Bibliographic citations
Linares, P., (2021). Integración de un sistema de visión por computador a un robot cartesiano controlado por un ordenador de placa de recursos computacionales limitados para la generación de trayectorias durante el proceso de repique de plantines de alcachofa en un entorno de interior [Tesis, Universidad Privada Antenor Orrego - UPAO]. https://hdl.handle.net/20.500.12759/7756
Linares, P., Integración de un sistema de visión por computador a un robot cartesiano controlado por un ordenador de placa de recursos computacionales limitados para la generación de trayectorias durante el proceso de repique de plantines de alcachofa en un entorno de interior [Tesis]. PE: Universidad Privada Antenor Orrego - UPAO; 2021. https://hdl.handle.net/20.500.12759/7756
@misc{renati/380953,
title = "Integración de un sistema de visión por computador a un robot cartesiano controlado por un ordenador de placa de recursos computacionales limitados para la generación de trayectorias durante el proceso de repique de plantines de alcachofa en un entorno de interior",
author = "Linares Otoya, Paulo Enrique",
publisher = "Universidad Privada Antenor Orrego - UPAO",
year = "2021"
}
In the present research, a base sequence of algorithms typical of computer vision systems was developed and its results were integrated into a Farmbot Cartesian robot, so that it was possible to automate the process of classifying/transplanting artichoke seedlings in an indoor environment (i.e. not in an open field). As a first step, a bibliographic review was carried out in order to identify the most relevant parameters when classifying an artichoke seedling according to its quality. In addition, other fellows and I made visits to nurseries specialized in the tasks of producing and quality control of horticultural seedlings and, such as Agrog´enesis, in order to obtain information related to the quality control process and classification of artichoke seedlings in La Libertad. As a second step, a bibliographic review was carried out regarding the state of the art of the use of computer vision systems in the agricultural industry, specifically those related to the quality control of horticultural plants in their early stage. Third, a camera calibration sequence was performed, using the “Zhang method” and a chess calibration pattern. Fourth, a base sequence of computer vision algorithms was developed in order to classify plants with an age of 20 days after sowing, into three classes according to quality. Fifth, the intrinsic and extrinsic parameters obtained during the camera calibration were used to locate the seedlings within the workspace of the Cartesian robot. Finally, the seedling classifying/transplanting process was executed 25 times, evaluating the performance of the classifier and the transplanting process in general in terms of time and hits. In order to evaluate the classifier, the F1-Score was considered as the main metric, obtaining values of 0.941 when classifying class A seedlings, 0.785 when classifying class B seedlings and 0.88 for class C seedlings. In addition, the percentage of success obtained during transplantation (i.e. correct handling of seedlings using the gripper) had values between 20 − 30 %.
This item is licensed under a Creative Commons License