Multi-resolution Representation of Lakes in Remote Sensing Images Considering Morphological Characteristics
Multi-resolution representation of remote sensing images plays an important role in data transmission,computational efficiency and privacy protection.The traditional multi-resolution representation of remote sensing images ignores the overall morphological features of the target,which is prone to information redundancy and affects the readability of the image.In this paper,a multi-resolution representation of remote sensing images taking into account morphological features is proposed by combining superpixel segmentation and texture transfer technologies.The method firstly analyses the morphological characteristics of lakes based on geographic feature differences and fractal dimensions.Then,the lake boundaries are simplified based on the superpixel segmentation algorithm.Finally,the original and simplified boundaries of the lakes are used for texture transformation to generate multi-resolution remote sensing images.The results show that the method can effectively remove redundant details and improve the readability of the object in the process of multi-resolution representation of remote sensing images.
万方数据
维普
