Semi automatic road extraction from digital images

Egyptian Journal of Remote Sensing and Space Sciences

Hani Rezaeian¹  Abolfazl Abdollahi²  Hamid Reza Riahi Bakhtiari³ 

[1] Faculty of Geography, Department of Remote Sensing and Geographic Information Systems, Kharazmi University of Tehran, Iran;Faculty of Geography, Kharazmi University of Tehran, Iran;Faculty of Natural Resources and Earth Science, University of Shahrekord, Shahrekord, Iran;

关键词: Road extraction;    SVM;    Full Lambda;    Morphological operators;   

DOI  :  10.1016/j.ejrs.2017.03.001

来源: DOAJ

【 摘 要 】

Road extraction from digital images is of fundamental importance in the context of automatic mapping, effective urban planning and updating GIS databases. Very high spatial resolution (VHR) imagery acquired by airborne and space borne sensors is the main source for accurate road extraction. Manual techniques are fading away as they are time consuming and costly. Hence, road extraction method that is significantly more automated has become a research hotspot in remote sensing information processing. This paper proposes a semi-automatic approach to extract different road types from high-resolution remote sensing images. The approach is based on edge detection and SVM and mathematical morphology method. First the outline of the road is detected based on Canny operator. Then, Full Lambda Schedule merging method combines adjacent segments. Then the entire image was classified using Support Vector Machine (SVM) and various spatial, spectral, and texture attributes to form a road image. Finally, the quality of detected roads is improved using morphological operators. The algorithm was systematically evaluated on a variety of satellite images from Worldview, QuickBird and UltraCam airborne Images. The results of the accuracy evaluation demonstrate that the proposed road extraction approach can provide high accuracy for extraction of different road types.

【 授权许可】

Unknown