Automated Evaluation System of Stress in Cattle
- Publication Type:
- Thesis
- Issue Date:
- 2020
Open Access
Copyright Clearance Process
- Recently Added
- In Progress
- Open Access
This item is open access.
Stress in animals can be defined as internal and external body response to environmental effects. Stressed cattle after slaughtering produce meat with less than the normal amount of glycogen with undesirable taste and colour, which is called 'dark meat’. Dark meat is not appropriate for human consumption which causes massive economic losses in the global meat industry. Cattle under stress exhibit high metabolic rate, heart rate, respiration rate, and skin temperature. These physiological changes have been used in many studies as stress indicators for measuring stress in cattle before slaughtering such as blood tests and rectal temperature. These current measurement methods are invasive and are considered as wasting time and effort when dealing with a very large number of cattle on a farm. Several researchers used Infrared thermography technology (IRT) as an alternative non-invasive method for detecting stress pre-slaughtering through measuring temperature for eyes region. So far little research has been carried out to detect pre-slaughter stress automatically in cattle with dark-meat prediction. Therefore, the main aim of this research is to develop a new fully automated system for detecting stress pre-slaughtering and predicting dark meat.
Multi-view face detection in cattle with enhancement of the accuracy of detection rate. Multi-view face detection is achieved through using three Support Vector Machine (SVM) classifiers, which are established by using Histogram Oriented Gradient (HOG) as features and SVM for classification. Detected face is used as the Area of Interest for eyes segmentation.
A novel segmentation approach to automatically identify the eyes of cattle regardless of the position of the animal in relation to the camera. This proposed novel method includes foreground identification using edge difference. A new method for thresholding based on histogram processing is also proposed. After eye segmentation, eye localization and temperature measurement will be the last stage of the proposed method.
Lastly, Support Vector Machine (SVM), Logistic Regression (LR), Naïve Bayes (NB), Decision Tree (DT) are developed as machine learning algorithms for stress assessment and predicting dark-cutting. The results show that the proposed system based on the Decision Tree model can be used to detect stress with a dark-meat prediction with significant accuracy in term of Specificity, Recall and F-measure.
Please use this identifier to cite or link to this item: