Research on lightweight Yolo coal gangue detection algorithm based on resnet18 backbone feature network

• Published in: Internet of things (Amsterdam. Online) Vol. 22; p. 100762
• Main Authors: Xue, Guanghui, Li, Sanxi, Hou, Peng, Gao, Song, Tan, Renjie
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.07.2023
• Subjects: Coal and gangue sorting robot; Coal gangue detection; Coal preparation; Lightweight network; Model compression; Unstructured pruning; You only look once (YOLO); Coal gangue detection; Coal and gangue sorting robot; Coal preparation; Lightweight network; You only look once (YOLO); Unstructured pruning; Model compression
• ISSN: 2542-6605; 2543-1536; 2542-6605
• DOI: 10.1016/j.iot.2023.100762

•Presents a coal gangue detection algorithm based on the improved YOLO algorithm with the backbone feature extraction network of ResNet18, and discusses its feature scale clip and unstructured pruning to enhance its lightweight and real-time.•Prepares coal gangue data sets under different conditions with consideration of the actual working condition of the coal gangue presorting.•Compared with other lightweight networks, ResNet18-YOLO coal gangue detection algorithm with ResNet18 has the better performance.•Compared with other deep learning algorithm, the ResNet18-YOLO coal gangue detection algorithm gives good consideration to both accuracy and real-time, and has better detection effect. Nowadays coal gangue presorting is still mostly conducted manually, with high labor intensity, low sorting efficiency, and potential safety hazards. A coal gangue sorting robot that is an effective device to complete the coal gangue presorting instead of the worker, for which the coal gangue intelligent detection is one of its key technologies. The lightweight and real-time performance of the detection model has an important impact on the performance of coal gangue sorting robot. Lots of models for the coal gangue detection have been proposed based on machine learning, but there are still some problems such as slow detection speed, structural redundancy and large model size. This paper adopts the lightweight network of ResNet18 as the backbone feature extraction network of YOLOv3 and proposes a coal gangue detection algorithm of ResNet18-YOLO, studied the feature scale reduction and unstructured pruning of the model to further improve its lightweight and real-time performance, prepared the coal gangue data set referring to the actual gangue sorting situation and discussed the performance of the model. The results show that the improved ResNet18-YOLO algorithm can detect the coal gangue at a speed of 45.5 ms/piece and the model size is only approximately 65.34 MB when the mAP of the coal gangue detection is 96.27%. It has better real-time performance and smaller model size with the condition of equivalent mAP performance to the other algorithms, and has good gangue detection performance, which is conducive to reducing the technical requirements for the gangue sorting robot.