A Novel Method for Detecting and Predicting Emerging Disease in Poultry Chickens Based on MobileNet Model

Abstract

The chicken business has a significant impact on the food manufacturing sector. Concerns over poultry birds’ quality have increased globally as demand has risen. Chicken eggs and chicken meat are both helped along by the industry’s commitment to quality control. The industry’s players are worried about the welfare of the birds because of the rising demand for poultry meat. The poultry sector is able to keep better tabs on the well-being of its chickens thanks to recent technology breakthroughs. With the use of Internet of Things (IoT)-based wearable sensing devices like accelerometers and gyro devices, avian diseases and chicken health may now be diagnosed via video surveillance, voice observations, and feces inspections. Placed atop a chicken, these motion detectors send the hen’s daily movements to the internet for examination. It’s a difficult problem to analyze such data and provide more precise forecasts regarding chicken health. In this research, we present a framework for an Internet of Things-based prediction service that can identify illnesses in chicken flocks at an early stage. The MobileNet approach has been shown to reach a 97% accuracy in both theoretical analysis and experimental findings. In addition, the suggested research compares the efficacy of several classification models to provide a more precise and top-performing classification strategy. The primary goal of the research is to provide predictive service architecture based on Industrial IoT that can more precisely categorize poultry hens in real time.