Research On Fast Detection And Tracking Of Specific Environment Target Based On Binocular Camera

At present,the detection and tracking of moving targets at close range mainly use traditional infrared and radar technologies,which cannot meet the technical requirements of higher requirements such as reviewing the background environment and simultaneous processing of target detection and tracking.The current image processing technology based on binocular camera technology can provide better technical solutions when dealing with the above application problems,but it also has limitations such as large data processing volume and the inability of the hardware platform to meet the computing needs.In order to introduce the image processing technology based on binocular cameras into the detection and tracking applications of moving targets,this article mainly focuses on how to reduce the image processing data as much as possible in the field of ranging applications,while ensuring accuracy and accuracy.The research is carried out on the scale of measurement and algorithm calculation,in order to realize the real-time binocular camera slow-moving target ranging in a specific simplified application scenario on a lower-cost hardware platform.In order to achieve the above goals,this paper first set up a test environment with a simple background,and set up a binocular camera based on a lower-cost computing platform,and designed a research technology route from a simple environment to a complex environment,and from a static camera to a dynamic camera.Carry out fast ranging research for slow targets,with the goal of reducing computing requirements and optimizing ranging methods,so that the application of binocular camera technology in special scenarios can be realized.The core content and main innovations of this paper include the following points:(1)Use the existing image processing technology based on binocular cameras to conduct technical verification for the special environment set up in this article.Based on the verification results,on the premise of increasing the calculation speed and reducing the amount of calculation,the research environment is designed pertinently And parameter settings provide reasonable preconditions for research.(2)Aiming at the set application scenarios,on the premise of improving the calculation efficiency and ensuring the accuracy of the ranging,a stereo ranging algorithm based on the improved Census transform is proposed.The research of the algorithm mainly includes optimizing the cost calculation of the algorithm;improving the support window of the calculation;reducing the cost aggregation path and so on.(3)Research on image matching based on the images produced by binocular cameras,mainly focusing on the refinement of the parallax matching area.On the premise of improving efficiency,reducing the amount of calculation,and ensuring accuracy,the global matching is replaced by regional matching,and block matching is used instead of full-frame matching,and the research results in the application scenarios set in this article are verified.