Olivier Sero-Guillaume

On the Evaluation of Segmentation Methods for Wildland Fire

This paper focuses on the study of fire color spaces and the evaluation of image segmentation methods commonly available in the literature of wildland and urban fires. The evaluation method, based on the determination of a segmentation quality index, is applied on three series of fire images obtained at the usual scales of validation of forest fire models (laboratory scale, fire tunnel scale and field scale). Depending on the considered scale, different methods reveal themselves as being the most appropriate. In this study we present the advantages and drawbacks of different segmentation algorithms and color spaces used in fire detection and characterization.

Comparison of two methods for estimating fire positions and the rate of spread of linear flame fronts

The aim of this paper is to compare two continuous methods for the identification of fire front positions and the rate of spread of linear flame fronts. The first technique is based on the heat fluxes measured by a specific thermal sensor and the use of the inverse method. The second one uses an image processing technique applying the simple direct linear transformation (DLT) algorithm and the fire segmentation algorithm to obtain fire front positions. The fire segmentation method has been selected by benchmarking four recently published studies and involves a supervised evaluation criterion. The two continuous approaches give closely similar results and it is shown that these two techniques can be respectively validated by each other.

Measurement of radiative absorption coefficient for a vegetal medium

The radiative heat flux received by vegetation is a leading process of the propagation of forest fires. This flux is a function of the radiative absorption coefficient. We have measured this coefficient for Quercus coccifera. The variation of the values of the absorption coefficient as a function of the load (surface density of vegetation) has been studied. We have shown that the absorption coefficient is proportional to the load; the proportionality coefficient depends upon the nature of the vegetation.

Investigation of a novel image segmentation method dedicated to forest fire applications

To face fire it is crucial to understand its behaviour in order to maximize fighting means. To achieve this task, the development of a metrological tool is necessary for estimating both geometrical and physical parameters involved in forest fire modelling. A key parameter is to estimate fire positions accurately. In this paper an image processing tool especially dedicated to an accurate extraction of fire from an image is presented. In this work, the clustering on several colour spaces is investigated and it appears that the blue chrominance Cb from the YCbCr colour space is the most appropriate. As a consequence, a new segmentation algorithm dedicated to forest fire applications has been built using first an optimized k-means clustering in the Cb-channel and then some properties of fire pixels in the RGB colour space. Next, the performance of the proposed method is evaluated using three supervised evaluation criteria and then compared to other existing segmentation algorithms in the literature. Finally a conclusion is drawn, assessing the good behaviour of the developed algorithm.

Asymptotic matching and genetic optimization applied to optimal aerospace design

For many optimum design problems, the cost function is the result of a complex numerical code, and the desired optimum is global. The global optimization code requires many evaluations of the cost function, which implies that the optimum cannot be obtained in a reasonable computation time. The first aim of this paper is to propose, on an example problem, a way of solving such complexity. The problem considered here concerns an aerospace vehicle which is intended to fulfil a given task. The control of the vehicle for achieving the task is obtained as the solution of an optimal time control problem. Our aim is to design the architecture of the vehicle so that it can complete the task in minimum time. For a given architecture the minimum time and the control are calculated using a very fast semi-analytical method, obtained by asymptotic matching. This minimum time is taken as the cost function for a given architecture design. The best architecture is then computed by a stochastic genetic-like global optimization algorithm. A presentation of this genetic algorithm is the second aim of the paper.

Shapes of liquid drops obtained using symbolic computation

The first aim of this paper is to show how two free boundary problems arising from fluid mechanics can be solved with a domain perturbation method. The second aim is to analyse the range of validity of the series solutions. The analysis will aim at identifying the location and the nature of the singularities characterizing these series. After expansion, the equations obtained are linear, but at each stage the length of the expressions grows exponentially. Herein we implement techniques for the automatic generation of hierarchical expression sequences and we present several tools for reducing the combinatorial blow-up of the expressions arising in these two problems.

Dependence of Levy Statistics With Reynolds Number Application to turbulence intermittency and spray modeling

In this poster, we show that Kidas log-stable law for the intermittency of turbulence can be extended over a wide Range of Taylor scale Reynolds number Re »elds a simple relation between the scale parameter of the law and Re » log-normal laws were firstly introduced by Kolmogorov for the size of particles under pulverization, log-stable laws are applied to the experimental drop size p.d.f in an annular gas-liquid flow. Though the fitting is quite good there does not seem to be any clear relationship between the parameters of the law and the external parameters of the experiment. This may be related to a breaking of the homogeneity hypothesis.

Extension of the Kida law in turbulence

Abstract We extend the validity range of Kidas log-stable law of stability index α =1.65 and intermittency parameter μ =0.2 to a new range of Reynolds number. This law describes intermittencies in fully developed turbulent flows or more precisely the p.d.f. of turbulence dissipation. Former measurements of the hyper-flatness factors of order 4, 5, 6 of turbulent velocity increments, coming from both experimental works and numerical simulations are used. We show that the power-law variation of these hyper-flatness factors with Taylor scale based Reynolds numbers Re λ can be fitted, for Re λ ranging from 35 to 750, by a log-stable law of stability index α =1.65 and intermittency parameter μ =0.21. To cite this article: N. Rimbert, O. Sero-Guillaume, C. R. Mecanique 331 (2003).