Research On SAR Image Change Detection Method Based On LBP And Deep Learning

With the continuous development of observation projects,the community has put forward higher requirements for the accuracy of remote sensing image change detection.Synthetic Aperture Radar(SAR)images can be obtained independently of atmospheric and sunlight conditions and have unique advantages in change detection.Hence,in recent decades,SAR images have been successfully applied to fields such as environmental monitoring,urban research,and disaster assessment.However,change detection based on SAR images becomes difficult due to the presence of speckle noise inherent in SAR images.Therefore,this paper studies how to effectively improve the detection accuracy of SAR image change and increase the practicality of the change detection algorithm.The main research results are as follows:(1)In order to make full use of the characteristics of Local Binary Pattern(LBP)with strong classification ability,low computational complexity,high feature discrimination,invariability to monotonous gray level changes and able to combine the overall characteristics of images,and enhance the robustness of the algorithm to SAR image speckle noise,improved LBP and FCM is proposed for change detection in SAR image.Experimental results show that the algorithm has good detection performance,which not only improves the robustness of the algorithm to speckle noise,but also significantly improves the accuracy of change detection.(2)Aiming at the problem that the speckle noise affects the detection accuracy of SAR image change,based on the deep learning theory,Stacked Sparse Automatic Encoder(SSAE)is proposed for change detection in SAR image.The SSAE is applied to the change detection of SAR image,used to extract the deep features of the image and reduce the influence of SAR image speckle noise on the change detection.Experimental results show that SSAE has high robustness to SAR image speckle noise,and can effectively improve the accuracy of change detection.(3)In order to increase the practicability of the change detection method in SAR image,this paper builds Graphical User Interfaces(GUI)based on MATLAB R2013 a platform using GUIDE,to design software for two kinds of proposed change detection methods in SAR image.Experimental results show that the design and application of SAR image change detection software effectively improves the automation ability of change detection,and increases the effectiveness and practicability of the algorithm.