COMPARISON OF AUTOMATIC FEATURE EXTRACTION METHODS FOR BUILDING ROOF PLANES BY USING AIRBORNE LIDAR DATA AND HIGH RESOLUTION SATELLITE IMAGE

Atik M. E., Donmez S. O., Duran Z., Ipbuker C.

7th International Conference on Cartography and GIS, Sozopol, Bulgaria, 18 - 23 June 2018, pp.857-863 identifier

  • Publication Type: Conference Paper / Full Text
  • City: Sozopol
  • Country: Bulgaria
  • Page Numbers: pp.857-863
  • Keywords: 3D Point Cloud Processing, Object-Based Image Analysis, Feature Extraction
  • Istanbul Technical University Affiliated: Yes

Abstract

Airborne LIDAR technology is one of the most widely used rapid solutions for producing 3D dense point cloud. Also, automatic feature extraction is under 3D Modelling studies in Geographic Information Systems (GIS). There are several techniques for reconstruction and classification in this field. RANSAC (RANdom SAmple Consensus) is one of the data processing method in order to use LIDAR data. Additionally, another technique for feature extraction is object based point cloud analysis. In this paper, RANSAC are used for feature extraction on 3D Airborne LIDAR data. Thestudy area is selected in Turkey. Roof planes are detected by using both methods. A code is written on MATLAB software for RANSAC algorithm. eCognition software is used for object based point cloud analysis. In this analysis, very high resolution satellite images are combined with 3D point cloud of the study area. In the result of the study, the roof geometry that is obtained using both methods are compared scientifically in terms of area and number of points. In the conclusion step, the results are discussed and future works are planned.