融合中值滤波与小波软阈值去噪模型的新元矿视频监控图像滤波方法

发布时间：2018-04-11 05:11

本文选题：井下视频监控系统 + 中值滤波　；参考：《金属矿山》2017年12期

【摘要】：井下视频监控系统的应用对于实时获取井下生产进度、机电设备运行状况等信息,及时有效开展井下应急救援工作发挥了重要作用,但井下光照不均匀、空气中大量粉尘导致监控系统获取的图像较模糊,影响了对井下各类信息进行有效采集和分析。以新元矿井下视频监控系统为例,将中值滤波算法与小波软阈值去噪模型相结合,提出了一种井下视频图像滤波方法。该方法预先对原始视频图像进行3层小波变换,对获取的低频小波系数和高频小波系数分别进行逆变换,得到3幅低频图像和3幅高频图像;针对低频图像,采用融合噪声判别准则的改进中值滤波算法进行去噪;对于高频图像采用改进型小波软阈值去噪模型进行处理。在此基础上,将滤波后的低频和高频图像进行融合,实现对视频图像的高效滤波。采用该矿井下两幅视频图像对所提方法进行试验,并与中值滤波算法、小波硬阈值去噪模型、小波软阈值去噪模型进行对比分析,结果表明,所提算法处理后的图像清晰度明显优于其余3类算法,且该方法耗时较短,适合于高效处理井下视频监控图像。
[Abstract]:The application of underground video surveillance system plays an important role in obtaining the information of downhole production progress, electromechanical equipment running condition, and carrying out underground emergency rescue work in time and effectively, but the underground lighting is not uniform.A large amount of dust in the air leads to blurred images obtained by the monitoring system, which affects the effective collection and analysis of all kinds of underground information.Taking the video monitoring system of Xinyuan mine as an example, a downhole video image filtering method is proposed by combining the median filtering algorithm with the wavelet soft threshold denoising model.In this method, the original video image is pre-processed with three-layer wavelet transform, and the obtained low-frequency wavelet coefficients and high-frequency wavelet coefficients are inversely transformed respectively to obtain three low-frequency images and three high-frequency images.The improved median filtering algorithm based on fusion noise criterion is used to remove noise, and the improved wavelet soft threshold denoising model is used to process high frequency images.On this basis, the filtered low-frequency and high-frequency images are fused to realize the efficient filtering of video images.The proposed method is tested by two video images under the mine and compared with median filtering algorithm, wavelet hard threshold denoising model and wavelet soft threshold denoising model. The results show that,The image sharpness of the proposed algorithm is obviously better than that of the other three kinds of algorithms, and the time consuming of the proposed algorithm is relatively short, so it is suitable for efficient processing of underground video surveillance images.
【作者单位】：江苏海事职业技术学院信息工程学院;
【分类号】：TD76;TP391.41

【相似文献】