Context Aware Modification on the Object Based Image Analysis

摘要: Spectral similarities, spatial adjacencies and complex relationships between various object types lead to the recognition difficulties ambiguities in urban areas. Using a knowledge base containing contextual information together with higher levels of modifications based image analysis may improve results such situation. In this paper, composed regions’ internal external context are utilized for classification modification on pre-defined segmented regions very high resolution (VHR) remotely sensed imagery. proposed methodology, defining region’s each regions, (OBIA) is performed. Then, wrongly classified due defined mutual dependencies neighboring image. Evaluation capabilities aware methodology performed WorldView-2 satellite imagery over Rio de Janeiro Brazil digital surface model. The obtained represent that using some spectral mistakes generating rules decreased.