Kohonen Self Organizing for Automatic Identification of Cartographic Objects

Automatic identification and localization of cartographic objects in aerial and satellite images have gained increasing attention in recent years in digital photogrammetry and remote sensing. Although the automatic extraction of man made objects in essence is still an unresolved issue, the man made objects can be extracted from aerial photos and satellite images. Recently, the high-resolution satellite images, typically at most 3 meters in panchromatic band ground sample distance (GSD) and up to four multispectral bands in the visible and near infrared spectrum, are suitable for detection and identification of objects. This paper presents a new algorithm for identification of cartographic objects based on Artificial Neural Network (ANN). The algorithm is divided in two modules: image simplification by the Wavelet transform, Mathematical Morphology (MM) operators, and identification of object by the Kohonen Self Organizing Map (KSOM) and split and merge method. The study area included two parts of an orthoimage from Kish, Iran.