The recent advances in sensor technologies and data analysis could improve our capacity to acquire long term and individual dataset on animal behavior. This is particularly interesting when behavioral data could be linked to zootechnical, physiological or genetical information, with the objective of improving animal management. In this article we proposed a framework, based on computer vision and deep-learning, to automatically estimate animal location inside pasture. We illustrated our framework for the monitoring of grazing goats. We were able to detect, in average, 87.95% of the goats and to identify the detected individuals with an average sensitivity of 94.9% and an average precision of 94.8%. Goats were allowed to graze an experimental plot, where infected feces with gastro-intestinal nematodes were previously dropped in delimited areas. Four animals were monitored, during two grazing weeks on the same pasture (Try 1 and Try 2), spaced from more than 2 months. Using the monitoring framework, we were able to study different aspect of animal behavior, relating to parasitology. First, we monitored the ability of the animal to avoid feces on pasture, and showed an important temporal and individual variability. Interestingly, the avoidance capacity of all animals increased during the second grazing week (Try 2), and the level of increase was correlated to the level of infection during Try 1. We also studied the relationship between the time spent on contaminated areas with the level of infection, and was not able to find clear relationship. We characterized social behavior using the inter-individual distance, but again, we were not able to find a link with the level of infection. Due to the low number of studied animals, biological results have to be interpreted with caution, but our framework can be used to explore the relationship between behavior and parasitism in routine experimentations.