Classifying Chewing and Rumination in Dairy Cows Using Sound Signals and Machine Learning

Area of Science:

• Animal Science
• Agricultural Engineering
• Bioacoustics

Background:

• Understanding dairy cattle feeding behavior is crucial for optimizing nutrition and productivity.
• Jaw movements during feeding provide key insights into dietary intake and feed processing.
• Accurate classification of jaw movements can help identify nutritional deficiencies and improve feeding strategies.

Purpose of the Study:

• To introduce a novel methodology for classifying dairy cattle jaw movements.
• To analyze jaw movements during feeding on different particle sizes of wheat straw and Alfalfa hay.
• To evaluate the effectiveness of various machine learning classifiers for this task.

Main Methods:

• Sound signals of jaw movements were recorded and transformed into images using short-time Fourier transform.
• Texture features were extracted using established image analysis methods (GLCM, SGLDM, GLRLM, GLDM).
• Genetic Algorithm (GA) was used for feature selection, followed by classification using six distinct algorithms (Naive Bayes, k-NN, SVM, Decision Tree, MLP, k-Means).

Main Results:

• The study successfully classified four distinct jaw movement categories: bites, exclusive chews, chew-bite combinations, and exclusive sorting.
• High classification precisions were achieved across all tested classifiers, with the Support Vector Machine (SVM) reaching 95.9%.
• The methodology demonstrated the potential for accurate analysis of cattle feeding behavior based on acoustic signals.

Conclusions:

• The developed methodology offers a valuable tool for livestock managers to assess dairy cattle nutrition and feeding practices.
• Accurate classification of jaw movements enhances understanding of dietary patterns, aiding in the identification of health issues or deficiencies.
• This approach can lead to improved feeding strategies, reduced waste, and improved overall dairy cattle well-being and productivity.