In addition to precautionary measures for avoiding forest fires, early warning and immediate response to a fire breakout are the only ways to prevent environment and cultural heritage from damages. To this end, a 3D predicted fire propagation system has been developed within FIRESENSE by using a computer vision based algorithm for wildfire detection and a 3D fire propagation estimation.
Firstly, in wildfires, smoke is first observed. In order to early identify fires, the project focused on smoke detection. A video based technology was chosen among other options. Smoke is difficult to model due to its dynamic texture and irregular motion characteristics. The main characteristics of smoke are described as follows: it moves slowly, its colour is grey and it moves vertically.
The main detection algorithm that has been developed is composed of four sub-algorithms that aim at detecting (i) slow moving objects, (ii) smoke-coloured regions, (iii) rising regions, and (iv) shadow regions. The first three sub-algorithms address the identification of smoke main properties while the last one allows discriminating smoke from shadows (shadows of the objects in the scene, especially clouds, may show the similar features than smoke). For each property, a decision value is generated. These sub-algorithm decisions are then used to give the final “smoke” or “not smoke” decision.
Secondly, once a wildfire is detected, the main focus deals with assessing its propagation direction and speed. In order to model the propagation, the monitored area is divided into cells and an algorithm is in charge of modelling how the fire propagates from one cell to another. The algorithm selected for this estimation considers fire propagation depending on a number of parameters: ignition points, fuel model, humidity, wind terrain data and other factors.
According to the algorithm, these parameters are either measured or estimated for each cell and assumed to be constant with respect to time. When a cell is ignited, the calculation of the ignition times for its neighbouring cells is performed. The propagation of fire from one cell to another depends on the ignitability of the cell, which is calculated only once per cell. However, as time increases and fire propagates further, some of the parameters in some cells may change. To cope with this problem, the dynamically changing parameter values within a recursive computation of the ignition times have been taken into account. This allows estimating the fire propagation in a changing landscape.
This propagation can be visualized in any 3D-GIS environment. The project aimed to visualize raw propagation data on Google EarthTM. The main reasons for choosing Google EarthTM are its public availability and its large utilisation by both experts and non-experts. Moreover, it allows creating impressive 3-D animations of the fire propagation, in addition to static views. The timeline of the fire propagation can be represented thanks to its layered design. Positions of the deployed equipment, observation posts, fire-fighting units etc. can also be visualized on the map.
Finally, the introduced solution focuses on quick and reliable detection and localization of the fire because it is much easier to suppress a fire when the starting the location is known, and while it is in its early stages. Furthermore, the solution provides information about the fire progress from which fire fighting staffs can rely on to manage their resources and protect cultural heritage sites.
Fire detection and 3D fire propagation estimation for the protection of cultural heritage areas, Kosmas Dimitropoulos, Kivanc Köse, Nikos Grammalidis and Enis Cetin