A review of several key techniques of target vision tracking in complex scenes

Abstract

Target vision tracking is a hot research issue in the field of machine vision. It has been studied by many countries and widely used in many fields. Its main task is to detect, identify and locate moving targets in the video sequence, and obtain the paramaters such as position, velocity, acceleration, direction, motion trajectory of the target, in order to provide the basis for further intelligent video analysis. Due to the complexity and unpredictability of the target itself and the external environment, robust real-time target tracking in complex scenes has become a key issue to be solved urgently. This paper aims to summarize key techniques such as motion detection, target classification and target tracking involved in video target tracking. And then the paper introduces the algorithms commonly used in several key techniques, compares and analyzes the advantages and disadvantages of various algorithms. Besides, the paper also summarizes the current research status and research issues, and propects the future research trends.