Visual target tracking method of bionic retina

Abstract

The invention discloses a visual target tracking method of a bionic retina, and belongs to the field of photoelectric target tracking. According to the invention, the variable resolution characteristic of bionic retina imaging is mainly utilized, and the field of view is mapped to the logarithm polar coordinates for target tracking. Firstly, logarithm polar coordinate mapping imaging of a bionic retina is utilized, a target tracking algorithm is improved through the staring effect of a fovea centralis, the information compression effect of an edge and real-time parameters, redundant data are compressed, detail information of an interested target is enhanced, and the information acquisition efficiency in a view field range is improved. And meanwhile, target information in a scene is accurately grasped by utilizing singular points and imaging threshold judgment, and the reliability of target screening is improved. And finally, real-time target tracking is realized by comparing an eccentricity judgment algorithm with higher speed with the motion control module. Compared with a traditional method, under the condition of a large field of view and high resolution, real-time target tracking is achieved on the basis of a bionic retina vision mechanism, efficiency is higher, and precision is higher.

Description

technical field [0001] The invention belongs to the field of photoelectric target tracking, in particular to a bionic retina visual target tracking method. Background technique [0002] Target recognition and tracking is a common technology. Its core is to locate and track specific or interested objects in real time. At present, this technology has been widely used in many occasions, such as: the field of robotics. Visual perception is one of many applications of robots. The prerequisites for (navigation, obstacle avoidance, etc.) usually require a large field of view, high resolution, and real-time requirements for visual perception, which imposes strict requirements on traditional visual perception. On the other hand, when a robot performs target recognition, the target size is relatively small under the condition of a large field of view. Therefore, the background other than the target becomes redundant data, which reduces real-time performance. Object tracking has becom...

AI Technical Summary

Problems solved by technology

[0003] At present, the typical target tracking is to detect the movement in the video sequence or a single object with certain characteristics through an algorithm, obtain its coordinates, speed and other parameters, and then track it. Therefore, the real-time performance is not good, especially under the conditions of shooting with a fixed-resolution camera.

[0004] The above shows that although the traditional target tracking method can achieve tracking to a certain extent, there are problems of poor real-time performance and poor reliability.

Images