Classification of carcass is conducted by trained and well prepared personal. However, it is criticized for being subjective and unreliable. Farmers fear on mistakes that may occur, so they do not trust much in such system. A lower carcass classification (one or two degrees) has a high impact in the received price. Automatization of this process with non-invasive system with an objective classification can be a solution for this problem. A pilot test was implemented in one slaughterhouse in Masovia Region, Poland.
There are currently very few system that collect technological solutions based on sensor systems and image analysis. For this reason, a non-invasive integral system has been proposed that can obtain a classification with a high degree of precision. A simple sensor system capable of measuring both the temperature and the humidity of the carcass
Environment was deployed in the meat treatment plant itself, as well as monitoring its weight throughout the emptying and cooling process. This system is equipped with a high-resolution camera and an advanced image processing device that can apply
Image processing in three phases. The first phase, correction and pre-processing of the image; the second phase, identification of the Landmark points for the segmentation of the carcass; and a final step of estimation of the fat level of the carcass. After the results were obtained from a set of real beef carcasses, it is possible to conclude that the system can predict, with high precision, the results of the classification made by a set of 6 experts. In the time that a human expert would take to classify a total of 40 pieces, the system can accurately classify up to 10 times more pieces and send the results in real-time to the control system
Farmers received fair and relevant prices to the meat quality they delivered to the slaughterhouse. It has impact on the economic situation of beef farms. Secondly most farmers were present in person during the classification made by human, here after one or two deliveries they convinced that classification is made correctly and they stay for results on the farm. the result were available to the farmers online.
Farmer's comment
It was something new receiving data of classification process automatically by mail. I was a little bit afraid that the machine also can make mistakes, but after two deliveries it flew away. It’s fair classification system with very good communication to farmer.