Home // International Journal On Advances in Systems and Measurements, volume 9, numbers 3 and 4, 2016 // View article

Combining spectral and spatial information for heavy equipment detection in airborne images

Authors:
Katia Stankov
Boyd Tolton

Keywords: remote sensing image processing; right-of-way threats detection; differential morphological profile; spectral information; Hausdorff distance

Abstract:
Unsupervised construction on the pipeline right-of-way may provoke pipe rupture and consequently gas leaks. Heavy equipment is seen as a clue for construction activity. Monitoring the pipeline right-of-way for heavy equipment is therefore important for environmental and human safety. Remotely sensed images are an alternative to expensive and time consuming foot patrol. Existing image processing methods make use of previous images and/or external data. Both are not always available. We propose a new method for image processing to detect heavy equipment without the need of auxiliary data. We first detect potential heavy equipment locations and then use spatial descriptors and spectral information to eliminate false alarms. The method was validated in different environments – urban, vegetation, open excavation – and in different seasons. The experiments demonstrated the capacity of the method to detect heavy equipment without the use of previous images and/or external data.

Pages: 199 to 209

Copyright: Copyright (c) to authors, 2016. Used with permission.

Publication date: December 31, 2016

Published in: journal

ISSN: 1942-261x