6533b828fe1ef96bd1287fe8

RESEARCH PRODUCT

Approche multi-critère pour la caractérisation des adventices

subject

description

The objective of this thesis is to develop a way to detect weeds in a field using multispectral images, in order to determine which weeds should be eliminated during the current crop cycle and more particularly at the early stages. The multi-criteria approach focuses on the spatial arrangement, the spec- tral signature, the morphology and the tex- ture of the plants located in the plots. This work proposes a method for selecting the best criteria for optimal discrimination for a given setup. Prior to the extraction of these crite- ria, a set of methods was developed in order to correct the errors of the acquisition de- vice, to precisely detect the vegetation and then to identify within the vegetation the in- dividuals on which the different criteria can be computed. For the individual detection step, it appears that leaf scale is more sui- table than plant scale. Vegetation detection and leaf identification are based on deep lear- ning methods capable of processing dense fo- liage. The introduction of these methods in a usual processing chain constitutes the ori- ginality of this manuscript where each part was the subject of an article. Concerning the acquisition device, a method of spectral band registration was developed. Then, new vege- tation indices based on artificial intelligence constitute one of the scientific advances of this thesis. As an indication, these indices offer a mIoU of 82.19% when standard in- dices ceil at 63.93%-73.71%. By extension, a leaf detection method was defined and is ba- sed on the detection of their contours, this method seems advantageous on our multis- pectral data. Finally, the best property pairs were defined for crop/weed discrimination at leaf level, whith classification performances up to 91%.