Parallel Processing Of Remote Sensing Image Filtering Algorithm Based On CUDA

Remote sensing image filtering is a key step in remote sensing image preprocessing process,which can effectively suppress the noise influence in remote sensing image and provide a good data base for the subsequent processing and application of remote sensing image.In recent years,with the continuous development of remote sensing technology,the amount of remote sensing data is increasing rapidly.On the other hand,the demands for real-time processing is also increasingly urgent,the traditional remote sensing image filtering algorithm is based on the CPU serial algorithm,which cannot meet the above demands.Therefore,how to improve the remote sensing image filtering processing speed is an urgent problem to be solved.On the basis of studying the characteristics of CUDA parallel programming technology and several typical remote sensing image filtering algorithms,the parallelism of mean filter,median filter and Gaussian filtering algorithm is analyzed,and the parallel algorithms is designed.The processing efficiency advantage of the designed algorithms is verified by comparing with the traditional CPU serial filter algorithms.The main work and innovation of the paper are as follows:(1)It described the research status of remote sensing image filtering algorithms,image processing algorithms based on GPU at home and abroad,and introduced the technical characteristics of CUDA parallel programming architecture.Systematically described the principle of mean filter,median filter and gaussian filter algorithm.And analyzed the denoising characteristics of these algorithms through experiments.(2)Analyzed the parallelism of the traditional remote sensing image filtering algorithms,and designed the parallel algorithms.The validity of the parallel algorithms based on CUDA is verified by comparing with the original serial filtering algorithms.(3)Taking the mean filter algorithm as an example,it studied the performance of the parallel acceleration algorithm.And verified the effectiveness of CUDA algorithm performance optimization strategy by comparing with the GPU image filtering algorithm provided by OpenCV.