Multi Feature Descriptors for Early Stage Detection of Mastitis using Multi-SVM Classifier

Abstract

The animal husbandry and livestock industries play key significance in the rural economy, particularly for the small and marginal farmers. Mastitis is the most frequent worldwide illness in cattle since it significantly impacts animal health, quality of milk and the economy of the nation. Early identification of mastitis is highly crucial to prevent the economic damage to the dairy farmers and dairy product business. A multi feature descriptor (Color, GLCM and Zernike moments) based mastitis detection using Infrared Thermography is suggested. Infrared thermography (IRT) is a noninvasive approach that detects skin surface temperature and several descriptors such as Color moments, GLCM features and Zernike moments are derived from obtained thermal pictures. The outcomes of suggested approach for mastitis detection are cross verified with the standard methods of mastitis detection. Experimental findings reveal that the suggested approach employing Multi-SVM classifier obtained accuracy of 99.22 percent from the gathered samples. The suggested approach is effective for monitoring individual animals in farms and diagnosing mastitis at early stage decrease the economic loss of the dairy producers.