中图法分类号tk223tn911文献标识码a文章编号论文引用.doc

中图法分类号：TK 223, TN 911 文献标识码：A 文章编号： 论文引用格式： Log-Gabor小波和分数阶多项式KPCA的火焰图像状态识别 宋昱1 ，吴一全1,2 1.南京航空航天大学电子信息工程学院，南京 210016；2.华中科技大学煤燃烧国家重点实验室，武汉 430074 摘 要：目的：为了进一步提高锅炉燃烧火焰图像状态识别的性能，提出了一种基于Log-Gabor小波和分数阶多项式核主成分分析(Kernel Principal Component Analysis, KPCA)的火焰图像状态识别方法。方法：首先利用Log-Gabor滤波器组对火焰图像进行滤波，提取滤波后图像的均值和标准差，并构成纹理特征向量。然后使用分数阶KPCA方法对纹理特征向量进行降维，并将降维后的纹理特征向量输入支持向量机进行分类。结果：实验结果表明，提出的方法能够准确地提取火焰图像纹理特征，与2种基于Gabor小波特征提取的方法相比，本文方法的分类识别正确率更高，分类精度为76%。同时，第1主分量方差比重与核函数参数d之间满足递增关系。结论：本文提出了一种对锅炉燃烧火焰图像进行状态识别的方法，对提取的火焰图像纹理特征向量进行降维并进行分类，可以获得较高的分类精度。实验表明本文提出的方法分类精度较高，运行时间较短，具有良好的实时性。燃烧监测；火焰图像；Log-Gabor小波；核主成分分析；支持向量机 State identification of boiler combustion flame images based on Log-Gabor wavelet and kernel principal component analysis with fractional power polynomial models SONG Yu1, WU Yi-quan1,2 1. School of Electronic and Information Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, 210016; 2. State Key Laboratory of Coal Combustion (Huazhong University of Science and Technology), Wuhan, 430074 Abstract: Objective: To improve the performance of state identification of the boiler combustion flame images, a state identification method based on the Log-Gabor wavelet and kernel principal component analysis (KPCA) with fractional power polynomial models was proposed. Method: First the flame images were filtered by the Log-Gabor filter bank and the texture feature vectors of the images were constructed using the mean and standard deviation of the filtered image. Then KPCA with fractional power polynomial models was used to reduce the dimension of the texture feature vectors and the dimension reduced texture feature vectors were classified by the support vector machine. Result: Experimental results show that the method proposed can accurately extract the texture features of the flame images. Compared with the feature extraction method based on the Log-Gabor wavelet and 2 other feature extraction methods based on the Gabor wavelet, the classification rat