Blurred and flickering imaging targets detection and tracking against bright background

摘要

Aerial target under the daytime condition shows complex imaging characteristics, like misshapen and blurred edges, even flickering and discontinuous imaging, which becomes a bottleneck of target detection and tracking in imaging tracking system. This paper presents a novel algorithm for detecting and tracking aerial targets with blurred and flickering imagery. Firstly, point spread function (PSF) is determined combining with the feedback information of tracking process and we use a new model of image restoration proposed by Ben-Ezra to deblur the moving target. Secondly, association network model with linking mechanism is established in terms of their similarity based on accompanying values of the target features to handle the situation of flickering imaging and feature losing for target. Simulations are conducted using the proposed method for image sequences, which can obtain the good results.