距离和视角对红外热像法测温影响与修正

Influence and Correction of Distance and Viewing Angle on Temperature Measurement by Infrared Thermal Imaging

红外热像仪与被测建筑表面的距离及相对视角会影响测温准确度,对此开展研究并提出了修正方法。以哈尔滨工业大学寒地楼内墙表面为测试对象,在红外热像仪分别距墙体表面1.0~7.0 m和30°~90°视角下,测试了28组工况下的墙体表面温度,同时利用热电阻温度传感器对测试期间的墙体表面温度进行监测。结果表明,红外热像仪测温值与距离、视角存在较强的非线性关系;以热电阻温度传感器测温值为基准,红外热像仪测温值随着距墙体变远、测温视角变小而逐渐降低;在不同视角下,距墙体表面1.0 m处的平均误差最小,为0.22℃,7.0 m处的平均误差最大,为1.16℃;在不同距离下,红外热像仪在墙体表面垂直方向(90°)位置处的测温平均误差最小,为0.44℃,与墙面夹角30°处的测温平均误差最大,为0.87℃;基于BP神经网络,构建了以红外热像仪测温值、视角和测温距离为输入量,热电阻温度传感器的测温值为输出量的测温修正模型。经过修正后,红外热像仪测温与热电阻温度传感器测温的误差减小至0.01℃,提高了墙体表面测温精度,对精准识别墙体热工性能有重要意义。

The distance and relative angle of view between InfraCAM and the measured building surface affect the accuracy of temperature measurement, which is investigated and a correction method is proposed. The temperature of the wall surface of the Handi building of the Harbin Institute of Technology was tested under 28 sets of working conditions with the InfraCAM at a distance of 1.0~7.0 m from the wall surface and at an angle of view of 30°~90°, respectively. The wall surface temperature was also monitored during the test period using Resistance Temperature Detector(RTD) temperature sensors. The results show that there is a strong non-linear relationship between the measured temperature values of InfraCAM and the distance and viewing angle. Taking the temperature measurement value as the base, the temperature measurement value of InfraCAM gradually decreases as the distance from the wall becomes farther and the shooting angle becomes smaller. The smallest mean error of 0.22 ℃ at 1.0 m from the wall surface and the largest mean error of 1.16 ℃ at 7.0 m at different viewing angles. InfraCAM has the smallest average error in temperature measurement at the vertical(90°) position of the wall surface at different distances, which is 0.44 ℃. Among them, the average error of temperature measurement at the angle of 30°with the wall surface is the largest, which is 0.87 ℃. According to the BP neural network, a temperature measurement correction model with InfraCAM temperature measurement value, viewing angle and distance as the input quantity and the temperature measurement value of RTD temperature sensor as the output quantity was constructed. With the correction, the error between the wall surface temperature and the actual measured value of RTD temperature sensor is reduced to 0.01 ℃, enhancing the accuracy of the wall surface temperature measurement, which is significant for the accurate identification of the thermal performance of the wall.