Wildlife detection algorithm based on Inv-YOLOv5m

Wildlife target detection based on infrared camera images is of great significance for biological protection. At present, with the gradual popularization of infrared camera detection, real-time wildlife detection has been widely studied. A series of target detection models have emerged, but the detection accuracy is always not high. In order to realize the effective detection of wildlife, a wild animal detection method based on improved YOLOv5m is proposed in this paper. First, the Focus module in the original feature extraction network is replaced by CBL module. Secondly, the self-attention mechanism is embedded in the feature extraction module to improve the feature extraction ability of the detection network. Finally, the involution operation is introduced into the feature fusion module to improve the detection ability of each feature scale. The experimental results show that the enhanced YOLOv5m wildlife detection model in this paper has greatly improved the accuracy, recall rate and map, and provides a new technical means for wildlife detection.