Remote Sensing Image Change Detection Based On Combined Difference Image And Kernel Fuzzy Clustering

Change detection in remote sensing images is a process that analyzes images of the same scene taken at different times in order to identify changes that may have occurred between the considered acquisition dates.Now,It is one of the most important technologies to update the basic geodatabase.And it has been widely used in a series of important applications,such as land cover and land use change monitoring,dynamic monitoring of urban development,natural disaster analysis and military reconnaissance and combat effectiveness evaluation and so on.So this paper studies on the thorough research to the key technology of the remote sensing image change detection.The content of this paper mainly includes the following three aspects:(1)This paper introduces the research background of remote sensing image change detection,analyzes various remote sensing image change detection algorithms at home and abroad,and systematically summarizes the concepts,mathematical models,processes and algorithms of remote sensing image change detection.Profoundly analyzes the theory of commonly used clustering algorithm,and expresses the basic concept of clustering algorithm in mathematical form.(2)A change detection method of remote sensing images based on combined difference image is proposed.Firstly,subtraction and log-ratio operations are used to generate two kinds of difference images.Then the combination of the phase of the subtraction difference image and the amplitude of the log ratio difference image obtained by Fourier transform is processed by inverse Fourier transform to generate the combined difference image.Finally,the change detection image is achieved by clustering new difference image using fuzzy C means(FCM)clustering algorithm.The combined difference image utilizes the amplitude information and phase information of the image,so the proposed method has good anti-noise performance,which further improves the detection accuracy.The validity of the method is verified by the experiment of real remote sensing data set.(3)A change detection method based on kernel-based fuzzy clustering for remote sensing images is proposed.Firstly,the new difference image is constructed by using the phase of the subtraction difference image and the amplitude of the log ratio difference image.And then use the kernel-based fuzzy clustering method to cluster the new difference image.The use of kernel functions maps the raw data to high?dimensional feature spaces,which can achieve more accurate classification.The results of real remote sensing data set show that the proposed method has better change detection effect.