Application of Improved YOLOv5 in Building Materials Warehouse Rat Detection

As an important part of the civil engineering field, the rodent problem in the environment of building material warehouses not only threatens the safety of the stored materials, but also may affect the quality and service life of the building materials. In order to solve the complex challenges in the detection of rodent infestation in building material warehouses, this paper proposes an intelligent detection method based on the improved YOLOv5. The building materials warehouse environment is characterised by dense objects, complex lighting and frequent shading, which makes the rodent pest detection task more difficult. In this paper, two improvements are made on the basis of YOLOv5: first, a local perception enhancement module is introduced to improve the model's ability to capture local features, especially in the case where the rodent target is occluded or hidden in the accumulation of building materials, which significantly improves the detection accuracy; second, an inter-frame feature fusion mechanism based on the temporal information is used to enhance the detection robustness in the dynamic scene by utilising the temporal information of consecutive frames of video sequences, and reduce the leakage. The second is to use the inter-frame feature fusion mechanism based on temporal information to enhance the robustness of detection in dynamic scenes by utilising the temporal information of consecutive frames in the video sequence, and to reduce the occurrence of leakage and false detection. The experimental results show that the improved YOLOv5 has excellent performance in the detection of rodent infestation in the building materials warehouse environment, with significant improvement in precision and recall, which provides an efficient and reliable technical solution for the intelligent management of building materials warehouse.