面向矿业摄像仪的图像增强与目标跟踪  

Image enhancement and target tracking for mining camera

作　　者：郭刚 李明利 杨波涛 白鑫 王喜升 王子陵 Guo Gang;Li Mingli;Yang Botao;Bai Xin;Wang Xisheng;Wang Ziling(China Coal Shaanxi Yulin Dahaizhe Coal Industry Co.,Yulin 719099,China;China Coal Information Technology Co.,Ltd.Beijing Dongcheng District 100013)

机构地区：[1]中煤陕西榆林大海则煤业有限公司,陕西榆林719099 [2]中煤信息技术(北京)有限公司,北京市东城区100013

出　　处：《现代科学仪器》2023年第4期148-153,共6页Modern Scientific Instruments

基　　金：中关村绿色矿山产业联盟课题:基于5G在煤矿深部复杂环境下的智慧安全开采研究和应用(中绿盟鉴字[2022]第202号)。

摘　　要：针对煤矿井下照度低导致的矿业摄像仪采集图像模糊,不利于运动目标跟踪的问题,提出一种面向矿业摄像仪的图像增强与目标跟踪方法。方法首先采用小波变换和逆变换增强矿业摄像仪图像,然后通过结合高斯背景模型和三帧差分算法提取精确的运动目标轮廓信息,实现了高检测效率和高检测精度的目标跟踪。仿真结果表明,所提方法可明显提高矿业摄像仪采集图像的亮度,且使图像边缘和纹理更加清晰,并可有效检测到矿业摄像仪采集图像中的运动目标,实现对运动目标的跟踪。相较于传统高斯背景模型和改进建模方法的高斯背景模型,所提方法检测精度更好,检测效率为105ms/帧,误检率和漏检率分别为5%和4%,可满足矿井下视频实时监测及精度的需求,具有一定的实际应用价值。Aiming at the problem that the image collected by mining camera is blurry due to low illumination in coal mine,which is not conducive to moving target tracking,an image enhancement and target tracking method for mining camera is proposed.Methods Firstly,wavelet transform and inverse transform are used to enhance the image of mining camera,and then the accurate contour information of moving target is extracted by combining Gaussian background model and three frame difference algorithm to achieve target tracking with high detection efficiency and high detection accuracy.The simulation results show that the proposed method can significantly improve the brightness of the image collected by the mining camera,and the image edge and texture are more clear,and can effectively detect the video moving targets collected by the mining camera,achieve the tracking of moving targets,with high detection efficiency.Compared with the traditional Gaussian background model and the Gaussian background model of the improved modeling method,the proposed method has better detection accuracy,detection efficiency of 105ms/frame,false detection rate and missing detection rate of 5%and 4%respectively,which can meet the requirements of real-time monitoring and accuracy of underground video,and has certain practical application value.

关 键 词：矿业摄像仪 图像增强 目标跟踪 小波变换 高斯背景模型 三帧差分算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]