Axis 1

Title : Characterization of Objects and Populations of Objects, Morphological Descriptors

Animators  : Johan Debayle – Fabrice Lamadie

 In this axis,   we seek to characterize the morphology of compact or deformable objects, individual entities, such as crystals or cells, faceted particles, agglomerates or flocs structured in a more or less compact form.

 The objective of this work is to arrive at a fine description of the evolution of the morphology of these objects under the effect of environmental constraints (exchanges with the surrounding environment in the case of biological systems for example), physicochemical (modification of ionic strength or pH, addition of additives) or hydrodynamics (elongation, shear, … …).

It is thus desired to acquire information on the morphology of objects in the stationary state (size, shape, facies, structure) and in dynamics (under the effect of a variation of the stresses) in order to quantify the ability to deformation, embrittlement (creating cracks or break points) or consolidation (structure, cohesion, porosity).

 The usual means of characterization of a population of objects (micronic or millimetric) are based on the diffusion of an electromagnetic wave in general and in particular light or the analysis of images obtained in situ by endoscopy or off line by light or electron microscopy. On-line or in-situ characterization techniques are rich sources of data. However, the means must be found to extract the relevant data from diffraction measurements, especially in a dense medium, or from particle images that often appear nested, superimposed or imperfectly contoured.

Moreover, it is possible to associate with each of the characterization techniques a set of morphological descriptors. These are usually chosen or imposed by the manufacturer of the equipment, usually independently of those used for modeling. In addition, each of the analyzes generates a great deal of information concerning the morphological characteristics of several hundreds of thousands of elements of the population, in particular aggregates. Another hard point lies in the implementation of statistical tools to find the appropriate quantitative descriptors. On this aspect, the most recent statistical techniques of multivariate analysis and learning in a context of large dimension (data mining ) could constitute lines of research for the analysis of the morphological data but a rapprochement of the communities of researchers in processes and in statistical mathematics is to operate.

 In this context, the different themes covered by this axis relate to the :

– development of methodologies derived from diffraction analysis, in situ or online, in order to extract relevant data on the morphology or structure of objects in dilute or dense suspensions and their spatial or temporal evolutions,

– development of experimental visualization techniques, in situ or online, allowing the acquisition of images of flocs, aggregates, or crystals under stationary or dynamic conditions in order to acquire data on changes in morphological properties over time,

– mathematical development for image processing in order to extract the morphological characteristics (parameter of size, shape, porosity) of the individual objects within the population,

– statistical analysis to exploit information on large data sets to identify morphological descriptors relevant to the targeted transformation process, calculate moments (simple, mixed or multiple) distributions of mono or multivariable properties of populations,

– the formalization, or the reformulation, of the relationship between means of characterization and morphological descriptors, and where appropriate, the development of new descriptors.