Detection of urban buildings is a pre-requisite in urban planning and management. Very High Resolution (VHR) images acquired from Unmanned Aerial Vehicle (UAV) can play a dominant role in extraction of urban features effectively. In recent years, Geographic Object Based Image Analysis (GEOBIA) has been highly utilized for classification of VHR images, than the traditional pixel based classification owing to its novel paradigm and very high accuracy. The present study aims at detecting the dense urban buildings more precisely and reliably through GEOBIA using the orthomosaic image, Digital Surface Model (DSM) and Digital Terrain Model (DTM) processed from VHR UAV images. Dense urban buildings in Khanjarpur area of Roorkee, covering an area of about 1.63 acres was selected for this experimental study. As the object-based classification involves both segmentation and classification, multi-resolution segmentation algorithm is utilized for segmentation and to select suitable values of parameters such as scale, compactness and shape for building detection and extraction. Classification has been executed after segmentation with a formulated set of rules. Further, the classification accuracy is verified through reference data obtained through heads-up digitization of buildings from the VHR UAV orthomosaic image. The extracted buildings achieved a overall accuracy of 88.1% and 76.3% as cross verified with reference buildings using object based and area based accuracy measures respectively.
Object Based Automatic Detection of Urban Buildings Using UAV Images
Lecture Notes in Civil Engineering
International Conference on Unmanned Aerial System in Geomatics ; 2019 ; Roorkee, India April 06, 2019 - April 07, 2019
2020-02-23
14 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Automatic aircraft object detection in aerial images
SPIE | 2003
|Automatic description of complex buildings from multiple images
British Library Online Contents | 2004
|Automatic aircraft object detection in aerial images [5253-116]
British Library Conference Proceedings | 2003
|Moving object detection using strip frame images
Online Contents | 2010
|