Remote sensing technology for detecting ships at sea can be widely used in maritime supervision, military reconnaissance, and ship rescue. It has very important application value. In this paper, we propose the YOLOv3-RS algorithm based on YOLOv3 algorithm to precisely detect small ships in remote sensing images. The YOLOv3-RS algorithm uses the K-medians algorithm to improve the clustering effect of the data set, and then uses DenseNet to increase the feature reuse rate, and finally improves the ResNet module to enhance the expressive ability of the network. The final results show that YOLOv3-RS algorithm has better detection performance than YOLOv3 algorithm.
Detection of Small Ships in Remote Sensing Images based on Deep Convolutional Neural Network
2020-10-05
754089 byte
Conference paper
Electronic Resource
English
Ship object detection in remote sensing images using convolutional neural networks
British Library Online Contents | 2017
|