Eventos Anais de eventos
COBEM 2021
26th International Congress of Mechanical Engineering
Autonomous navigation within corn plantation using computer vision
Submission Author:
Gabriel Araujo , BA , Brazil
Co-Authors:
Gabriel Araujo, Vitor Akihiro Hisano Higuti, Marcelo Becker
Presenter: Gabriel Araujo
doi://10.26678/ABCM.COBEM2021.COB2021-0806
Abstract
Food shortages resulting from the population increase is a projection that have been of concern to a large number of researchers, generating a demand for more efficient methods of increasing food production while also addressing the concerns of environmental footprint. To increase this efficiency, the automation of the field has been a trending research area, however it has been limited only to the phases of direct interaction with the plants. Since it is still a heavily manual task, the crop scouting and monitoring is a time-consuming and labor intensive job, which makes it expensive, inefficient and often overlooked for the aforementioned reasons. In the quest to increase efficiency and reduce costs and time, the application of autonomous robots for phenotyping research and monitoring of plantations is a very attractive solution. In view of this, this research investigates a method that makes a small robot capable of navigating within rows of a corn crop, using computer vision. The proposed method makes the positioning calculations from images treated by filters and color separation, binarization, edge detection and Hough transform, where the latter returns the possible lines to be chosen by the code as the desired path. Once the desired path is obtained, it will be returned to the integrated computer, which will be running ROS(Robot Operating System), to guide the robot back to the desired path. The project stands out from the others by presenting navigation within the trails, as opposed to most researches that feature navigation over the top of the crops. The expected results are that of a system with greater precision than a GPS system, since the latter can present flaws within plantations in more advanced growth states, and must be, at least, comparable to navigation by a LIDAR system, current baseline for under canopy navigation in corn crops. The images used by the system to guide can also be used to study the phenotypes presented by the plants, making the system more efficient. The research is part of a larger project that seeks to build a robust full path navigation system.
Keywords
Computer Vision, Autonomous Navigation, Corn Plantation
