矿井视频监控图像改进模糊中值滤波算法

由于矿井空气粉尘较多且光照不均匀，导致井下视频监控系统获取的图像含有大量噪声且明暗不均，在一定程度上影响了对井下生产状况的实时有效监控。为此，提出了一种改进模糊中值滤波算法。首先设计了一种自适应改进模糊隶属度系数计算方法，该方法将图像滤波窗口内像素点的灰度最小值、最大值、均值等作为阈值来计算不同像素点灰度值的模糊隶属度系数，克服了经典模糊隶属度系数计算时需设置大量阈值的不足；然后根据模糊隶属度系数进行模糊加权中值滤波，并对滤波后图像的灰度直方图进行了均衡化处理。采用C#语言编写算法运算程序，试验数据为2幅某矿井视频监控图像，试验中引入了模糊中值滤波、加权中值滤波、开关中值滤波等算法进行对比分析，并对各算法的试验结果采用峰值信噪比（Peak signal noise to ratio,PSNR）和算法耗时2个指标进行评价，结果表明：改进模糊中值滤波算法相对于其余3类算法而言不仅去噪效果较优，而且耗时较少，对于批量处理矿井视频监控图像有一定的参考价值。

Improved Fuzzy Median Filtering Algorithm of Mine Video Monitoring Images

Due to the existing factors of dust and uneven illumination in mine,so,the images obtained by mine video monitoring system are usually containe a lot of noise and the brightness of the image is uneven,it affect the real-time effective monitoring of the main production status to some extent.In orde to process mine video monitoring images with high effects,an improved fuzzy median filtering algorithm is proposed.Firstly,an adaptive calculation method of improved fuzzy membership degree coefficient is put forward,the minimum,maximum and average grey values of the pixels of each filtering window of mine video monitoring image are taken as the thresholds to calculate the fuzzy membership degree coefficients of the grey values of different pixels,so the deficiencies of setting many thresholds of the classical calculation method of fuzzy membership degree coefficient is overcame;then,the fuzzy weighted median filtering is conducted based on the fuzzy membership degree coefficients;finally,in view of the different fuzzy degree of the filtered image,the gray histogram of fitered image is conducted equalization processing to enhance the overall visual effects of the image.The program of the algorithm proposed in this paper is written by C# language,the experimental data are the two mine video monitoring images acquired from a mine.The fuzzy median filtering algorithm,weighted median filtering algorithm and switch median filtering algorithm are used to conduct contrast analysis of the filtering effects,the above experimental results of the algorithms are evaluated by peak signal to noise ratio (PSNR) and time-consuming,the experimental results show that the performance of the improved fuzzy median filtering algorithm is superior to the fuzzy median filtering algorithm,weighted median filtering algorithm and switch median filtering algorithm,besides that,the time-consuming of the improved fuzzy median filtering algorithm is less than the orther three algorithms,it can provide the reference for processing a large number of mine video monitoring images effectively.