A Research On Long-term Target Tracking Method Based On Hybrid Swarm Optimization

Visual tracking is a fundamental problem in computer vision due to its important role in a wide range of applications such as video surveillance,autonomous vehicle navigation,human-computer interaction,medical imaging,and many more.Although it has attracted increasing interests in recent decades and significant progress have been achieved,visual tracking remains a challenging task due to various factors such as partial occlusion,motion blur,illumination variation,abrupt motion,to name a few.Therefore,there are still many key technical problems to be solved how to design a long-term tracking algorithm under complex tracking conditions.In this dissertation,under the research of long-term target tracking,The main research is to improve or mix the swarm optimization algorithm,so as to enhance the convergence performance.The main research work and innovations are described as follows:(1)An object tracking method based on hybrid extended ant lion optimizer with sine cosine algorithm(EALO-SCA)is proposed.Firstly,in the standard ant lion optimizer(ALO),multiple elites are used instead of a single elite to improve the extremum information.Extend ant lion optimizer(EALO)algorithm to improve global search ability.Secondly,considering that the sine cosine algorithm(SCA)has a strong local search operator,a hybrid EALO-SCA tracker is proposed combining the advantages of both EALO and SCA.EALO-SCA tracker can improve tracking accuracy and efficiency.Finally,extensive experimental results in both quantitative and qualitative measures prove that the proposed algorithm is very competitive compared to 7 state-of-the-art trackers,especially for abrupt motion tracking.(2)An object tracking method based on hybrid teaching-learning-based optimization with adaptive grasshopper optimization algorithm(TLGOA)is proposed.First,the non-linear strategy of tangent function is used to replace the linear mechanism of the standard grasshopper optimization algorithm(GOA).The adaptive grasshopper optimization algorithm(AGOA)can avoid the local trapping problem,improve the ability of global optimization,and AGOA can deal with the problem of abrupt motion.Secondly,considering that the algorithm of teaching-learning-based optimization(TLBO)has obvious local exploitation operator and fast convergence speed,AGOA and tlbo are mixed in series to form a hybrid TLGOA,and design a tracking framework based on TLGOA is designed.Finally,a large number of experimental results show that TLGOA has obvious advantages over other optimization algorithms,and prove that TLGOA tracker has strong competitiveness compared with other state-of-the-art trackers.