Development of an Integrated Safety Management Model in Traditional Markets Based on Deep Learning

The roofs installed in most traditional markets in Korea are made of flammable materials, making them highly susceptible to large fires from small sparks. The enclosed structure further exacerbates the vulnerability to safety accidents. Therefore, this study proposes a unified model that employs a modular approach for accurate detection of fires and smoke. The modular method enables optimized learning for each event, as opposed to learning all events simultaneously, and maintains superior performance even with a smaller dataset. Through this study, we propose a model that can be easily implemented in aging facilities, mitigating indiscriminate fire department dispatches due to IoT sensor malfunctions