基于 MHSW 特征融合的火焰检测-《计算机技术与发展》

文章信息/Info

作者:: 张健¹ ; 钟中志² ; 柯艳国¹ ; 凡远柱³; 1. 国网安徽省电力有限公司,安徽合肥 230061; 2. 安徽大学电子信息工程学院,安徽合肥 230039; 3. 安徽南瑞继远电网技术有限公司,安徽合肥 230088

Author(s):: ZHANG Jian¹ ; ZHONG Zhong-zhi² ; KE Yan-guo¹ ; FAN Yuan-zhu³; 1. State Grid Anhui Electric Power Co. ,Ltd. ,Hefei 230061,China; 2. School of Electronic Information Engineering,Anhui University,Hefei 230039,China; 3. Anhui Nanrui Jiyuan Power Grid Technology Co. ,Ltd,,Hefei 230088,China

摘要:: 为了提高火焰特征使用效率和对光照强度变化的鲁棒性,结合 MHSW(Maximal HSV and SILBP of Windows)和圆形度特征融合给出了一种基于图像的火焰检测算法。该算法对图像进行超像素分割,并采用 YCbCr 颜色空间对超像素进行处理,分割出疑似火焰区域;然后对疑似火焰区域图像提取 MHSW 和圆形度特征,其中 MHSW 特征是同一水平内局部窗口中两个 SILBP(尺度不变的局部三元模式)统计直方图和 HSV 颜色直方图对应模式最大值组合而成;最后融合 MHSW特征和圆形度特征,并使用交叉网络搜查法的 SVM 实现火焰的识别。 MHSW 特征减少火焰特征使用的冗余性和火焰特征训练识别的复杂性,降低火焰识别的误检率;MHSW 特征对噪声和光照强度变化具有鲁棒性。实验结果证明,该算法是有效的,且具有较高的识别率和较低的误减率。

Abstract:: In order to improve the efficiency of flame feature and robustness to light intensity changes,a flame detection algorithm based on image is presented,which combines? MHSW (Maximal HSV and SILBP of Windows) and circularity feature fusion. The algorithm is used to perform super-pixel segmentation on the image and process the super-pixels in YCbCr space to segment the suspected flame region. Then MHSW feature and circularity feature are extracted from the suspected flame region images. The MHSW feature is a combination of two SILBP (scale invariant local ternary pattern) statistical histograms and HSV color histogram corresponding mode maximum values of local windows in the same level. Finally,by means of SILBP and circularity feature fusion,the SVM of cross network search method is used to identify the flame. The MHSW feature reduces the redundancy of flame feature and the complexity of flame feature training and recognition,and decreases the false detection rate of flame recognition. The MHSW feature is robust to noise and light intensity changes. The experiment shows that the algorithm is effective and has a high recognition rate and a low false negative rate.