基于区间直觉模糊集的遥感图像滤波方法研究地图制图学与地理信息工程专业论文.docxVIP

II II Abstract Remote sensing image filtering as an important part in remote sensing image preprocessing. Filtering effect will directly respond to the validity and reliability of the subsequent image processing and analysis. Therefore, the study of remote sensing image filtering is a critical step for the further processing and analysis. Fuzzy set theory is introduced into remote sensing image filtering based on its superiority in dealing with uncertainty. And some remote sensing image filtering methods are proposed based on Interval-valued intuitionistic fuzzy sets(IVIFSs).The study content and results of this article are as follow: 1．Analyzing the formation mechanism and classification of remote sensing image noise. Giving the remote sensing image filtering performance evaluation criteria and methods, including subjective evaluation and objective evaluation. Then, review median filtering and mean filtering which are two commonly used image filtering algorithm. And outline three image filtering methods based on Fuzzy theory: fuzzy weighted mean filter, median - fuzzy filtering and intuitionistic fuzzy entropy filter. 2．Make an introduction of the concept of Interval-valued intuitionistic fuzzy sets and Interval-valued intuitionistic set metrics, including the Interval-valued intuitionistic fuzzy entropy and similarity measure, as the theoretical basis of the proposed filtering method. 3．Impulse noise filter based on Interval-valued intuitionistic fuzzy entropy and the constructor of Interval-valued intuitionistic fuzzy set membership degree interval are presented. Through the window extremes screening to detect noisy image, use Interval-valued intuitionistic fuzzy entropy to calculat the noise gray value and obtain target pixel gray value to complete the whole filter. 4．Designing a IVIFSs similarity measure Gaussian noise filtering algorithm. On the basis of the fuzzy weighted filtering, using the similarity measures of the target pixel and its neighbor as weights