锅炉火焰图像特征分析

Features Analysis of Boiler’s Flame Image

当今国内电站锅炉和大型工业锅炉的火焰监测直接影响钢材料生产质量,而近年来考虑到锅炉内传感器的频繁更换,人工看火的可靠性,工业生产的智能化管理等[1-2],以火焰图像为研究对象,提取火焰图像多种特征值(平均亮度、平均对比度、平滑度、第3阶矩、一致性和熵)。基于灰度共生矩阵的统计方法,通过图像预处理,图像纹理特征提取,对火焰硬图像、软图像进行对比分割实验,可有效地判别图像的硬、软。

Nowdays, in domestic power station boilers and large industrial boilers, the monitoring of the flame plays an increasingly important role in tht Steel material＇s production quality, meanwhile, considering the frequent replacement of the sensor in the boilers in recent years, the reliability of the flame of artificial observation and the intelligent management of industrial production etc, this paper makes flame image as the research object, and extracts flame image＇s six features, which are average brightness,average contrasts, smooth the third matrics, homogeneity and entropy. The statistics method based on the gray level co-occurrence matrix, through the image preprocessing, image texture feature extraction, with the experiments of comparing and separating of the hard image and the soft image of the flame ,could realize the effective discrimination of image hardness and softness.