Monitorização automática por análise de imagem da condição corporal em vacas leiteiras
Data
2020-06-29
Autores
Título da revista
ISSN da revista
Título do Volume
Editora
Resumo
A avaliação da condição corporal é reconhecida como uma técnica essencial de gestão e maneio das vacas leiteiras. O conhecimento das reservas corporais de gordura, assume um papel determinante no maneio nutricional que, como consequência, permite prevenir distúrbios metabólicos, produtivos e de reprodução. Apesar de, estar comprovado que a avaliação da condição corporal manual por visualização, é uma técnica que apresenta benefícios, a sua utilização de forma consistente pode ser de difícil implementação. Pelo que, a constante necessidade de melhoria do desempenho das explorações leiteiras, associadas à crescente disponibilidade de soluções tecnológicas, tem levado vários autores a estudar o desenvolvimento de metodologias de classificação automática através da análise de imagem. Como tal, pretende-se com este trabalho estabelecer uma comparação entre os dados recolhidos pelo sistema comercial de avaliação da condição corporal através de imagem a três dimensões (3D) – DeLaval Body Condition Scoring (BCS)TM, e a avaliação da condição corporal de forma visual e ainda a recolha e análise dos dados do sistema de avaliação automático juntamente com software de gestão que integra o funcionamento da BCSTM - DelProTM FarmManager, no qual ficam registadas as pontuações de condição corporal de cada indivíduo numa base diária. Os dados de avaliação visual da condição corporal classificaram-se por base de uma escala de 5 pontos com intervalo de 1 ponto. Esta avaliação foi realizada nos mesmos animais que foram avaliados pela uma camara BCS DeLavalTM, na respetiva análise dos dados concluiu-se que os valores de avaliação da condição corporal de forma automática foram em média significativamente superiores (P <0,05) quando comparados com a classificação obtida de forma manual. A correlação entre a condição corporal (CC) obtida pelas duas abordagens em análise foi elevada (r= 0,73; P<0,01). Por conseguinte, na análise da relação entre os dois métodos, salienta-se a necessidade de recorrer a uma escala classificativa mais abrangente, a classificação com 1 ponto de intervalo não permite uma discriminação precisa dos resultados quando comparado ao sistema automático que pontua em notas com intervalo de 0,1. Ademais, foram ainda analisados e discutidos os dados de CC obtidos pela camara BCS DeLavalTM e a sua integração no software DelProTM FarmManager. Com os resultados obtidos pode concluir-se que a avaliação da condição corporal de forma automatizada, através da análise por sistema de imagens de três dimensões, funciona de forma precisa, fornecendo dados detalhados ao operador o que lhe permite tomar decisões atempadas para o maneio das vacas.
The assessment of body condition is recognized as an essential technique for the management of dairy cows. The knowledge of body fat reserves assumes a determining role in nutritional management, which, consequently, allows to prevent metabolic, productive and reproductive disorders. Although, it has been proven that manual body condition assessment is a simple and inexpensive technique that has benefits, its consistent use can be challenging to implement. Therefore, the constant need to improve the performance of dairy farms, associated with the increasing availability of technological solutions, has led several authors to study the development of the automatic classification of body condition through image analysis. As such, this work aimed to establish a comparison between the data collected by the commercial body condition assessment system through a three-dimensional image (3D) - DeLaval Body Condition Scoring (BCS)TM, and the visual body condition assessment. It is also objective of the present work the analysis of data obtained from the automatic assessment system DeLaval Body Condition Scoring (BCS)TM together with management software DelProTM FarmManager, in which the body condition scores of each individual are recorded daily. The visual assessment data for body condition was based on a 5-point scale with a 1-point interval. This evaluation was carried out on the same animals that were evaluated by a BCS DeLavalTM camera. Thus, in the analysis of the data, it was concluded that the values of evaluation of the body condition automatically were on average, significantly higher (P <0.05) when compared with the classification obtained visually. The correlation between body condition obtained by the two approaches under analysis was high (r = 0.73; P <0.01). In the analysis of the relationship between the two methods, the need to use a more comprehensive classification scale is highlighted, the classification with 1 point of interval does not allow precise discrimination of the results when compared to the automatic system that scores in notes with an interval of 0.1. The body condition data obtained by the BCS DeLavalTM camera and its integration in the DelProTM FarmManager software were also analyzed and discussed. With the results obtained, it can be concluded that the evaluation of the body condition in an automated way, through the analysis by a three-dimensional image system, works in a precise way, providing detailed data to the operator, which allows him to make timely decisions for the management of the dairy cows.
The assessment of body condition is recognized as an essential technique for the management of dairy cows. The knowledge of body fat reserves assumes a determining role in nutritional management, which, consequently, allows to prevent metabolic, productive and reproductive disorders. Although, it has been proven that manual body condition assessment is a simple and inexpensive technique that has benefits, its consistent use can be challenging to implement. Therefore, the constant need to improve the performance of dairy farms, associated with the increasing availability of technological solutions, has led several authors to study the development of the automatic classification of body condition through image analysis. As such, this work aimed to establish a comparison between the data collected by the commercial body condition assessment system through a three-dimensional image (3D) - DeLaval Body Condition Scoring (BCS)TM, and the visual body condition assessment. It is also objective of the present work the analysis of data obtained from the automatic assessment system DeLaval Body Condition Scoring (BCS)TM together with management software DelProTM FarmManager, in which the body condition scores of each individual are recorded daily. The visual assessment data for body condition was based on a 5-point scale with a 1-point interval. This evaluation was carried out on the same animals that were evaluated by a BCS DeLavalTM camera. Thus, in the analysis of the data, it was concluded that the values of evaluation of the body condition automatically were on average, significantly higher (P <0.05) when compared with the classification obtained visually. The correlation between body condition obtained by the two approaches under analysis was high (r = 0.73; P <0.01). In the analysis of the relationship between the two methods, the need to use a more comprehensive classification scale is highlighted, the classification with 1 point of interval does not allow precise discrimination of the results when compared to the automatic system that scores in notes with an interval of 0.1. The body condition data obtained by the BCS DeLavalTM camera and its integration in the DelProTM FarmManager software were also analyzed and discussed. With the results obtained, it can be concluded that the evaluation of the body condition in an automated way, through the analysis by a three-dimensional image system, works in a precise way, providing detailed data to the operator, which allows him to make timely decisions for the management of the dairy cows.
Descrição
Esta dissertação foi elaborada com vista à obtenção do Grau de Mestre em Engenharia Zootécnica apresentada à Universidade de Trás-os-Montes e Alto Douro
Palavras-chave
condição corporal , vacas leiteiras