一种基于神经网络算法的高温测量方法

A Kind of High Temperature Measurement Method Based on Neural Network Algorithm

在高温测量的过程中,被测物体光谱发射率稳定性差,采用传统算法进行高温测量,在建立测温模型时,存在测温精确度降低的问题。为解决上述问题,提出了一种采用神经网络算法的颜色法高温测量系统的测温方法。利用彩色数码相机对高温物体在各种温度散发的颜色光进行标定,利用图像识别技术对图像进行识别,融合变换光圈快门组合的方法将测温范围扩大到高温生产的常用温度范围,通过组建RBF神经网络测量模型和标定减少CCD光谱响应带宽和被测对象光谱发射率三方面对测温精度的影响,建立精确的颜色法高温测量系统的测温模型。实验结果证明,神经网络颜色法高温测量系统测温精确性高,适应性强。

In the process of high temperature measurement,using the traditional algorithm for high temperature measurement,the temperature measurement model is set up,the object being measured spectral emissivity of stability is poor,temperature measurement precision reduce problems. Therefore,an algorithm based on neural network is presented in the color of the high temperature measuring system of temperature measurement method. Using color digital camera to the high temperature object in all sorts of color light for calibration of temperature,using image recognition technology of image recognition,fusion transform aperture shutter combination method of measuring temperature will extend to the commonly used temperature range of producing high temperature,by a RBF neural network measurement model and calibration CCD spectral response bandwidth and the object being measured spectral emissivity of three sides facing the influence of the temperature measurement accuracy,establish the color of the precise method of high temperature measuring system of temperature measurement model. The color of the simulation experiment proves that the algorithm based on neural network method of high temperature measuring system of temperature measuring is of high accuracy and strong adaptability.