HYBRID METHOD FOR CATTLE RECOGNITION USING UNIQUE IRIS ID NUMBERS

Abstract

Cattle identification plays a pivotal role in various agricultural and veterinary applications, ranging from livestock management to disease control. Traditional methods of identification, such as ear tags and branding, often pose challenges in terms of reliability, longevity, and human intervention. In response, biometric identification systems have emerged as promising alternatives, with iris recognition standing out as a particularly robust and accurate modality. This thesis presents a comprehensive exploration of iris recognition technology tailored specifically for cattle. Leveraging advancements in computer vision, machine learning, and biometrics, the research proposes a hybrid method that combines feature extraction techniques, deep learning architectures, and unique ID generation algorithms. Through extensive experimentation and validation on diverse datasets, the model demonstrates superior accuracy and reliability in identifying individual cattle based on their iris patterns. Furthermore, the thesis focuses on the system’s consistency in generating the same unique ID number for different images of the same individual. This aspect enhances the accuracy of identification. In conclusion, this thesis contributes to the growing body of knowledge in agricultural biometrics and sets a foundation for the future development and adoption of intelligent and responsible identification technologies for livestock management and beyond.