摘要
井下光源不充足,机械运行振动以及大量煤尘等因素的存在导致视频监控系统实时获取的图像较模糊且参杂了大量的颗粒噪声。为改善矿井视频图像质量,将空间域直方图匹配方法引入小波变换域中,提出了一种基于小波分解子带自适应直方图匹配的图像增强方法。首先对矿井视频图像与经过直方图均衡化处理后的矿井视频图像分别进行小波变换,以直方图均衡化处理后的图像的低频小波分解子带灰度直方图为基准,将矿井视频图像的低频小波分解子带直方图与之进行匹配,获得增强后的低频小波分解子带;然后针对均值滤波方法存在的缺陷,提出了一种基于滤波窗口自适应划分的加权改进均值滤波算法,并将其应用于去除矿井视频图像高频小波分解子带中的颗粒噪声;最后对直方图匹配后的低频小波分解子带和滤波后的高频小波分解子带进行逆小波变换,得到了视觉效果较佳的矿井视频图像。利用C++语言分别对直方图均衡化、同态滤波、均值滤波以及所提方法进行了编程实现,采用实地获取的山西阳泉某煤矿井下视频监控图像进行试验,对试验结果引入边缘保持指数(Edge protection index,EPI)进行评价,结果表明,所提方法对矿井视频图像的处理效果相对于其余方法而言优势较明显。
The existing factors of inadequate underground light source,mechanical vibration and a large number of coal dusts make the real-time images obtained by the mine video monitoring system are fuzzy and mixed with a large number of granular noise. It is difficult to conduct analysis and interpretation of the mine video images. In order to improve the quality of the mine video images,the histogram matching algorithm in spatial domain is introduced into wavelet transform domain,a image enhancement method based on the histogram matching method in wavelet transform domain is proposed. Firstly,the mine image and the mine video image processed by histogram matching method are conducted wavelet transform respectively,based on the gray histogram of the wavelet low-frequency sub-band processed by the histogram matching method,the wavelet low-frequency sub-band of the mine video image is matched with it,the enhanced wavelet low-frequency sub-band is acquired; then,based on analyzing the basic principle of the average filtering method,according to the deficiencies of the average filtering method,the weighted improved improved average filtering method based on adaptive division of filtering windows is put forward,and it is applied to filtered the granular noise distributed in the wavelet high-frequency sub-bands of the mine underground video image; finally,the wavelet lowfrequency sub-band processed histogram matching method and the wavelet high-frequency sub-bands filtered by the improved average filtering method proposed in this paper are conducted inverse wavelet transform,the mine video image with perfect visual effects is obtained. Programs of the histogram equalization,homomorphic filtering,average filtering and the method proposed in this paper are obtained by C + + language respectively. The experimental data is the underground video image obtained in a mine of Yangquan city,Shanxi province,the edge protection index( EPI) is used to evaluate the effects of the above methods,the results show that,the performance of the method proposed in this paper is superior to the other methods.
出处
《金属矿山》
CAS
北大核心
2016年第6期130-133,共4页
Metal Mine
关键词
矿井视频监控系统
图像增强
小波变换
直方图均衡化
直方图匹配
均值滤波
边缘保持指数
Mine video monitoring system
Image enhancement
Wavelet transform
Histogram equalization
Histogram matching
Average filtering
Edge protection index