DBS-YOLO: A High-Precision Object Detection Algorithm for Hazardous Waste Images

The adverse effects of hazardous waste on the natural environment and human health are evident. In real-life scenarios, the dense distribution, mutual occlusion, and complex image backgrounds of hazardous waste present significant challenges for existing object detection algorithms to achieve a high detection accuracy. In order to enhance the efficiency of hazardous waste sorting, this article proposes an improved detection model, DBS-YOLO, based on the YOLOv8n network. This model strikes a balance between accuracy and lightweight design. First, we replace part of the convolution modules in the C2f module with deformable convolutional networks version 3 (DCNv3) modules, proposing the DC2f module. Incorporating this module into YOLOv8n not only achieves a lightweight network design but also enhances the model's adaptability to occluded hazardous waste. In addition, we introduce a bi-level routing attention (BRA) module to construct a global receptive field, capturing long-range dependencies and improving the model's focus on targets. Finally, we employ a soft non-maximum suppression (Soft-NMS) algorithm to reduce instances of false positives (FPs) and false negatives (FNs) caused by mutual occlusion, further enhancing model accuracy. Experimental results demonstrate that compared to the original YOLOv8n network, DBS-YOLO achieves improvements of 2.0% in precision, 1.0% in recall, 2.1% in mean average precision (mAP)@0.5, and 4.9% in mAP@0.5:0.95, reaching a mAP@0.5 of 95.4%. The model also mitigates instances of FPs and FNs. Comparative analyses with SDD, Faster R-CNN, and YOLO series networks confirm the effectiveness of the proposed model.