Research on the Improvement of Wheel Target Detection Algorithm for Anti Lifting of Container Trucks
In order to prevent safety hazards caused by container trucks being lifted together with containers during unloading operations,it is necessary to accurately identify the position and status of the container trucks during the operation.The feature extraction object for target detection is the collection truck wheel,and the YOLOv5s algorithm in neural networks is used as the basic algorithm.To address the shortcomings of insufficient feature extraction ability,weak feature connection,and limited scene application ability in the algorithm,three improvement measures are proposed:fusion structure heavy parameters,adaptive weighted pyramid,and detection head pruning.The experimental results indicate that improving the algorithm can enhance the object detection performance of the model.
万方数据
维普
