A fast moving target tracking method based on hybrid goa and tlbo search

Abstract

The invention provides a rapid moving target tracking method for GOA and TLBO mixed searching, which is used for solving the problems of low tracking precision and low efficiency when a traditional optimization algorithm is applied to video target tracking. The method comprises the steps that target state parameters and optimization model parameters are initialized; Searching a candidate region similar to the target by adopting the global situation of GOA; Searching a candidate image block most similar to the target image by using the local part of the TLBO; Calculating a similarity value of the candidate image block and the target image block; Selecting the optimal candidate image block as an iteration target image block and a target state parameter of a next frame to track the next frameof image; And repeating the above steps to realize video target tracking. According to the invention, the characteristics of strong GOA global exploration capability and strong TLBO local developmentcapability are combined, the continuous tracking of the inter-frame rapid or abrupt change motion can be effectively realized, the adaptability in a complex scene is improved, and the method has important theoretical significance and practical value.

Description

technical field [0001] The present invention relates to the technical field of target tracking, in particular to a fast moving target tracking method of hybrid GOA and TLBO search, which realizes the continuous tracking of targets, especially the sudden movement of targets due to fast motion between adjacent image frames Under the circumstances, the continuous tracking ability of the target can still be maintained. Background technique [0002] As an important branch of computer vision, target tracking refers to the process of continuously inferring the target state in the video sequence. Its task is to generate the trajectory of the target by locating the target in each frame of the video, and Provides complete target area in one moment. Video object tracking technology has been used in a variety of fields, such as: intelligent transportation systems, intelligent video surveillance, human-computer interaction, medical imaging, unmanned aerial vehicle and precision guidance...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the technical problems of low tracking accuracy and low efficiency in the application of traditional optimization algorithms in video target tracking, the present invention proposes a fast moving target tracking method that combines GOA and TLBO search, and uses the GOA (Grasshopper Optimization Algorithm) search strategy for global exploration and search The area where the target may appear; further develop the target position through TLBO (teaching and learning optimization algorithm) local search, improve the tracking accuracy and efficiency of the algorithm, and finally realize the smooth and fast moving video target tracking method at the same time

Images