Autonomous Navigation in Maize Crop Fields

Scope The autonomous navigation in agricultural environments allows also to carry out autonomous operations of special interest in agriculture environments. The goal of an autonomous navigation is to cover all the required areas where the operation has to be made. We take into account different navigation strategies using proper sensors being also able to detect references which makes possible the autonomous navigation. The operation chosen is weeding and due to its increasing importance, maize is chosen as case study. Weeding operations are very important to avoid direct damages caused in the crops due to the competition with weeds for resources like water, nutrients, light or space. The maize crop fields present a semi structured environment, whereas the crop are disposed along rows, the distance betweens crops in the same row (also between rows) is not regular in practice due to soil irregularities and sowing errors. Weeding operations are important during the first six weeks of life of maize crop (height up to 25 cm), presenting the maize different morphologies and size. Maize plants as living being will have mortality, growth differences, they can also have mud or some plant diseases which degrade its properties. As the maize is sown the crop and weeds will have the same opportunities during the first weeks. At the same time some weeds with similar morphology to the crop can also cause overlaps. A state of the art is done to identify perception systems, usuaully visionbased, and technological approaches related to the topics to be dealt [1], [2], [3], [4]. An outdoor environment implies also very variable light conditions with the existence of shadows and strong light contrasts. The work presented in this paper is in the context of the research project SAAPIN (Autonomous System for Precision and Integrated Agriculture), whose aim is to develop an autonomous system able to navigate through maize field and perform weeding operations.