Image Processing, Analysis and Modeling of Particle Populations

Johan Debayle (1, 2, 3, 4, 5, 6)

1 Mines Saint-Étienne MSE - École des Mines de Saint-Étienne
2 Université de Lyon
3 CNRS - Centre National de la Recherche Scientifique
4 LGF-ENSMSE - Laboratoire Georges Friedel
5 SPIN-ENSMSE - Centre Sciences des Processus Industriels et Naturels
6 PMMG-ENSMSE - Département Procédés de Mise en oeuvre des Milieux Granulaires

Particle populations are widely used in many industrial applications and fields of science from physics to biology or agronomy. In chemical engineering, in particular, it is generally desired to extract information on geometrical characteristics and on spatial distribution from 2D images of the population of particles involved in the process. For example in pharmaceutics, the size and the shape of crystals of active ingredients are known to have a considerable impact on the final quality of products, such as drugs. Hence, it is of main importance to be able to control in real time the granulometry (size and shape) of the crystals during the process. The purpose of this talk is then to show different ways (deterministic and stochastic methods) of image processing, analysis and modeling to geometrically characterize the particles from a sequence of 2-D images acquired by a camera (visualizing the particles during a particular process). The developed methods will be presented by addressing different issues: the perspective projection of the 3-D particle shape onto the image plane, the blurred appearance of unfocused particles, the degree of agglomeration or overlapping, and the random variation in size/shape of the observed particles. The methods are mainly based on image enhancement, restoration, segmentation, tracking, modeling, feature detection, stereology, stochastic geometry, pattern analysis and recognition. The methods will be particularly illustrated on real applications of crystallization processes (for pharmaceutics industry) and multiphase flow processes (for nuclear industry).

Mots clés: image analysis; stochastic geometry; granular media; pattern recognition

Domaines: Génie des procédés; Traitement du signal et de l'image [eess.SP]; Vision par ordinateur et reconnaissance de formes [cs.CV]; Modélisation et simulation

Johan Debayle. Image Processing, Analysis and Modeling of Particle Populations. 2018 The 11th International Conference on Machine Vision (ICMV 2018), School of Electronics - Si Chuan University, Chine - Halmstad University, Suède - University of Barcelona, Espagne, Nov 2018, Munich, Germany. ⟨hal-01936548⟩