基于视觉的煤矿井下带式输送机异常状态监测方法研究

发布时间：2018-03-30 07:04

本文选题：煤矿安全生产　切入点：计算机视觉　出处：《太原科技大学》2017年硕士论文

【摘要】：视频监控是安全防范的重要手段,近年来视频监控技术在矿井生产中得到广泛应用,为煤矿安全生产提供技术上的支持和保证。目前我国矿井下的视频监控系统主要以人工监视为主,经常出现人为的错漏现象。随着计算机视觉技术的发展,煤矿智能视频监控将取代人工监控的方式实现对异常状态的实时自动监测。带式输送机作为矿井生产中应用最多的运输设备,在长时间高强度运转下极易出现各种故障,也是煤矿井下视频监控的重点。本文研究了基于视觉技术的矿井下带式输送机异常状态监测方法,以实现带式输送机的智能监控,具体研究内容如下:首先,为解决煤矿井下视频监控系统采集到的图像对比度低、光照不均且伴有大量噪声等视觉效果差的问题,给出了一种基于加权引导滤波同步去噪的单尺度Retinex算法实现煤矿井下图像增强。利用加权引导滤波代替单尺度Retinex算法的高斯滤波对图像的低频分量进行照度估计,再经对数域转换到实数域得到反射图像,最后利用带有保边去噪功能的加权引导滤波对图像的高频分量进行去噪处理得到增强后的图像。然后针对带式输送机的输送带跑偏故障,给出了一种基于计算机视觉的输送带跑偏监测方法。首先将视频监控中采集到的视频图像设置感兴趣区域(Region Of Interest,ROI)以减少计算量,同时对ROI进行图像预处理。然后采用改进的Canny边缘检测算法得到ROI边缘二值图像,利用累计概率霍夫变换(PPHT)提取输送带边缘直线特征,最后根据所得直线特征来判断输送带是否跑偏。最后针对带式输送机在运送物料过程中的打滑故障,给出了一种基于OpenCV的带式输送机打滑检测方法。首先利用背景差分法和连通区域标记法检测出输送带上的多个运动目标,其次使用最小外接矩形获得目标的宽高比,结合基于质心特征的轨迹跟踪方法获得多个运动目标的位移及对应的时间间隔,最后利用速度公式获得输送带的运行速度,从而判断是否存在打滑故障。在经过了大量实验并对实验数据进行全面分析后,实验结果表明本文提出的基于视觉技术的矿井下带式输送机异常状态监测方法可以实现对跑偏和打滑故障的自动监测,对提高我国煤矿安全生产监控信息化水平具有重要意义。
[Abstract]:Video surveillance is an important means of safety prevention. In recent years, video surveillance technology has been widely used in mine production, providing technical support and guarantee for coal mine safety production.At present, the video surveillance system under the mine in our country is mainly manual surveillance, which often appears the phenomenon of human error and leakage.With the development of computer vision technology, intelligent video surveillance in coal mine will replace manual monitoring to realize real-time and automatic monitoring of abnormal state.Belt conveyor, as the most used transportation equipment in mine production, is prone to various faults under long time and high intensity operation, which is also the focus of video monitoring in coal mine.A single scale Retinex algorithm based on weighted guided filter for simultaneous de-noising is presented to enhance the underground image of coal mine.Using weighted guided filter instead of Gao Si filter of single-scale Retinex algorithm to estimate the illumination of the low-frequency component of the image, the reflected image is obtained by converting the logarithmic domain to the real domain.Finally, the enhanced image is obtained by using the weighted guided filter with edge-preserving denoising function to Denoise the high-frequency components of the image.Then, a method of belt deviation monitoring based on computer vision is presented in view of belt deviation fault of belt conveyor.Firstly, the region of interest (region of interest) is set up in the video surveillance to reduce the computational cost, and the ROI is preprocessed at the same time.Then the improved Canny edge detection algorithm is used to get the binary image of the ROI edge, and the cumulative probability Hough transform is used to extract the edge line feature of the conveyor belt. Finally, according to the obtained straight line feature, the conveyor belt is judged whether it is running out of direction or not.Finally, aiming at the slip fault of belt conveyor in the process of transporting materials, a method of skid detection based on OpenCV is presented.First, the background difference method and the connected area marking method are used to detect multiple moving targets on the conveyor belt, and then the minimum outer rectangle is used to obtain the ratio of the width to the height of the target.Combined with the trajectory tracking method based on the centroid feature, the displacement and the corresponding time interval of multiple moving objects are obtained. Finally, the speed of the conveyor belt is obtained by using the velocity formula, so as to judge whether there is a slip fault or not.After a large number of experiments and a comprehensive analysis of the experimental data, the experimental results show that the method proposed in this paper can automatically monitor the deviation and slip faults of the mine belt conveyor based on visual technology.It is of great significance to improve the information level of coal mine safety production monitoring in China.
【学位授予单位】：太原科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TD528.1;TP391.41

【参考文献】