摘要
分析了影响燃油炉燃烧效率的因素，提出一种基于火焰彩色图像像素分类的表征燃烧状况的指标评价函数，用于识辨燃烧状况．该函数基于贝叶斯（Ｂａｙｅｓ）统计概率分类器原型，在取得先验概率和条件概论的密度函数后，算出后验概率，最终以后验概率为判断依据，建立判断模型．从实际炉膛拍摄火焰图象选取学习样本，得到分类决策函数，并应用检验样本分析图象分类精度进而估计类别选取的合理程度．实际试验结果表明，该方法不仅能够得到与目视判别一致的结果，而且判断的可靠性更高．
The fuel combustion status evaluation function was established by analyzing the factors having effect on the efficiency of industrial furnaces. Based on the flame color image classification by Gauss classification theory and the Bayes classification rule, the density function of prior probability and conditional probability of the flame image classification can be obtained. The posterior probability can be used to establish the judge model. The learning specimen selected from the practical furnace flame image to establish the classification decision function specimen are used to analyze the classification precision for accounting the reasonable degree of the image category. The test results show that the method not only coincides with human judgement but also eliminates the man-made interference.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2001年第2期224-226,共3页
Journal of Harbin Institute of Technology
