Detection of Adulterated Pork Meat via Color and Chemical Characteristics

Pork in the Philippines is mostly sourced locally, and not all these meats come from well-sanitized slaughterhouses. Some carcasses of pigs that have died other than by slaughtering, called adulterated or double-dead meat, make their way to markets and get mixed up with legal and safe meat, only to be sold to unsuspecting consumers. These pose health hazards to meat handlers and consumers. One of the longtime practices in preventing the proliferation of double-dead meat in the markets is the manual inspection that analyzes the meat based on its organoleptic properties. This solution, however, is prone to limitations of subjectivity and human evaluation error. In this study, the proponents designed a system that will identify whether the meat sample is adulterated or not. The system is aimed at providing fast and accurate detection of double-dead meat using color and chemical determinants that is at least comparable to, or better than, humans. The application of the system can be utilized to replace the traditional human evaluation using electronic sensors as to avoid bias in the evaluation of the meat by different assessors as well as to eliminate the health risks on the meat inspectors and the consumers. The system utilizes three sensors: color sensor, methane sensor and pH level sensor. A microcontroller implements a logistic regression classifier. The model performs well, with an accuracy of 91.40% after validation, with an area of 0.95 under its ROC. When realized, the model has an accuracy of 93.18% compared to 84.09% of the human assessor's evaluation of the same meat samples.