Visão computacional para identificação de entrelinhas de plantio: uma contribuição para o controle de ervas daninhas

Abstract

One of the great opportunities for national economic advantage is creating innovation projects with technology for the agribusiness sector. One of the main demands for the sector consists of using computer vision applications for autonomous vehicles. In this context, there is a need in the agricultural market to produce technology capable of identifying crop rows to guide agricultural implements that work on those rows. These implements are usually driven by operators, who are prone to errors and low accuracy. This project aims at allowing a cooperation between industry and academia and it has an important role in the development and application of high tech technologies. The main project’s goal is to develop, test, and implement a system, based on a computer vision, that should be capable of identifying the lines between of crops. In the first stage of the project, a systematic literature mapping was carried out by us in order to identify the best technology approach for the proposed context. This system will be the basis for controlling agricultural implements that are based between the lines to carry out crop management. In the context of this project, the software to be developed will be used in a weeder for organic soybeans. The results obtained show that, in addition to promoting the development of software systems of this nature, public-private partnerships such as the one carried out in this study can generate technological transfer from academia to industry.