A Deep Learning-Based Point Cloud Target Recognition Method for Inland Waterway Transmission Line Vessel Overheight Overruns

Conventional inland waterway ship over-height limit recognition methods are based on laser rangefinder measurements, and the recognition resultsa are biased by the influence of the marine environment. Therefore, a deep learning-based point cloud target recognition method is designed to recognize the over-height limit of ships in inland waterways. The advanced laser scanning technology is used to obtain the three-dimensional point cloud data of inland waterways, and the contour of the point cloud is extracted by meshing and boundary delineation. Based on deep learning, we constructed the target recognition model of ship overrun point cloud, selected convolutional neural network to process the point cloud data, and learned the features of ship overrun point cloud contour to avoid the problem of recognition error. Aligning the point cloud target with the point cloud contour data of the ship in the waterway, the point cloud target of the ship identified by the model and the point cloud contour data are aligned to ensure the accuracy and reliability of the identification. Using the example analysis, it is verified that this method has higher recognition accuracy and can be applied in real life.