Poster Abstract: Intelligent Heating Monitoring Method Based on Infrared Image Segmentation and Target Detection

With the development of smart cities, the information of the heating system of buildings can be highly integrated and shared. However, the current intelligent heating system adopts the scheme of installing temperature sensors in users' rooms to collect data, which has problems such as low acceptance by users, high equipment cost, difficult management, and poor robustness. To solve these problems, this paper proposes a new type of intelligent heating system: an intelligent heating monitoring method based on infrared image segmentation and target detection. An infrared cameraequipped unmanned aerial vehicle is used to acquire thermal images of buildings and create datasets. SegNet semantic segmentation model is used to semantically segment the infrared image and segment windows that can infer the indoor temperature. A portion of the indoor temperature is collected using a temperature sensor to analyze the temperature acquisition points that can represent the indoor temperature. In this way, we can achieve the purpose of inferring the indoor temperature based on the infrared images acquired through the outdoors.