YOLOv5 Garbage Target Detection Based on Lightweight Convolution in Marine Environment

In the Marine environment, in view of the great threat to the ecological environment and Marine life caused by underwater environment, it is necessary to develop effective detection methods to realize the automatic garbage target detection. This paper proposes an improved underwater garbage detection algorithm based on YOLOv5. Aiming at the problems such as uneven illumination, blurred image and noise interference in underwater environment. The original model in this study is optimized and replaced with a lightweight GSConv convolution, thereby reducing the computational power required by the model. This modification enables the model to achieve an excellent balance between accuracy and speed, effectively addressing the challenges of detecting complex backgrounds or small targets. Moreover, extensive testing of our improved method is conducted using the publicly available real underwater environmental garbage dataset TrashCan1.0. The final experimental results demonstrate that compared to the original YOLOv5 model, our enhanced algorithm exhibits higher detection accuracy and robustness in underwater garbage detection tasks. Consequently, this algorithmic research holds significant potential for installing and deploying detectors on autonomous underwater robots, thus playing a crucial role in marine ecological preservation and environmental safety.