Research On Tracking Method Of Six-Rotor Aircraft Based On Image Detection

In recent years,with the rise of UAV aerial photography,UAV has been widely used in the fields of military reconnaissance,environmental detection,disaster relief,etc.,because of its small mass,flexibility and convenience.However,it has the limitation on the image size,and it is difficult to reflect the actual situation on site through single image.Especially in the process of fixed-distance tracking,UAV cannot effectively identify and extract target information,which may lead to unsatisfactory or even ineffective tracking.It is necessary to mosaic two successive images taken by UAV,which can effectively increase the proportion of the tracking route.Therefore,the research of image mosaic technology for UAV has become increasingly important.This paper mainly introduced the tracking method for the six-rotor aircraft based on image detection,including the MATLAB mosaic for the image taken by the six-rotor aircraft,the detection and identification of the track target line,the image acquisition and extraction of the six-rotor aircraft and the final tracking control.In the research process of MATLAB mosaic technology for six-rotor aircraft images,the following problems were solved.Firstly,aiming at the problems of different shooting angles and image blurring caused by aircraft jitter,image graying and de-noising were carried out,which made SIFT matching to complete better feature point matching.Because there may be mismatches in this process,the four-point RANSAC algorithm was integrated to purify the matching of SIFT.The comparative experimental data showed that the effect was remarkable.Secondly,based on the analysis of the Image warping theory,a fast inverse image warping was proposed to meet the real-time requirements of UAV images.The gradual-in and gradual-out image fusion also eliminated the brightness difference and splicing gap of the final image mosaic.During the process of detecting and identifying tracing lines,a fast fusion algorithm of MATLAB image mosaic and Hough transform was proposed,aiming at the fact that the Hough transform detection of fixed distance cannot effectively capture and extract the route information under interference.The experimental results showed that the improved algorithm in the present study not only guaranteed the effectiveness of the final route extraction,but also greatly increased the ratio of the target route in the image and reduced the mismatch rate of the route.Finally,the actual flight experiments were carried out on the experimental platform,including fixed altitude flight experiments and tracking experiments.The experimental results showed that tfhe proposed algorithm in this study is effective and feasible for the tracking research of the six-rotor aircraft.