With the rapic developments in most China cities, urban environment monitoring is very important, for example, for the subway safety, illegal drilling and construction in subway field should be quickly detected. Monitoring techniques with high precision and efficiency are vital to prevent the accidents and reduce losses, and UAV has great potential and advantages to conduct such task compared to human daily inspection. To quickly get the illegal operation information from UAV image, a method of change detection based on elevation difference and local binary pattern (LBP) is proposed to monitor the land surface along subway by unmanned aerial vehicle remote sensing (UAVRS). After accurate registrations without GCPs complete, the two DSMs and DOMS are mutually matched by the same points in their own model. Gaussian smoothing is used in gray-scale map to eliminating noise jamming before making change detection based on elevation difference. Pix4D is used to generate 2 DSMs of the study area, and texture feature is measured by LBP which is advanced in its rotational invariance and brightness invariance. Comparing with former researches, two DSMs are matched by invariant points in their own models instead of GCPs which are usually collected by GPS, with the registration precision less than 0.1m both in XY and Z directions, which meets the requirement of illegal operation detection in subway safety monitoring, and the adoption of LBP works well for images collected in different climate and illumination.

Date Published: 6 August 2018
PDF: 7 pages
Proc. SPIE 10773, Sixth International Conference on Remote Sensing and Geoinformation of the Environment (RSCy2018), 107731V (6 August 2018); doi: 10.1117/12.2323087

Published in SPIE Proceedings Vol. 10773:
Sixth International Conference on Remote Sensing and Geoinformation of the Environment (RSCy2018)
Kyriacos Themistocleous; Giorgos Papadavid; Silas Michaelides; Vincent Ambrosia; Diofantos G. Hadjimitsis, Editor(s)