基于矿井监控的视频质量矫正算法

A Video Quality Correction Algorithm Based on Mine Monitoring

煤矿井下光照条件复杂、空间可见度低,导致采集到的监控视频质量差,不利于基于视觉的智能化功能展开。通过对数字图像处理技术的研究,提出了一种针对矿井监控视频的图像矫正算法。首先,对输入的视频帧使用直方图均衡化处理,提高图像的全局对比度。其次,将输入的图像转换至HSV频道,在亮度V频道下,借助伽马映射矫正图像的亮度值,并结合矫正后的亮度值转换至RGB频道。最终,将前两步处理后的结果在像素维度进行灰度值叠加。实验证明,算法平均灰度值相比对比算法最高,信息熵相比对比算法最大,在全局亮度和细节信息丰富度方面,优于其他图像矫正算法。

The complex lighting conditions and low spatial visibility in coal mines result in poor quality of collected monitoring videos,which is not conducive to the deployment of visual based intelligent functions.Through research on digital image processing technology,an image correction algorithm for mine monitoring videos was proposed.First,histogram equalization was applied to the input video frame to improve the global contrast of the image.Secondly,the input image to the HSV channel was converted,and in the brightness V channel,gamma mapping was used to correct the brightness value of the image,and the corrected brightness value was combined to convert it to the RGB channel.Finally,the results of the first two steps would be overlaid with gray scale values in the pixel dimension.Experiments showed that the average gray scale value of the algorithm was the highest compared with the contrast algorithm,and the information entropy was the largest compared with the contrast algorithm.It was superior to other image correction algorithms in terms of global brightness and detail information richness.