【摘要】 研究如何利用计算机图像处理与模式识别技术建立一套实时监测系统,从而有效地对锅炉结焦状态图像进行处理和识别。通过结焦图像处理系统获取位图格式的结焦图像,从中提取用于结焦状态诊断的特征量。设计了一个最小距离分类器,对其用大量样本进行训练和测试,并将待识别图像的特征向量作为分类器的输入量对其进行结焦状态的诊断。更多还原

【Abstract】 How to diagnose thestates of coke by a monitoring system isintroduced, whichisconstructedby using digitalimage processing and pattern recognition technique. From the image system a BMP coke image is derived, and its characteristics are extracted. Then a minimal distance classifier is designed. Lots of samples are used to train and test the classifier. Therefore the classifier can be used to diagnose the states of the coke image by the characteristic vector of the coke image.更多还原