Research on Object-Oriented Approach to ADS40 Image Classification

ADS40 images with High spatial resolution have more spatial characteristics as well as spectral characteristics than low-resolution data. In this paper,an object-oriented classification method based on multi-scale segmentation is introduced to classify ADS40 image of Taiyuan city. Firstly,a multi-scale segmentation algorithm is applied to get objects.Then,the features of objects,such as spectral, geometrical and topological characteristics,are measured.The hierarchy structure for classification is built.Finally,a fuzzy rule classifier is applied to extract the classification information of ADS40 images.The research shows that the object-oriented method compared with traditional classification approaches can improve the overall classification accuracy of ADS40 images,reduce the "Pepper and Salt Phenomenon"effectively,and meet the need of ADS40 image classification.