Arbitrary direction remote sensing image ship detection method and system based on rotation candidate box

Abstract

The invention provides an arbitrary direction remote sensing image ship detection method and a system based on a rotation candidate box, and the method comprises the steps: firstly obtaining a remotesensing image training data set, carrying out the preprocessing of all remote sensing images in the obtained training data set, and obtaining a preprocessed training data set; then, establishing a first Faster-RCNN (Recurrent Convolutional Neural Network); taking each remote sensing image in the preprocessed training data set as training data of a first Faster-RCNN network, and training the firstFaster-RCNN network by adopting a back propagation algorithm to obtain a trained first Faster-RCNN network; and finally, utilizing the trained first Faster-RCNN network to detect the ship in the actual remote sensing image. The beneficial effects of the invention are that the technical scheme provided by the invention avoids the manual design of features; a network based on a detection box in anydirection is provided; meanwhile, for dense objects, the detection effect is better than that of a traditional horizontal frame detection network.

Description

technical field [0001] The invention relates to the technical field of remote sensing image processing and information extraction, in particular to a method and system for detecting a ship in an arbitrary direction remote sensing image based on a rotating candidate frame. Background technique [0002] The method of object detection in remote sensing image based on manual extraction not only consumes a lot of manpower, but also consumes a lot of time, and the work efficiency is low. In addition, the interactive surface object detection method is greatly affected by human factors, and the accuracy is generally not high. [0003] In recent years, with the continuous development of deep learning technology, remarkable achievements have been made in many fields. Among them, deep learning technology performs target detection in pictures of natural scenes, which has achieved achievements that are difficult to achieve with traditional methods. The traditional ship detection method...

AI Technical Summary

Problems solved by technology

The traditional ship detection method uses the classic horizontal frame detection network, and needs to manually design features, the recognition accuracy is not high, and the efficiency is low. Therefore, it is necessary to use the method of deep learning to detect the ship target in the remote sensing image , so as to realize the automation and intelligent processing of remote sensing surface object detection tasks, and realize fully automatic remote sensing image surface object detection

Images