Research On Remote Sensing Image Water Segmentation Method Based On Fuzzy Clustering

Counting the area of water resources is a prerequisite for understanding and protecting water resources,and it can also provide support for water area detection.The rapid development of remote sensing technology has gradually made remote sensing technology an important way to obtain ground data in general.Because remote sensing images have the advantages of high accuracy and strong real-time performance.The dynamic analysis and research of remote sensing images have been widely used and developed in the application and detection of water resources.In this way,advanced remote sensing technology and image processing technology are used to obtain information such as river banks and water area,and real-time observation of river information can be realized,and river information can be obtained quickly and effectively.All work first needs to effectively segment the water body in the remote sensing image.Fuzzy clustering is more suitable for processing complex remote sensing images,so this article will focus on how to use fuzzy clustering algorithms and effectively segment remote sensing images.The main work of the paper is as follows:(1)Summarize and compare the current situation of remote sensing image water segmentation at home and abroad,as well as the advantages and disadvantages of various algorithms,and analyze the theoretical basis of remote sensing image water extraction.(2)Aiming at the problem of the reduction of segmentation accuracy caused by excessive background information of remote sensing images,a remote sensing image denoising algorithm is proposed.Traditional bilateral filtering is an algorithm that can retain image edge information while removing image noise,but it can obtain limited image edge detail information in noisy images.Therefore,this paper introduces an improved bilateral filtering denoising algorithm based on rough set.According to the advantages of the rough set,which can do better in some area,such as deal with the uncertainty datas,the rough set theory is used to obtain the classes label of each pixel and the rough edge map of the image,and the two sets of information are imported into the traditional bilateral filtering framework to improve The performance of the filter to maintain the edge and denoise.Experimental results show that compared with traditional bilateral filtering,this method can improve the signal-to-noise ratio by 1.2-5.1Db,and the structural similarity can be increased by about 0.11-0.37.It can better retain the detailed information of the water body while denoising.(3)Aiming at the problem that traditional clustering algorithms have weak noise suppression ability and too many parameters that need to be set manually in remote sensing image water segmentation,this paper proposes a fuzzy clustering remote sensing image water segmentation method combined with gravity model.First,the fuzzy membership matrix obtained by the fuzzy C-means algorithm is used as the initial membership matrix of the algorithm,and then the ratio of the normalized local standard deviation to the local mean is used as a weighting factor to reflect the degree of influence of the neighboring pixels on the central pixel.Finally,combined with the spatial attraction model,a trade-off weighting factor is introduced in the relationship between the local space and the fuzzy membership to segment the remote sensing image.Experimental results show that in recent years,the segmentation performance is better than the traditional fuzzy C-means clustering algorithm and related representative algorithms.The segmentation performance has obvious advantages,the segmentation accuracy is the highest and the highest is 97.1%,and the false alarm rate is reduced by about 15%-30%.