Marine Target Detection Based on Improved Faster R-CNN for Navigation Radar PPI Images

Marine target detection is always an important issue in the field of target detection, and scholars at home and abroad have achieved fruitful research success, but there are still some problems while using traditional detection method such as Large interference of sea clutter on target detection and limited detection performance. Recently, deep learning has been rapidly developed and widely applied in the field of target detection, However, there is currently no related research on the application of deep learning to navigation radar. In this paper, we propose a marine target detection method based on improved Faster R-CNN for navigation radar PPI (Plane Position Indicator) images. We built our own maritime target dataset by collecting radar measured data under different conditions, and improved the Faster R-CNN target detection method in four aspects. The maritime target detection model was obtained through training and optimization. The experimental results proved that our marine target detection method based on Faster R-CNN shows better detection performance in accuracy and reliability compared with the traditional Faster R-CNN method.