基于纹理粗糙度的回转窑火焰图像FCM分割方法

在深入研究氧化铝回转窑火焰图像特点的基础之上,提出了一种将基于图像灰度值的模糊C-MEANS(FCM)算法与图像纹理粗糙度特征相结合的图像分割方法.利用加窗自相关系数表征图像中火焰区与物料区在纹理粗糙度方面的差异,对FCM聚类的结果隶属度矩阵进行去模糊化运算,改善了火焰区与物料区的分割效果.实验结果表明,图像灰度值信息和纹理粗糙度特征的融合对于提高氧化铝回转窑火焰图像的分割精度具有重要的研究价值.

FCM Segmentation for Flame Image of Rotary Kiln Based on Texture Coarseness

Based on the in-depth study on the flame images of calcining zone of the alumina rotary kiln, a new image segmentation method was proposed, which combined the standard gray scale Fuzzy C-MEANS(FCM) algorithm with the texture coarseness character of the image. This method used sliding window autocorrelation coefficient to describe the texture coarseness difference between flame area and material area, in order to discern the result membership matrix of FCM algorithm, as a result, the segmentation result of flame image was improved. Experiments prove that the combination of gray scale and texture coarseness character has important effect for flame image segmentation of alumina rotary kiln.