Research on Measuring Vehicle Outline Dimensions by Fusing Vision and Laser Point Clouds

The modification and overloading of vehicles have been persistently prohibited in highway transportation, becoming a significant factor affecting traffic safety. It severely disrupts public transportation order and directly threatens the safety of people's lives and properties. To address the low accuracy of measuring vehicle outline dimensions and the difficulty in segmenting vehicle components using purely image or point cloud data, a method is proposed that integrates laser point clouds and vision for measuring car outlines and components. The approach involves mapping the 3D point cloud model of the vehicle collected by the laser radar to a 2D plane, identifying and extracting feature points from the 2D projected image, matching the color information of the vehicle captured by the coaxial camera with the corresponding laser point cloud to merge the measured vehicle's laser point cloud model with the camera-captured image for 3D model reconstruction. This enables the identification, segmentation, and measurement of vehicle and external features.