六盘水师范学院学报Issue(4)：1-5,5.DOI:10.16595/j.1671-055X.2015.04.001
基于灰度共生矩阵的烟雾图像纹理分析
Analysis of Smoke Image Texture Based on GLCM
摘要
Abstract
The accuracy of smoke detection plays an important role on the real-time detection and the early warning of fire. In order to overcome the influence of the factors such as smoke at low concentration, far from monitoring ground as well as the wind speed when the fire burning, we present a video smoke detection method based on gray level co-occurrence matrix. At first, it applies the method of background difference to extract the movement area which will be divided blocks for local information. Then, it uses gray level co-oc-currence matrix to extract the textural features of each block. Finally, it uses the smoke’s textural features (en-ergy, entropy, correlation) to train for its internal relations,completing the extraction of smoke image. The ex-periments results show that the textural features of the gray level co-occurrence matrix have good properties under the disturbance of the moving-target in similar color. Both the related experimental data and compared results show that the textural feature extraction is effective for smoke detection.关键词
烟雾图像纹理/灰度共生矩阵/纹理特征/图像分析Key words
smoke image texture/GLCM/texture features/image analysis分类
信息技术与安全科学引用本文复制引用
邓兴,李金兰,郁钟铭..基于灰度共生矩阵的烟雾图像纹理分析[J].六盘水师范学院学报,2015,(4):1-5,5.基金项目
贵州省科学技术基金项目(NO.黔科合J字LKLS[2013]01号)；贵州民族大学2014年科研基金资助项目。 （NO.黔科合J字LKLS[2013]01号）
