Dr. Xu Zhong will present within the PROTECT stream at the Locate16 Conference on 12-14 April 2016. The 3-day Conference will present leading industry professionals and researchers focusing on disruptive technology and how it is building a smarter society.
Abstract Title: Real-time estimation of wildfire perimeters from curated crowdsourcing
Natural disasters, such as wildfires, have significant impacts upon human lives, critical infrastructure, and delicate environments worldwide. Timely and accurate information about the areas affected by ongoing natural disasters assists emergency responders and the general public to minimise these impacts. However, in the unexpected and extreme events surrounding an emergency, reliable and up-to-date information about the affected areas often becomes scarce. During the devastating 2009 Black Saturday wildfires in Victoria, Australia, for example, information about the current extents of multiple wildfires was not always available. Authoritative information sources can suffer from bottlenecks and delays, while user-generated social media data usually lacks the necessary structure and trustworthiness for reliable automated processing. We developed an automated data mining technique for realtime tracking of wildfire perimeters based on “curated” crowdsourced data about telephone calls to the emergency services. Being crowdsourced, this data is more plentiful and more rapidly available than authoritative information sources during an emergency.
On the other hand, because emergency calls are handled by trained professional operators, the data generated is more structured and reliable than social media data. The Country Fire Authority (CFA) is providing anonymized emergency calls data within its publicly available RSS feed. Using this feed, our data mining algorithm can automatically detect and track wildfire perimeters, in real time, and with relatively high accuracy. Accuracy can be further increased through combination with other authoritative demographic and environmental information, such as population density and dynamic wind fields.
Our wildfire perimeter estimator has been validated against the post-fire reconstructed perimeter of the devastating Black Saturday wildfires (2009) and Mickleham-Dalrymple wildfires (2014) in Victoria, Australia. The results show how such publicly available “curated” crowdsourced data can provide an important source of up-to-date information about an evolving emergency.