Research On Moving Target Tracking System Based On Maching Vision

Machine vision technology is a hot research topic in recent years,it is to use machine to replace the human eye to detect and identify the target.As the core issue in the field of machine vision,target tracking has been widely used in military navigation,intelligent traffic control,industrial testing,medical science diagnosis and public life safety and other fields.Although many effective target tracking algorithms have been proposed so far,these algorithms basically have certain pertinence,that is for a certain aspect(high-speed movement,occlusion,target deformation,etc.)put forward the corresponding solution,and most of the target tracking algorithms have high complexity,low real-time performance,prone to track loss,jitter and other phenomena.So we need to solve a lot of problems to develop a truly robust and practical tracking system.In this paper,an embedded mobile target tracking system based on machine vision is built with intelligent baby carriage as the research platform,and the applied tracking algorithm is deeply researched and analyzed.The main contents and achievements of this paper are as follows:(1)This paper analyzes the TLD target tracking algorithm.The algorithm combines the traditional tracking algorithm and the detection algorithm.When the target disappears,it has good re-detection ability,and through an online learning mechanism to correct error about the results of tracking and testing,to improve the performance of classifier continuously,the final output more accurate tracking target box,so the algorithm has good robustness.However,through the detailed analysis of each module in the TLD algorithm,the tracking module is sensitive to the interference of light intensity and occlusion,which will affect the tracking performance of the algorithm;The detection module needs to scan the whole video frame,and the scanning window needs to pass through the variance classifier,the set classifier and the nearest neighbor classifier to detect the position of the target to be tracked,which makes the real-time performance of the algorithm decline.(2)Aiming at the low real-time performance of TLD target tracking algorithm,this paper proposes to improve the TLD tracking algorithm using the detection module based on the compression sensing theory to replace the cascade classifier in the TLD tracking algorithm to improve the real-time performance of the algorithm.This paper analyzes the basic principles of compression sensing theory in detail,and through detailed analysis of using the sparse measurement matrix in the compression perception theory to extract the Haar-like feature of the image and using the naive Bayesian classifier to classify the target and background,It is proved that the detection module based on the compression sensing theory can extract the target feature of the image efficiently and greatly improve the detection speed of the image,which makes it possible to improve the real-time performance of the algorithm while ensuring the accuracy and robustness of the tracking result after using the compression-aware detection module for the TLD tracking algorithm.(3)An embedded mobile target tracking system based on machine vision is designed with intelligent baby carriage as the research platform,and the improved algorithm of TLD is applied to the system to realize the real-time tracking of the moving target of the baby carriage.Finally,the experimental results of the improved algorithm are compared with the experimental results of the original algorithm.The results show that the improved tracking algorithm has better real-time performance and higher robustness.