Vehicle speed measurement method based on binocular vision

• Published in: Ji suan ji ke xue Vol. 48; no. 9; pp. 135 - 139
• Main Authors: Chang, Zi Ting, Shi, Yuqing, Wang, Jun, Yu, Minghe, Yao, Lan, Zhao, Zhibin
• Format: Journal Article
• Language: Chinese
• Published: Chongqing Guojia Kexue Jishu Bu 01.09.2021; Editorial office of Computer Science
• Subjects: Algorithms; Binocular vision; Displacement measurement; Frames (data processing); Highways; Measurement methods; Template matching; Traffic speed; Trucks; vehicle speed measurement|alignment on region matching|binocular vision|template matching; Velocity measurement; Video data; Weighing
• ISSN: 1002-137X
• DOI: 10.11896/jsjkx.201000047

In order to cooperate with the weighing of trucks at the entrance of the expressway, it is necessary to measure the real-time speed of the vehicles when they pass the weighing platform. Speed ​​measurement using binocular vision technology has the advantages of low cost, simple deployment and high stability, and has broad application prospects. The technical difficulty of binocular visual velocity measurement is the displacement measurement of the target. The core problem is the precise matching of the target between multiple frames of images. This paper proposes a matching region alignment algorithm based on spatial location and a spatial displacement calculation method based on template matching. Specifically, using the spatial position relationship of the wheel to limit the matching area of ​​the wheel can effectively reduce the mismatch problem of similar wheels; use the template matching method to track the key points of the wheel, and then obtain the spatial displacement of the wheel between multiple fr