Models for Determining Distances to Objects by the Technical Vision System
摘要
This article proposes a technology for determining the distance from a single video camera to recognized objects. YOLO neural network is employed for object recognition and classification. Subsequently, the contours of recognized objects are delineated, and the number of pixels within these contours is counted. As an object moves, both the contour volume and pixel count increase. This information is utilized to formulate a regression equation, which accurately estimates the distance to the targeted object.