Application Research Of Remote Sensing Image Registration Based On Deep Residual Network In UAV System

Remote sensing image registration technology is an important link in the field of image processing,it has a wide range of applications in remote sensing system platforms such as drones,artificial satellites,and spacecraft.Because of the difficulty of remote sensing image registration,it limits its application and development to a certain extent.In order to meet the high precision requirements of remote sensing images in various fields,this paper proposes a remote sensing image registration method based on multi-layer feature fusion of deep residual network,and a remote sensing image registration method based on spatial transformation and dense convolution fusion.1.This paper proposes a remote sensing image registration method based on multi-layer feature fusion of deep residual network.The deep residual network Res Net50 is used to extract the features of the remote sensing image,and the features of multiple convolutional layers are weighted and fused to enhance the robustness of the feature points.Random sampling consistency algorithm is used to eliminate mismatched points,improve the matching accuracy of feature points,and complete the remote sensing image registration according to the calculated transformation model.2.This paper proposes a remote sensing image registration method based on spatial transformation and dense convolution fusion.The STN-Dense Net model with spatial transformation structure is used for feature extraction to enhance the feature extraction effect of the image deformation area.The improved grid-based motion statistical algorithm is used to match the features extracted by the model,and the method of homography is used to eliminate the mismatched point pairs to improve the matching accuracy,thereby achieving accurate registration of remote sensing images.Experiments show that the method in this paper can effectively increase the number of correct matching points,and has higher registration accuracy and stronger robustness.It can provide accurate remote sensing image registration results for UAV systems,and has good performance in practical applications such as urban planning,disaster monitoring,and farmland protection.