Improving Optical Flow Algorithms In UAVs Applied Research In Positioning

With the advancement of science and technology,more and more attention has been paid to UAV application research,especially the precise navigation and positioning technology of UAV to target position has become a hot spot in recent years.Vision algorithms based on image processing are the key technologies for UAV tracking and navigation.For complex flight environments,conventional optical flow algorithms cannot achieve better positioning and tracking results.For example,when the light is dark,the UAV positioning cannot be It is stable and has a large error.When the drone lens is blocked by obstacles,the error rate of the conventional optical flow algorithm is high.In response to the above problems,the paper mainly carries out innovative research and practical application work from the following aspects:(1)Research and improve a feature point extraction algorithm.For UAVs in complex areas and complex environments such as occluded objects,the adaptive area threshold is first set to improve the feature point allocation,and then the feature points are further screened according to the Shi-Tomasi algorithm to further ensure the reliability of feature point extraction.Then,the gray-scale centroid algorithm is used to add direction information to the extracted feature points,and rotation sampling is used to perform binary encoding output on the sampling area,which reduces the error caused by the rotation of the UAV body.(2)Research and improve an optical flow adaptive insertion algorithm based on feature point extraction.Aiming at the complex environments such as low light and occlusion that are common in UAV image positioning and tracking,the matching rules are improved on the basis of the original optical flow matching,and the extracted feature points are further screened by the proposed forward and backward bidirectional tracking strategy.Then add the adaptive insertion method for drones to reduce the impact of complex environments on the rapidity of the algorithm,and finally correct the optical flow calculation to reduce the error caused by the optical flow pyramid scaling and increase the optical flow algorithm in different environments.stability.The experimental results show that the improved feature point extraction algorithm designed in this paper ensures the flight stability of the UAV in conventional and special environments,and effectively improves the robustness of the UAV for visual positioning using optical flow.Compared with the conventional feature point extraction optical flow algorithm,in the occlusion and low-brightness environment,the algorithm speed is increased by nearly 20%,and the positioning error is reduced by about 40%.