A cloud point methodology for evaluating the integrity risk of arboreal crop during field coverage of agricultural machinery

Abstract

Several countries around the world are currently facing the need to have a fully automated farming process. In the United Kingdom, for example, the government has acknowledged this need and named it farmgate. As part of such process, machinery is designed to traverse the field performing a previously given task, such as harvesting, seeding, herbicide management, among others, following an also previously given path (or way points). However, when traversing, and due to environment layout or vehicle manoeuvres, machinery might collide, damaging itself or affecting the crop’s health. To address this problem, in this work, we develop a methodology for the evaluation of the expected damage in a crop, when a path has been planned for the agricultural machinery. To this end, we use point cloud processing tools that allow us to account the hitting risks per each manoeuvre and, therefore, to make the proper corrections in the path planning process. Hence, after evaluation, we are able to minimise the damage of the crop. Our proposal is tested on an existing and publicly available dataset, named CitrusFarm dataset, using the dimensions of several commercially available tractors, such as John Deere 9R, New Holland T8.435 (76.2 cm SmartTrax), Case IH Magum^TM 400 Rowtrac, and it can be extended to other platforms. The statistical results, show for example, that for the John Deere 9R tractor on the tested field, there is a 49.71% (sequence 04 from the dataset) and 87.27% (sequence 06) of risk of severely damaging the crop, whereas New Holland T8.435 tractor shows 34.34% (sequence 04) and 55.96% (sequence 06) and Case IH Magum^TM 400 Rowtrac shows 33.17%(sequence 04) and 52.41% (sequence 06). A final case study is implemented where our approach is successfully part of the decision process in the placement of waypoints in an olive grove to minimise impacts. The later shows that our methodology can benefit machinery design and path planning for full coverage practices in agriculture.