Cattle detection and counting using unmanned aerial vehicles

• Grant number: 18/12845-9
• Support Opportunities: Research Grants - eScience and Data Science Program - Regular Program Grants
• Start date: February 01, 2019
• End date: January 31, 2021
• Field of knowledge: Physical Sciences and Mathematics - Computer Science
• Principal Investigator: Jayme Garcia Arnal Barbedo
• Grantee: Jayme Garcia Arnal Barbedo
• Host Institution: Embrapa Informática Agropecuária. Empresa Brasileira de Pesquisa Agropecuária (EMBRAPA). Ministério da Agricultura, Pecuária e Abastecimento (Brasil). Campinas , SP, Brazil
• Associated researchers: Andréa Roberto Bueno Ribeiro ; Inácio Henrique Yano ; Luciano Vieira Koenigkan ; Maria Fernanda Moura ; Patricia Menezes Santos ; Romis Ribeiro de Faissol Attux ; Thiago Teixeira Santos

Abstract

Monitoring livestock population is an essential part of the farm management. However, this may not be a trivial task, especially in large properties adopting extensive stockbreeding, which is common in countries like Brazil. In this context, aerial surveys arise as a potential solution. Satellite images are not well suited for this task because most sensors do not have enough spatial resolution to resolve individual animals, and cloud contamination can obscure features of interest. Using manned aircraft for surveying cattle farms, although technically feasible, has a number of drawbacks associated, such as operation costs, noise levels, and risk of accidents. Unmanned aerial systems (UAS), although often pointed out as a potential solution for the problem, are still not being successfully used with this purpose. In this context, this project aims at investigating technical and practical factors that prevent a more effective use of UAS for livestock monitoring, having two goals in mind: 1) to develop a methodology for surveying livestock using UAS equipped with visible spectrum cameras; 2) to propose a new algorithm to automatically recognize and count cattle in the images captured using the suggested methodology. Two distinct approaches will be investigated, the first using a conventional camera and the second using a 360 degree camera. The main difference between both methodologies is that the first usually requires multiple flights to cover large areas, while the second may be able to image much larger regions with a single flight, at the cost of losing resolvability, as animals are farther away from the camera. In order to achieve the project's goals, the team of researchers includes people with vast experience in image processing, machine learning, unmanned aerial vehicles, software development and animal behavior. The project will produce advancements in computer science, in the form of a new image-based algorithm for animal detection, and in animal management, in the form of a new methodology for cattle counting using UAS. This methodology may also be used in the future to address other aspects of cattle monitoring, such as animal health, body measurements, etc. (AU)

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

ARNAL BARBEDO, JAYME GARCIA; KOENIGKAN, LUCIANO VIEIRA; SANTOS, THIAGO TEIXEIRA; SANTOS, PATRICIA MENEZES. A Study on the Detection of Cattle in UAV Images Using Deep Learning. SENSORS, v. 19, n. 24, DEC 2 2019. (18/12845-9)

ARNAL BARBEDO, JAYME GARCIA; KOENIGKAN, LUCIANO VIEIRA; SANTOS, PATRICIA MENEZES. Cattle Detection Using Oblique UAV Images. DRONES, v. 4, n. 4, DEC 2020. (18/12845-9)

ARNAL BARBEDO, JAYME GARCIA; KOENIGKAN, LUCIANO VIEIRA; SANTOS, PATRICIA MENEZES; BUENO RIBEIRO, ANDREA ROBERTO. Counting Cattle in UAV Images-Dealing with Clustered Animals and Animal/Background Contrast Changes. SENSORS, v. 20, n. 7, APR 2020. (18/12845-9)