King's College London

Research portal

A Machine Learning Approach to Waterbody Segmentation in Thermal Infrared Imagery in Support of Tactical Wildfire Mapping

Research output: Contribution to journalArticlepeer-review

Jacqueline A. Oliver, Frédérique C. Pivot, Qing Tan, Alan S. Cantin, Martin J. Wooster, Joshua M. Johnston

Original languageEnglish
Article number2262
JournalREMOTE SENSING
Volume14
Issue number9
DOIs
Published1 May 2022

Bibliographical note

Funding Information: Funding: Aspects of the field campaign providing the data for this study were part of the Fire Detection Experiment (FIDEX) conducted under a program of, and funded by, the European Space Agency (ESA Contract No. 4000122813/17/I-BG), from NERC National Capability funding to the National Centre for Earth Observation (NE/Ro16518/1) and Leverhulme Trust grant number RC-2018-023 to the Leverhulme Centre for Wildfires, Environment and Society. Aspects of this study were supported through a Collaborative Research Agreement between AFFES and the Canadian Forest Service (CFS), as well as the Canadian Safety and Security Program Project Charter CSSP-2019-TI-2442 between CFS and Defence Research and Development Canada. Funding Information: Aspects of the field campaign providing the data for this study were part of the Fire Detection Experiment (FIDEX) conducted under a program of, and funded by, the European Space Agency (ESA Contract No. 4000122813/17/I-BG), from NERC National Capability funding to the National Centre for Earth Observation (NE/Ro16518/1) and Leverhulme Trust grant number RC-2018-023 to the Leverhulme Centre for Wildfires, Environment and Society. Aspects of this study were supported through a Collaborative Research Agreement between AFFES and the Canadian Forest Service (CFS), as well as the Canadian Safety and Security Program Project Charter CSSP-2019-TI-2442 between CFS and Defence Research and Development Canada. The NERC Airborne Research and Survey (Airborne Remote Sensing Facility) and British Antarctic Survey are thanked for their support of the flight campaign, without which this research would not have been possible. The authors would like to thank the Ontario government’s Aviation Forest Fire and Emergency Services (AFFES) division for their logistical and operational support in executing the field campaign, with special thanks to the Red Lake Fire Management Headquarters and Colin McFayden. We also thank Alexander Charland for providing preliminary data management support. Publisher Copyright: © 2022 by the authors. Licensee MDPI, Basel, Switzerland.

King's Authors

Abstract

Wildfire research is working toward near real-time tactical wildfire mapping through the application of computer vision techniques to airborne thermal infrared (IR) imagery. One issue hindering automation is the potential for waterbodies to be marked as areas of combustion due to their relative warmth in nighttime thermal imagery. Segmentation and masking of waterbodies could help resolve this issue, but the reliance on data captured exclusively in the thermal IR and the presence of real areas of combustion in some of the images introduces unique challenges. This study explores the use of the random forest (RF) classifier for the segmentation of waterbodies in thermal IR images containing a heterogenous wildfire. Features for classification are generated through the application of contextual and textural filters, as well as normalization techniques. The classifier’s outputs are compared against static GIS-based data on waterbody extent as well as the outputs of two unsupervised segmentation techniques, based on entropy and variance, respectively. Our results show that the RF classifier achieves very high balanced accuracy (>98.6%) for thermal imagery with and without wildfire pixels, with an overall F1 score of 0.98. The RF method surpassed the accuracy of all others tested, even with heterogenous training sets as small as 20 images. In addition to assisting automation of wildfire mapping, the efficiency and accuracy of this approach to segmentation can facilitate the creation of larger training data sets, which are necessary for invoking more complex deep learning approaches.

View graph of relations

© 2020 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454