Truck Compartment detection algorithm based on 3D point cloud for automated loading

With the development of the economy, automated loading is becoming a trend in the transportation industry. Carriage detection is a key technology for automated loading. To meet the requirements of automated loading, we propose a carriage detection solution based on LiDAR-generated 3D point clouds. Our solution consists of three main steps. Firstly, we extract the carriage floor plane to provide a basis for projecting the fence point clouds. Secondly, we combine image-based detection with 3D detection to achieve preliminary carriage detection. Finally, we detect the carriage by extracting the fence height. Experimental results show that our method has an error rate below 2% and a processing time of less than 2s. This indicates that our solution successfully performs detection of the entire carriage, meeting the requirements of automated loading.