The Modification of Building Wall Temperature Measurement Based on Infrared Thermal Image of Machine Vision

The outer wall is an indispensable core component of the building envelope and the heat transfer coefficient is tested in the field is the key link of energy saving evaluation and renovation. In this paper, the research target is limited to the cold area and cold area exterior wall.However, there are still many defects in the current field test method of heat transfer coefficient of external walls of buildings. Firstly, Based on the theoretical research of unsteady heat conduction inverse problem in building, a systematic identification method is proposed by using machine learning algorithm to identify the thermal resistance of external wall. Through theoretical analysis, the technical route of identification modeling of exterior wall thermal resistance is given, which provides sufficient research basis for building exterior wall thermal image recognition system. At the same time, the influence of variable thermal conductivity on external wall thermal resistance is studied. Then the machine learning algorithm is used to complete the establishment of the external wall thermodynamics model. In addition, the rapid extraction program of the temperature data of the effective area of the wall surface was written. Through this program, the effective wall temperature information after the exclusion of foreign bodies could be obtained, and the visualization of the infrared data and the effective area image after calibration was realized.