Research On Target Positioning Method Of UAV Based On Improved Mean Shift Algorithm

With the development and progress of modern society,the goal orientation has been paid more and more attention.Traditional target positioning methods such as radar method and base station layout method require too much external hardware and are easily affected by the surrounding environment.Due to its high maneuverability and ease of operation,drones play an increasingly important role in target positioning.Image matching plays an important role in drone target positioning.The traditional Mean Shift matching algorithm based on color histogram feature,gradient direction histogram feature and gradient magnitude histogram feature has various defects,which cannot be used for UAV target positioning,so it needs to be improved to allow Can be used in the target positioning of the drone.This article is an in-depth study on this issue.This paper first reviews the development of target location methods and image matching algorithms at home and abroad.Then the traditional Mean Shift algorithm based on color histogram feature,gradient direction histogram feature and gradient amplitude histogram feature is studied.The feasibility of their target location is demonstrated by experiments.On this basis,this paper proposes a new method.The feature is the gradient magnitude histogram feature based on gradient vector flow(GVF).Using the Mean Shift algorithm based on this feature can solve the problem of the above method,it does not have a certain robustness due to color,illumination loss of the target;nor does it mean because the target has a spatial position change in the image.The figure can’t be reflected;it shows good continuity and smoothness in the similarity matching surface.The accuracy of the proposed algorithm in target location is demonstrated by simulation experiments and error analysis.