Detection Method for Transport State of C80B Coal Open Wagons Based on Improved YOLOX Algorithm

When transporting coal under the effect of various factors, open wagons often encounter issues, such as red ore, coal residues, foreign objects, and frozen and snowy bottoms, which increase consumption and costs.The inspection of openwagon transport status mostly relies on manual monitoring, which has low automation and is labour intensive and inefficient.This study proposed a You only look once X (YOLOX) model based on a parameter-free attention residual structure and lightweight adaptive feature fusion pyramid to improve the accuracy of the detection of the state of openvehicle transport.A parameter-free attention residual structure (SARM) and self-adaptive spatial feature fusion module (SASFF) modules were introduced into the neck network of YOLOX.The improved YOLOX model was built by using a highly perceptive detection structure (DCBS) module proposed in this study in the detection process, and the accuracy of the model was verified by experiments.Subsequently, a high-quality open-truck-transportation image dataset was constructed for the open-truck transportation target detection task and accurately labelled.Results demonstrate that, in experiments on the Common Objects in Context 2017 dataset, the improved YOLOX algorithm shows improvements of 11.9%, 4.4%, and 1.9% over the YOLOX, YOLO-V5, and YOLO-V8 algorithms, respectively, and the proposed algorithm exhibits considerable improvements in missed and wrong detections.The proposed method provides the rapid detection of the transport status of open wagons for subsequent automated reporting and logging.