A new algorithm for object-oriented multi-scale high resolution remote sensing image segmentation

The rich spatial structure information and geographic information in a high-resolution remote sensing image are need to be extracted in different scales. However, the traditional image segmentation methods based on pixels spectral characteristics and single-scale image information extraction methods have obvious flaws in this respect. In order to utilize the rich scale-dependent information contained in high resolution remote sensing images, the geo-science applications of remote sensing image and geographical information extraction must be carried out under multi-scale condition. Region-based object-oriented image analysis method provides a new idea for high-resolution remote sensing image information extraction. The key issue is to realize multi-scale high resolution remote sensing image segmentation. In this paper, an object oriented multi-scale image segmentation method is introduced based on minimum heterogeneity criterion of neighbouring region growing. Segmentation results show that this method can easily adapt its scale parameter to different scale image analysis tasks and any chosen scale object-extraction of interest. In a word, it can provide enormous object characteristics for further object-oriented processing or analysis.