Underwater Target Detection Based on Improved YOLOv4

As one of the underwater image processing methods, the underwater target detection algorithm is becoming increasingly important in daily life. An underwater target detection algorithm is proposed in this paper. The difference between this algorithm and the existing algorithms is that it is based on improved YOLOv4. K-means++ algorithm uses re-clustering method to obtain anchor boxes in underwater scene. Image enhancement is also performed using MSRCR with dark channel fusion, and the dataset is trained using a segmented training method. The experimental results show that the detection effect of this method is better than several other methods.