Research On Key Algorithms Of Remote Sensing Digital Image Processing

With the rapid development of remote sensing images,aerial remote sensing has been widely used because of its low cost and high flexibility.Because of the limited flight height,aerial remote sensing images often need to pass through clouds during the shooting process,which leads to the noise of fog,haze and other noise in the aerial remote sensing images,which reduces the quality of aerial remote sensing images.At the same time,in order to get a high-resolution panoramic image that contains all the information of the area,it is necessary to use image stitching to synthesize several successive aerial remote sensing digital images into one image.1.An image defrosting algorithm based on deep learning is presented.First,the original atmospheric scattering model is distorted so that many unknown parameters are unified into one unknown parameter,and then the unknown parameters are estimated by using multiscale convolution neural network.Finally,the parameter estimates are brought into the defrosting model to get the restored fog-free image.2.A new fast stitching algorithm for UAV aerial images is presented.The improved A-KAZE algorithm is used to extract and describe the feature points of aerial remote sensing images.The feature descriptors are simplified and reduced by principal component analysis.Finally,the images to be stitched are stitched into a panorama,and the stitching seams are balanced by a fusion algorithm.3.Design a wetland tour platform of Panjin red Beach,Liaoning,based on WeChat applet.The mosaic of the defrosted and stitched remote sensing image with the Golden Map API will lay a foundation for the subsequent industrialization development of remote sensing image.Compared with the related algorithms,the proposed algorithm performs better in the visual effect and objective indicators of remotely sensed digital image defogging,and effectively improves the clarity of remotely sensed image under haze weather conditions.In the aspect of remote sensing image stitching,the proposed algorithm has no obvious stitching seams,no color balance and strong real-time performance compared with the improved algorithm.