基于改进YOLOv8s的矿用输送带异物检测方法

打开文本图片集

中图分类号：TD528.1 文献标志码：A

Abstract： Inlow-illumination mine environments，conveyor belt foreign object detection algorithms suffer frominsufficient extractionof global image features andan excessive numberof model parameters.A method for detecting foreign objects on mine conveyor belts based on an improved version of YOLOv8swas proposed. YOLOv8s was improved using VMamba and MobileNetv4：MobileNetv4 was employed to enhance the backbone network by integrating the Universal Inverted Bottleneck （UIB） module.The efcient inverted residual structure reduced the overallnumber of model parameters，and a dynamic feature adaptation mechanism was used to strengthen feature robustness in smal-object scenarios.The core feature extraction and fusion module C2f was improved by VMamba's Visual State Space （VSS）module，which eficiently captured global contextual informationin images througha state space model and four-directional scanning mechanism，enhancing the model'sunderstanding of global imagestructure.Aparameter-sharing lightweight detection head was designed， using Group Normalization （GN）as the basic convolutional normalization block to compensate for accuracy loss caused by model lightweighting.Experimental results showed that the improved YOLOv8s model achieved an mAP@0.5 of 0.921and an mAP @ 0.5：0.95of 0.6o1 on aself-built dataset，reduced the number of parameters by

27.7% compared to original YOLOv8s， outperformed mainstream object detection models such as YOLOv11s and YOLOv10s， and met the requirements for foreign object detection on mine conveyor belts.

Key words： conveyor belt foreign object detection; YOLOv8s; VMamba; MobileNetv4; lightweight; group normalization

0引言

煤矿井下高负荷运行的输送带常处于昏暗环境，易混入锚杆、矸石等异物。（剩余11137字）