The Earthquake Image Change Detection Algorithm Research And Implementation Based On NSCT Of Texture Analysis

In the study about the change detection of complex land surface Images in different periods from geological disasters,most of the image domain change detection method is simply to the image of the spectral or texture information for simple data statistics,in the low resolution of seismic image recognition accuracy of rivers,lakes and other surface contour obvious deficiencies,as well as by noise causes the change detection results in high bit error rate and false alarm rate.Therefore,in order to better compensate for lack of detection of the method in the earthquake disaster,this paper combines the multiscale techniques and features in texture analysis method to optimize the change detection algorithm model to improve detection accuracy of earthquake disaster areas in the image in rivers,lakes and other types of damage,for residents in the disaster areas as soon as possible to reduce the disaster loss and post disaster reconstruction provides the reliable scientific basi.In this paper,research on image change detection technology based on the logarithmic difference function edge texture enhancement of,the main target is improve recognition accuracy of seismic image object types of edge profile changes in the region,to compensate for the most detection method in the insufficient precision of image edge feature change detection.Firstly,the texture analysis model based on directional log difference function is designed to recognize and enhance the main edges of objects in different periods.Then,the texture similarity detection model VSSIM based on log difference function is designed to extract the features of texture difference between different scales in NSCT transform space.Finally,based on the relational fuzzy c-means clustering method to differences in the characteristics of accurate classification,and the use of regional growth method to optimize the classification results,extracted from the final seismic image change detection results.In this paper,image change detection method based on NSCT adaptive to the local texture analysis,the main target is based on texture features of different local neighbourhood of the image to carry out according to the features in texture analysis and enhancement to compensate for most detection method in image features local detail features change detection accuracy is not enough.First use the mean and the variance of the image of local to describe the difference of different regional features in texture,and in combination with the multiscale technique to construct based on NSCT adaptive local image texture enhancement method can not only keep the basic texture features of the source image invariant of,you can also highlight the image object types of edge information.Is conducive to the extraction of NSCT transform decomposition of different scales on the image features of the texture difference,aswell as changes in the seismic region detection results output.Experiment results show that the two methods in terrain types of seismic image difference detection has an advantage each,variation function texture analysis can accurately identify the edge details of the image in the rivers,lakes and other surface and adaptive to the local texture analysis can also keep the details of the changes in the region and clear edge structure,while also reducing the value of noise interference to the object feature detection and for people in the disaster prevention,disaster assessment and post disaster emergency response and other aspects provide a reliable basis for decision making.