Binocular Vision-Based Minimum Distance Determination Between Persons and Warning Lines

Measurement of minimum safe distance between warning line and target person has been one of challenges in safety monitoring in hazardous areas. To address this problem, a method based on binocular vision was proposed in this paper for minimum distance determination. Effective feature points of the warning line and target person were first screened through feature point matching, with 3D coordinates acquired via triangulation. An infinite plane was subsequently generated by planar fitting of warning line feature points. The 3D coordinates of the target person were then projected onto this plane, converting the 3D measurement to the 2D planar distance calculation to achieve high-precision determination of minimum safe distance. Experimental results demonstrated a minimum average error of 3.26% and a maximum average error of 5.60%, indicating that the proposed method has the potential to be applied in safety monitoring, transportation, manufacturing, and other complex environments.