Abnormal Behavior Analysis for Surveillance in Poultry Farms using Deep Learning

The poultry farming sector plays a vital role in supplying sustenance to an expanding population. Nevertheless, the birds' well-being remains a crucial issue, as inadequate living conditions result in abnormal behaviors that impact the productivity and health of the entire crowd. In order to improve and sustain the health of the birds, an automated surveillance system that monitors and analyzes chickens behaviors needs to take in place. The study of abnormal behavior analysis for surveillance in poultry farms is of essential importance in ensuring the health, welfare, and productivity of the birds. This research aims to develop a comprehensive understanding of the various factors contributing to abnormal behavior patterns and the methods for effectively monitoring and detecting these behaviors. By identifying and addressing issues related to illness, stress, or discomfort at an early stage using the proposed system, farmers can implement targeted interventions to improve the overall well-being of the birds, leading to enhanced production efficiency and profitability. Furthermore, this research contributes to the development of sustainable poultry farming practices, protecting public health, and safeguarding food safety, highlighting the significance of abnormal behavior analysis in the poultry industry. This paper proposes a computer vision based system that monitors and analyzes the chickens' behaviors in poultry farms. The system takes video input and segments them into 10-second segments. The proposed system achieves an accuracy of 96.43% using a convolutional neural network for heatmaps classification.