Font Size: a A A

Research And Application Of Remote Sensing Image Clustering Based On The Improved Fuzzy C-means Algorithm

Posted on:2018-12-24Degree:MasterType:Thesis
Country:ChinaCandidate:J K ZhangFull Text:PDF
GTID:2348330518498248Subject:Electronics and Communications Engineering
Abstract/Summary:PDF Full Text Request
In recent years, as the number of satellite increases and the resolution of remote sensing image improves constantly, our country has obtained a large amount of satellite image data. The traditional method that deals with remote sensing images through manual way has many disadvantages, such as more work?low efficiency?poor stability and it is bad that the remote sensing images can't be extracted and used better. Due to the technology about intelligent image processing has some advantages in the aspect of processing speed and stability, it has received extensive attention and development in the field of artificial intelligence.The paper, with the meteorological remote sensing image as the research object,has carried out research and application about remote sensing image clustering, the main work is as follows:1 Proposes a multi-scale exponential fuzzy c-means (FMEFFCM) algorithm based on feedback information. In the remote sensing image, the gray interval length of different clouds is differ. This paper proposes a multi-scale exponential function fuzzy c-means (FMEFFCM) algorithm based on feedback information via the membership degree. The algorithm determines the feedback factor through membership degree, then the factor changes the clustering scope of related classes,which can make the fuzzy clustering algorithm cluster accuracy higher for remote sensing image.2 Since each class of the clustering result contains large amounts of data object of not clouds, according to the physical information and distribution characteristics of clouds, the paper puts forward a density classification algorithm about clouds based on Radon transform. Firstly the algorithm determines the relative center of cloud by Radon transform; then calculating density of the data object; finally to judge the data which is within the scope of the center is the cloud data object or not by density.3 Combine with the fuzzy-density algorithm and the FMEFFCM algorithm, the paper exploits a interface about remote sensing image clustering by MATLAB language. The software is mainly divided into density module?the fuzzy clustering module?classification module and the result output module. The paper has carried on a brief explanation to each module and demonstrates the operation for the experimental parts about remote sensing image.
Keywords/Search Tags:Fuzzy c-means algorithm, The tighten clustering degree, Multi-scale, Fuzzy- density algorithm
PDF Full Text Request
Related items