6533b832fe1ef96bd129a055

RESEARCH PRODUCT

Échantillonnage adaptatif optimal dans les champs de Markov, application à l’échantillonnage d’une espèce adventice

subject

description

This work is divided into two parts: (i) the theoretical study of the problem of adaptive sampling in Markov Random Fields (MRF) and (ii) the modeling of the problem of weed sampling in a crop field and the design of adaptive sampling strategies for this problem. For the first point, we first modeled the problem of finding an optimal sampling strategy as a finite horizon Markov Decision Process (MDP). Then, we proposed a generic algorithm for computing an approximate solution to any finite horizon MDP with known model. This algorithm, called Least-Squared Dynamic Programming (LSDP), combines the concepts of dynamic programming and reinforcement learning. It was then adapted to compute adaptive sampling strategies for any type of MRF distributions and observations costs. An experimental evaluation of this algorithm was performed on simulated problems. For the second point, we first modeled the weed spatial repartition in the MRF framework. Second, we have built a cost model adapted to the weed sampling problem. Finally, both models were used together to design adaptive sampling strategies with the LSDP algorithm. Based on real world data, these strategies were compared to a simple heuristic and to static sampling strategies classically used for weed sampling