摘要
本文分析了彩色CCD摄像机测量机理,建立了由火焰图像求取燃烧温度场的数学模型,推导和分析了相应的三色温度测量模型,论述了图像噪声的滤除方法,分析了目前较先进的BP神经网络测温误差大的原因,并提出一种新的基于比值输入的BP神经网络法来计算温度。实验证明,该算法具有较高的计算精度和实用性。
This paper analyzed the measuring mechanism, built the mathematical model of calculate burning temperature field from flame image, deduced and analyzed corresponding three color temperature measuring model, discussed filtering method of image noise, analyzed the reason of big measuring errors when measured by BP neural network which is more advanced in present, and developed a new BP neural network method, which based on ratio input, to calculate the temperature. Experiments show that the method has more high measurement precision and practicability.
出处
《中国仪器仪表》
2006年第6期66-68,共3页
China Instrumentation
