The problem of scene matching for video satellite remote sensing images is studied. The principle of image classification based on deep convolutional neural network is analyzed. It is pointed out that deep convolutional neural network can be used to extract the feature vector of an image. Based on this, a scene matching method based on deep convolutional neural network is proposed, which uses the feature vector extracted by deep convolutional neural network to match, and has good robustness to illumination, scale and viewing angle differences. The algorithm was verified by images shot by video satellites.
Scene Matching Method of Video Satellite Remote Sensing Image based on Deep Convolutional Neural Network
01.10.2019
309273 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch