Research On Object-oriented Change Detection Algorithm Of High-resolution Optical Image

Remote sensing image change detection is a technology for determining surface changes by using the images acquired in the same geographical area but at different times.It has been widely used in urban planning,agricultural surveys,environmental protection,and disaster assessment.With the development of remote sensing satellite,the resolution of optical remote sensing images has been greatly improved,thus,the image is more clearer and abundant.Traditional pixel-based change detection algorithm has some limitations due to the fact that it cannot adapt to the rich information characteristics of high-resolution images.However,the object-oriented change detection algorithm can fully consider the spectral,texture and structure information of the image to improve the ability for distinguishing whether the image objects have changed.Therefore,the object-oriented change detection algorithm has become a hot research topic in the field of remote sensing.In this paper,we research the problems of multi-feature optimization and difference image classification for object-oriented change detection algorithms,and propose two object-oriented change detection algorithms in high-resolution optical image.The main contents are summarized as follows:In order to solve the problem of feature redundancy in existing object-oriented algorithms,a object-oriented change detection algorithm based on feature optimization is proposed by combining with the locally linear embedding(LLE)theory.Firstly,the two temporal remote sensing images are divided into objects by using multi-scale segmentation algorithm.Secondly,the spectral and texture features of objects are extracted,and the feature change vector is constructed.Then,the improved LLE algorithm is designed to optimize the feature change vector,so as to the high-quality features are automatically mined while the feature amounts are compressed.Finally,the difference image is generated according to optimized feature,and the FCM algorithm is carried to cluster the difference image to obtain the final change detection results.Experimental results construct on real GF-1 datasets confirm the validity of the proposed algorithm in improving the accuracy of change detection.To alleviate the problem that the FCM algorithm is easily trapped into local optimization in the process of difference image classification,a object-oriented change detection approach based on FCM clustering with the pigeon-inspired optimization(PIO)theory is proposed.Firstly,the difference image is generated by feature optimization,and the PIO algorithm is carried to search the optimal solution on the difference image.Secondly,the optimal solution is regarded as the initial clustering center of the FCM algorithm to achieve difference image classification,thus,the change detection results is obtained.Experimental and analysis results on the GF-1 datasets verify that the proposed approach offers great contributions on improving the robustness and accuracy of change detection.