts forward the effective target area based on the target measured image extraction algorithm,the firststep for UAV measured concentration of images on a large scale matching,the second step record matching results,and use effective object extraction algorithm based on improved RANASC optimization algorithm to achieve the extraction of the target area effectively.In order to realize the existence of large Angle difference of satellite calibration images and UAV measured image matching,this paper further studies the big Angle of view image matching algorithm based on perspective transformation,puts forward a kind of image matching algorithm based on POS data,after complete the matching,using RANSACA optimization algorithm to solve a single should sex between image matrix,for later target positioning basis in this paper.In order to achieve the target positioning,is proposed in this paper on the basis of the above research target localization algorithm,the effective extraction in the measured images within the target area for selecting target image point as pending a pixel,image matching for UAV measured through big view image and calibration chart of single should transform matrix,the final target image point is mapped to a calibration who goal orientation on the drawing.