Automatic Digital Terrain Model Extraction using strategies based on segmentation and classification

• Grant number: 08/04431-8
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: September 01, 2008
• End date: August 31, 2009
• Field of knowledge: Physical Sciences and Mathematics - Geosciences - Geodesy
• Principal Investigator: Antonio Maria Garcia Tommaselli
• Grantee: Juliana Takahashi Miyoshi
• Host Institution: Faculdade de Ciências e Tecnologia (FCT). Universidade Estadual Paulista (UNESP). Campus de Presidente Prudente. Presidente Prudente , SP, Brazil

Abstract

Generation of Digital Terrain Models from optical images is still a topic of interest both in the scientific and industry fields due to the crucial importance of this information in the current Photogrammetric processes. Existing commercial Photogrammetric software usually generate the DTM without a previous classification and recognition of the scene objects and, as a consequence, building corners, rooftops and trees are measured as being terrain points. This model is known as Digital Surface Model. Due to this, it is mandatory to edit the automatic extracted data. Another problem is the quality of the contours generated by interpolation from this data, which are unsatisfactory, requiring manual editing. The aim of this project is the development of a methodology to introduce a previous stage in the process of Digital Terrain Modeling, which is to use different strategies for each land cover element, using images previously classified images. This classification will drive the correspondence process by applying different strategies for each segment instead of using a generic process. The obtained results will be assessed and the contribution of this proposal to the automation of the Photogrammetric pipeline will be evaluated as well.