육계농장의 ICT장비가 생산한 빅데이터 분석을 활용한 질병발생예측모델(DOPM) 개발연구

Abstract

Since 2014, the government has promoted the ICT equipment support project for livestock farmers as an alternative system to address the rapid aging of livestock farmers and sharp decline in labor force due to population decline. This project aims to realize more precise and scientific smart livestock farming through the analysis of big data produced by ICT equipment in real time. Prior to the introduction of ICT equipment, farmers relied on their long-term experience in farm operations to implement emergency measures by using a sudden increase in the number of dead birds or sudden change in the amount of water consumption data as an indicator of the occurrence of disease. However, damage and loss to farmers have not been reduced with these measures. Therefore, this study aims to address the farmers' consistent desire to develop a disease outbreak prediction model (DOPM) that can be easily utilized in the field and has high precision. The current study developed a disease outbreak prediction model (DOPM) that can find a singularity (golden time) in the data even if only one type of ICT data (Water & Feed) is properly monitored and analyzed. We found that if the time when a sharp decrease (more than 100% decrease from the previous day), confirmed through the daily feed intake data for two consecutive days (golden time = DOPM), is set as the start of the disease outbreak and preventive measures are taken quickly, this model can be used to minimize farm losses and damages better than the existing method that relies on analog data and experience. Additionally, the DOPM was confirmed to be the fastest, most sensitive, and most accurate model in the field poultry farms. As a result, we have found a method to interpret non-verbal signals of the chickens with the transformed ICT-Data in the AI-age.