基于数字图像处理的二维温度场测量

MEASURING TWO-DIMENSIONAL TEMPERATURE FIELD BASED ON DIGITAL IMAGE PROCESSING TECHNIQUE

研究了CCD摄像机进行二维温度场测量问题。分析了彩色CCD摄像机测量机理,推导和分析了相应的三色温度测量模型。由于测量模型的复杂性和非线性,因此利用BP神经网络在函数逼近方面的特性,提出了一种用BP神经网络计算二维温度场方法。试验表明这种方法是可行的、并有较高测量精度,具有工业应用前景。

The problem of measuring two-dimensional temperature field by CCD video camera is researched. The measurement mechanism of CCD color video camera is analyzed, and tri-color thermometry model is developed and analyzed. Because of complexity and non-linear of the model, a method of calculating two-dimensional temperature field by BP neural network is developed, which make use of the characteristic of BP neural network in function approximation. Experiments indicated the method is feasible and has a high degree of accuracy. The method has bright prospect in industrial application.