|Title||Multi-temporal change image inference towards false alarms reduction for an operational photogrammetric rockfall detection system|
|Author(s)||Partsinevelos, Panagiotis; Kallimani, Christina; Tripolitsiotis, Achilleas|
|Source||In: Proceedings of SPIE - The International Society for Optical Engineering. - SPIE - ISBN 9781628417005|
|Event||3rd International Conference on Remote Sensing and Geoinformation of the Environment, RSCy 2015, Paphos, 2015-03-16/2015-03-19|
|Department(s)||Laboratory of Geo-information Science and Remote Sensing|
|Publication type||Contribution in proceedings|
|Keyword(s)||false alarms - multi-temporal - rockfall|
Rockfall incidents affect civil security and hamper the sustainable growth of hard to access mountainous areas due to casualties, injuries and infrastructure loss. Rockfall occurrences cannot be easily prevented, whereas previous studies for rockfall multiple sensor early detection systems have focused on large scale incidents. However, even a single rock may cause the loss of a human life along transportation routes thus, it is highly important to establish methods for the early detection of small-scale rockfall incidents. Terrestrial photogrammetric techniques are prone to a series of errors leading to false alarm incidents, including vegetation, wind, and non relevant change in the scene under consideration. In this study, photogrammetric monitoring of rockfall prone slopes is established and the resulting multi-temporal change imagery is processed in order to minimize false alarm incidents. Integration of remote sensing imagery analysis techniques is hereby applied to enhance early detection of a rockfall. Experimental data demonstrated that an operational system able to identify a 10-cm rock movement within a 10% false alarm rate is technically feasible.