Avaliação da composição e da qualidade da carcaça e da carne de animais de produção: utilização de metodologias não invasivas e não destrutivas
Ficheiros
Data
2017-10-25
Título da revista
ISSN da revista
Título do Volume
Editora
Resumo
Na última década tem havido uma crescente preocupação com a obtenção de informação
sobre a composição e qualidade das carcaças, de peças e da carne. Esta preocupação está
relacionada com a possibilidade de desenvolver um sistema de pagamento dos produtos com
base no seu valor (do inglês value-based payment and marketing systems). Para isso têm sido
desenvolvidas e aplicadas técnicas não invasivas e não destrutivas sobretudo baseadas na
imagem. É neste panorama que se enquadra a presente tese e no qual será dado destaque às
técnicas de ultrassonografia em tempo real (do inglês real time ultrasonography- RTU) e
análise de imagem vídeo (do inglês vídeo image analysis- VIA) aplicadas à estimativa da
composição e qualidade da carcaça e da carne de bovinos, suínos e ovinos. O presente trabalho
está organizado em três partes.
Na Parte I foi elaborada uma breve revisão sobre a utilização da RTU para estimar a
gordura intramuscular (do inglês intramuscular fat- IMF), seguida de trabalho experimental em
que foram obtidas imagens de RTU de amostras do músculo Longissimus thoracis et lumborum
(LTL) de bovinos(n=54), usando um scanner de RTU equipado com uma sonda de 7,5 MHz. A
IMF química foi determinada em laboratório. As imagens RTU foram submetidas à análise de
imagem por computador para determinar as características das partículas de marmoreado.
Foram realizadas regressões simples e múltiplas entre a IMF química (Y) e as características
das partículas de marmoreado determinadas por RTU (X). Com esta metodologia foi possível
explicar 93% da variação da IMF do músculo LTL.
A Parte II contém uma revisão de literatura sobre a aplicação de sistemas VIA em
carcaças seguido de trabalhos experimentais. Relativamente aos capítulos envolvendo o sistema
VIA, foram utilizados dois trabalhos. Um com trinta cordeiros da raça autóctone Churra Galega
Mirandesa e outro com cinquenta e cinco carcaças de cordeiros da raça Churra da Terra Quente.
Em ambos os trabalhos foram obtidas imagens usando uma câmara digital com sensor de 8
megapixels. Foram registadas medições de área (A), perímetro (P), comprimento (L), largura
(W) e ângulo (Ang). Foram realizadas análises de regressão passo a passo (Stepwise) para
estimar a composição da carcaça em tecidos e peças para ambos pesos e rendimentos usando
as medições VIA e o peso de carcaça fria (PCF) como variáveis independentes. Nestes trabalhos foi possível verificar que é possível explicar a maior parte (R2>0,80) da variação do peso das
peças e da composição das carcaças de cordeiros leves. Todavia, para a estimativa de
rendimentos ou proporções verificou-se que os resultados não foram tão expressivos. O melhor
modelo para explicar a variação do peso do músculo incluiu o PCF e uma medição VIA e
explica cerca de 98% dessa variação. Os modelos desenvolvidos para estimar o rendimento dos
tecidos na carcaça utilizando o PCF e medições VIA têm precisão moderada, com valores de
R2 variando entre 0,472 a 0,731 (P <0,01).
Na Parte III foi elaborada uma breve revisão sobre a utilização do VIA para avaliar a
IMF, acompanhado de trabalhos experimentais em que se estima a IMF e características físicoquímicas
da carne de bovino e ainda se analisa a cor do músculo LTL de bovinos e de suínos.
Para estimar a IMF e as características físico-químicas da carne foram determinadas as
características físico-químicas como a mioglobina, capacidade de retenção de água,
comprimento dos sarcómeros, força de corte (WBSF), colagénio total, parâmetros de cor, pH
final e a IMF em 106 amostras do músculo LTL de bovinos mestiços. Coeficientes de correlação
e de determinação foram estabelecidos entre as variáveis obtidas por um sistema VIA simples
e as características físico-químicas da carne bem como com a IMF. Com a IMF obtida por
imagem e as características físico-químicas da carne e a IMF foi possível explicar a IMF e a
WBSF em 92% (DPR= 0,45; P< 0,001) e 82% (DRP= 0,63; P<0,001), respectivamente. Para
avaliar a relação entre os valores de cor da carne de bovinos e suínos no espaço CIELab foram
utilizadas 89 amostras do músculo LTL de bovinos e 110 amostras LTL de suínos. As
características colorimétricas L* (luminosidade), a* (vermelho) e b* (amarelo) foram medidas
através de dois processos: por colorimetria, com um colorímetro Minolta CR-10 e com um
sistema de visão assistida por computador (VAC). Relativamente aos parâmetros de cor, as
equações de regressão foram sempre maiores para os bovinos em todos os parâmetros avaliados
com valores de R2 variando entre 0,771 e 0,911. Os resultados observados nesta tese sugerem
que a RTU associada a análise de imagem foi uma ferramenta valiosa para estimar a IMF
química do músculo LTL sem destruir a costeleta; e que os resultados envolvendo o sistema
VIA sustentam as conclusões de outros trabalhos sobre a capacidade do sistema VIA em estimar
os cortes primários, a composição da carcaça em tecidos e a gordura intramuscular.
In the last decade, there has been a growing concern with obtaining information on the composition and quality of carcasses, cuts, and meat. This concern is related to the possibility of developing a value-based payment and marketing systems. For this, non-invasive and nondestructive techniques, mainly based on the image, have been developed and applied. It is in this panorama that the present thesis is framed and in which the real-time ultrasonography (RTU) and video image analysis (VIA) techniques were applied to the prediction of composition and quality of the carcass and meat quality of bovine, porcine and ovine species. This thesis is organized into three parts. In Part I, a brief review of the use of RTU to predict intramuscular fat (IMF) was made. This review is followed by experimental work in which RTU images were obtained from samples of the Longissimus thoracis and lumborum (LTL) muscle from bovine (n = 54) using an RTU scanner equipped with a 7.5 MHz probe. The RTU images were submitted to computer image analysis to determine the characteristics of the marbling particles. Simple and multiple regressions were performed between the chemical IMF and the characteristics of the marbling particles determined by RTU. With this methodology, it was possible to explain 93% of the IMF variation in the LTL muscle. Part II contains a literature review on the application of VIA systems to carcasses followed by experimental work. For the chapters involving the VIA system, two papers were used. One with thirty lamb carcasses of the autochthonous breed Churra Galega Mirandesa and the other with fifty-five carcasses of lambs of the breed Churra of Terra Quente. In both works images were obtained using a digital camera with a sensor of 8 megapixels. Measurements of area (A), perimeter (P), length (L), width (W) and angle (Ang) were recorded. Stepwise regression analyzes were performed to estimate carcass composition in tissues and cuts for both weights and yields using VIA measurements and cold carcass weight (CCW) as independent variables. In these studies, it was possible to verify that it is feasible to explain the variation of the weight of the cuts and the carcass composition of light lambs (R2 > 0.80). However, for the prediction of proportions, it was found that the results were less expressive. The best model to explain muscle weight variation included the CCW, and a VIA measurement explained about 98% of that variation. The models developed to predict yields in the carcass using CCW and VIA measurements have moderate accuracy, with R2 values ranging from 0.472 to 0.731 (P <0.01). In Part III, a brief review of the use of the VIA to evaluate the IMF, together with experimental work evaluating the IMF and the physical-chemical characteristics of the beef, was carried out, with another study with concerning the LTL muscle color of bovine and swine. To predict the IMF and the physical-chemical characteristics such as myoglobin, water holding capacity, sarcomere length, shear force (WBSF), total collagen, color parameters, final pH using a VIA approach it was used 106 LTL muscle samples from crossbred cattle. Correlation and determination coefficients were established between the variables obtained by VIA system and the physical-chemical characteristics of the meat as well as with the IMF. With the VIA approach it was possible explain 92% (DPR = 0.45, P <0.001) and 82% (DRP = 0.63, P <0.001), of IMF and the WBSF respectively. To evaluate the relationship between the color values of bovine and swine meat in the CIELab space, 89 LTL muscle samples from bovine and 110 LTL swine samples were used. The colorimetric characteristics L* (luminosity), a* (red) and b* (yellow) were measured through two processes: by colorimetry, with a Minolta CR-10 colorimeter and using simple a computer assisted vision system (VAC). Regarding the color parameters, the regression equations were always higher for bovine LTL in all parameters evaluated with R2 values ranging from 0.771 to 0.911, whereas for swine the results observed were less accurate. In an overall conclusion, the image approaches (RTU and VIA) allows good results in the prediction of carcass, cuts and meat composition and quality of different species. Also, the results observed in this thesis suggest that the simple and low-priced RTU and VIA approaches used associated with image analysis were valuable tools to predict the carcass, cuts and meat quality and composition in low input production systems adding a value of the products.
In the last decade, there has been a growing concern with obtaining information on the composition and quality of carcasses, cuts, and meat. This concern is related to the possibility of developing a value-based payment and marketing systems. For this, non-invasive and nondestructive techniques, mainly based on the image, have been developed and applied. It is in this panorama that the present thesis is framed and in which the real-time ultrasonography (RTU) and video image analysis (VIA) techniques were applied to the prediction of composition and quality of the carcass and meat quality of bovine, porcine and ovine species. This thesis is organized into three parts. In Part I, a brief review of the use of RTU to predict intramuscular fat (IMF) was made. This review is followed by experimental work in which RTU images were obtained from samples of the Longissimus thoracis and lumborum (LTL) muscle from bovine (n = 54) using an RTU scanner equipped with a 7.5 MHz probe. The RTU images were submitted to computer image analysis to determine the characteristics of the marbling particles. Simple and multiple regressions were performed between the chemical IMF and the characteristics of the marbling particles determined by RTU. With this methodology, it was possible to explain 93% of the IMF variation in the LTL muscle. Part II contains a literature review on the application of VIA systems to carcasses followed by experimental work. For the chapters involving the VIA system, two papers were used. One with thirty lamb carcasses of the autochthonous breed Churra Galega Mirandesa and the other with fifty-five carcasses of lambs of the breed Churra of Terra Quente. In both works images were obtained using a digital camera with a sensor of 8 megapixels. Measurements of area (A), perimeter (P), length (L), width (W) and angle (Ang) were recorded. Stepwise regression analyzes were performed to estimate carcass composition in tissues and cuts for both weights and yields using VIA measurements and cold carcass weight (CCW) as independent variables. In these studies, it was possible to verify that it is feasible to explain the variation of the weight of the cuts and the carcass composition of light lambs (R2 > 0.80). However, for the prediction of proportions, it was found that the results were less expressive. The best model to explain muscle weight variation included the CCW, and a VIA measurement explained about 98% of that variation. The models developed to predict yields in the carcass using CCW and VIA measurements have moderate accuracy, with R2 values ranging from 0.472 to 0.731 (P <0.01). In Part III, a brief review of the use of the VIA to evaluate the IMF, together with experimental work evaluating the IMF and the physical-chemical characteristics of the beef, was carried out, with another study with concerning the LTL muscle color of bovine and swine. To predict the IMF and the physical-chemical characteristics such as myoglobin, water holding capacity, sarcomere length, shear force (WBSF), total collagen, color parameters, final pH using a VIA approach it was used 106 LTL muscle samples from crossbred cattle. Correlation and determination coefficients were established between the variables obtained by VIA system and the physical-chemical characteristics of the meat as well as with the IMF. With the VIA approach it was possible explain 92% (DPR = 0.45, P <0.001) and 82% (DRP = 0.63, P <0.001), of IMF and the WBSF respectively. To evaluate the relationship between the color values of bovine and swine meat in the CIELab space, 89 LTL muscle samples from bovine and 110 LTL swine samples were used. The colorimetric characteristics L* (luminosity), a* (red) and b* (yellow) were measured through two processes: by colorimetry, with a Minolta CR-10 colorimeter and using simple a computer assisted vision system (VAC). Regarding the color parameters, the regression equations were always higher for bovine LTL in all parameters evaluated with R2 values ranging from 0.771 to 0.911, whereas for swine the results observed were less accurate. In an overall conclusion, the image approaches (RTU and VIA) allows good results in the prediction of carcass, cuts and meat composition and quality of different species. Also, the results observed in this thesis suggest that the simple and low-priced RTU and VIA approaches used associated with image analysis were valuable tools to predict the carcass, cuts and meat quality and composition in low input production systems adding a value of the products.
Descrição
Tese de Doutoramento em Ciência Animal
Palavras-chave
Composição da carcaça , Carne , Qualidade , Gordura intramuscular , Ultrassonografia , Ultrassonografia em tempo real , Análise de imagem vídeo , Animais de produção