Lucile Rossi

A 3D vision system for the measurement of the rate of spread and the height of fire fronts

This paper presents a three-dimensional (3D) vision-based instrumentation system for the measurement of the rate of spread and height of complex fire fronts. The proposed 3D imaging system is simple, does not require calibration, is easily deployable in indoor and outdoor environments and can handle complex fire fronts. New approaches for measuring the position, the rate of spread and the height of a fire front during its propagation are introduced. Experiments were conducted in indoor and outdoor conditions with fires of different scales. Linear and curvilinear fire front spreading were studied. The obtained results are promising and show the interesting performance of the proposed system in operational and complex fire scenarios.

Dynamic fire 3D modeling using a real-time stereovision system

This work presents a new framework for three-dimensional modeling of dynamic fires present in unstructured scenes. The proposed approach addresses the problem of segmenting fire regions using information from YUV and RGB color spaces. Clustering is also used to extract salient points from a pair of stereo images. These points are then used to reconstruct 3D positions in the scene. A matching strategy is proposed to deal with mismatches due to occlusions and missing data. The obtained data are fitted in a 3D ellipsoid in order to model the enclosing fire volume. This form is then used to compute dynamic fire characteristics like its position, dimension, orientation, heading direction, etc. These results are of great importance for fire behavior monitoring and prediction.

Automatic fire pixel detection using image processing: a comparative analysis of rule-based and machine learning-based methods

This paper presents a comparative analysis of state-of-the art image processing-based fire color detection rules and methods in the context of geometrical characteristics measurement of wildland fires. Two new rules and two new detection methods using an intelligent combination of the rules are presented, and their performances are compared with their counterparts. The benchmark is performed on approximately two hundred million fire pixels and seven hundred million non-fire pixels extracted from five hundred wildland images under diverse imaging conditions. The fire pixels are categorized according to fire color and existence of smoke; meanwhile, non-fire pixels are categorized according to the average intensity of the corresponding image. This characterization allows to analyze the performance of each rule by category. It is shown that the performances of the existing rules and methods from the literature are category dependent, and none of them is able to perform equally well on all categories. Meanwhile, a new proposed method based on machine learning techniques and using all the rules as features outperforms existing state-of-the-art techniques in the literature by performing almost equally well on different categories. Thus, this method, promises very interesting developments for the future of metrologic tools for fire detection in unstructured environments.

Benchmarking of wildland fire colour segmentation algorithms

Recently, computer vision-based methods have started to replace conventional sensor-based fire detection technologies. In general, visible band image sequences are used to automatically detect suspicious fire events in indoor or outdoor environments. There are several methods which aim to achieve automatic fire detection on visible band images, however, it is difficult to identify which method is the best performing as there is no fire image dataset which can be used to test the different methods. This study presents a benchmarking of state of the art wildland fire colour segmentation algorithms using a new fire dataset introduced for the first time. The dataset contains images of wildland fire in different contexts (fuel, background, luminosity, smoke etc.). All images of the dataset are characterised according to the principal colour of the fire, the luminosity, and the presence of smoke in the fire area. With this characterisation, it has been possible to determine on which kind of images each algorithm is efficient. Also a new probabilistic fire segmentation algorithm is introduced and compared to the other techniques. Benchmarking is performed in order to assess performances of 12 algorithms that can be used for the segmentation of wildland fire images.

Three-dimensional tracking for efficient fire fighting in complex situations

Each year, hundred millions hectares of forests burn causing human and economic losses. For efficient fire fighting, the personnel in the ground need tools permitting the prediction of fire front propagation. In this work, we present a new technique for automatically tracking fire spread in three-dimensional space. The proposed approach uses a stereo system to extract a 3D shape from fire images. A new segmentation technique is proposed and permits the extraction of fire regions in complex unstructured scenes. It works in the visible spectrum and combines information extracted from YUV and RGB color spaces. Unlike other techniques, our algorithm does not require previous knowledge about the scene. The resulting fire regions are classified into different homogenous zones using clustering techniques. Contours are then extracted and a feature detection algorithm is used to detect interest points like local maxima and corners. Extracted points from stereo images are then used to compute the 3D shape of the fire front. The resulting data permits to build the fire volume. The final model is used to compute important spatial and temporal fire characteristics like: spread dynamics, local orientation, heading direction, etc. Tests conducted on the ground show the efficiency of the proposed scheme. This scheme is being integrated with a fire spread mathematical model in order to predict and anticipate the fire behaviour during fire fighting. Also of interest to fire-fighters, is the proposed automatic segmentation technique that can be used in early detection of fire in complex scenes.

Estimation of spreading fire geometrical characteristics using near infrared stereovision

In fire research and forest firefighting, there is a need of robust metrological systems able to estimate the geometrical characteristics of outdoor spreading fires. In recent years, we assist to an increased interest in wildfire research to develop non destructive techniques based on computer vision. This paper presents a new approach for the estimation of fire geometrical characteristics using near infrared stereovision. Spreading fire information like position, rate of spread, height and surface, are estimated from the computed 3D fire points. The proposed system permits to track fire spreading on a ground area of 5mx10m. Keywords: near infrared, stereovision, spreading fire, geometrical characteristics

Estimation of fire volume by stereovision

This paper presents a new approach for the estimation of fire front volume in indoor laboratory experiments. This work deals with fire spreading on inclinable tables. The method is based on the use of two synchronized stereovision systems positioned respectively in a back position and in a front position of the fire propagation direction. The two vision systems are used in order to extract complementary 3D fire points. The obtained data are projected in a same reference frame and used to build a global form of the fire front. An inter-systems calibration procedure is presented and permits the computation of the projection matrix in order to project all the data to a unique reference frame. From the obtained 3D fire points, a three dimensional surface rendering is performed and the fire volume is estimated.

Measurement of the geometric characteristics of a fire front by stereovision techniques on field experiments

This paper presents stereovision techniques for measurement of the geometrical properties (position, rate of spread, fire height, fire inclination angle, fire base width, view factor) of fires obtained by experimental burnings at field scale. The system consists of two synchronized and pre-calibrated multi-baseline stereo cameras operating in the visible spectrum. The cameras are positioned in the back and the lateral positions relatively to the direction of fire propagation. Algorithms have been developed in order to (i) register these cameras, (ii) model in three dimensions the fire front from the back stereoscopic images and (iii) estimate some geometrical properties of fire such as the inclination angle and the fire base width from the lateral stereoscopic images. A user graphical interface was developed as a practical tool to estimate fire propagation features and to display the obtained results. Fire spread experiments were conducted at field scale (about 20 m wide and 3 m high). The fuel consists of Mediterranean shrub vegetation. The obtained results are promising and show interesting performance achieved by the proposed system in operational and complex fire scenarios.

Dynamic fire modeling in Three-dimensional space

This work presents a new framework for three-dimensional reconstruction of dynamic fire fronts found in outdoor unstructured scenes. The proposed approach addresses the problem of segmenting fire front regions using color attributes and clustering techniques in order to extract salient points from stereo images. These points are then used to reconstruct their 3D position in the scene. A matching strategy is proposed to deal with mismatches due to occlusions and missing data. The proposed framework was successfully used to build 3D data part of dynamic fire fronts. This 3D information was used efficiently to estimate the approximate position of the fire front and its heading direction. The obtained results are promising and show the possibility of tracking dynamic objects with changing shapes like fire fronts in a three-dimensional space.

Relationship Between Flame Length and Fireline Intensity Obtained by Calorimetry at Laboratory Scale

The difficulties in measuring Byrams fireline intensity have led many researchers to derive an empirical relation between the fireline intensity and flame length, which is easier to measure at the practical (firefighting) level. In this article, we address both the estimation of Byrams fireline intensity by comparison with oxygen consumption calorimetry (OC) measurement and the test of formulations for fireline intensity versus flame length. We directly measured the fireline intensity for spreading fires on a laboratory scale under conditions of no wind and no slope by OC. The fires were set across beds of Pinus pinaster needles, Avena fatua straw, and Genista salzmannii spines. Fireline intensity obtained by OC ranged from 28 to 160 kW/m, depending on species and load. Byrams index of fire intensity ranged from 38 to 185 kW/m. It was observed that Byrams intensity overestimated the fireline intensity measured by OC. Combustion efficiency was introduced in Byrams formulation, which led to a better estimate of fireline intensity. The indirect estimation of fireline intensity through the observation of the flame length was carried out using a stereovision system. Mean flame length ranged from 0.35 to 0.84 m according to the species and the fuel load. A new relationship was established between the fireline intensity and the flame length.