Research On Coal Flow Measurement Method Based On TOF Feature Images

It is an inevitable trend for the development of the coal mining industry to build an informatized,digital,intelligent and unmanned modern smart mines.In all aspects of the smart mines,intelligent production,intelligent transportation and intelligent dispatching management all require the support of coal flow data.Whether from the aspect of coal production or from the aspect of safety,the measurement of coal flow in the conveyor belts is very important.Aiming at the problems of current non-contact coal flow measurement method cannot obtain coal depth information quickly and intensively,a coal flow measurement method in conveyor belts based on TOF(time-of-flight)feature images is proposed.First,in the TOF feature images preprocessing part,for the convenience of observation and subsequent processing,the original depth images and intensity images are histogram equalized,and median filtering is adopted to remove the impulsive interference signal in the depth images while protecting the edge information of the image.To facilitate the establishment of the subsequent coal volume measurement model,coordinate transformation is used to convert the depth images from the image coordinate system to the world coordinate system.Aiming at the interference of background noise in the process of coal images acquisition and processing,the coal region recognition algorithm is proposed.The algorithm utilizes the advantage that TOF depth images can directly obtain the depth information of the object surface to obtain the coal ROI(region of interest).Then,in the coal flow measurement part,aiming at the problem that the depth image cannot accurately represent the depth value at the edge of the object where the depth value is discontinuous,a depth image inpainting algorithm is proposed.The algorithm utilizes the respective advantages of the depth image and the intensity image and fuses the depth information of the depth image and the reliable edge position information of the intensity image to obtain a high-precision depth map.An algorithm for coal flow velocity measurement based on TOF feature images is proposed,and then,based on the high-precision depth map and coal flow velocity information obtained above,a coal flow calculation model is designed to calculate the current coal flow.Finally,to explore the feasibility,accuracy,precision and real-time performance of the proposed measurement method,an experimental platform for coal flow measurement is established in the laboratory.The characteristic images captured by the TOF camera are processed under different belt speeds,coal loads and ambient light conditions,and the coal flow is calculated.The results show that the maximum error of the coal flow measurement method proposed does not exceed 3.54%,the maximum standard deviation does not exceed0.487,and it is suitable for use in the case of complex ambient light in the mine.Moreover,the average processing time for each frame of image does not exceed 83 ms for each frame of image,which meets the real-time requirements.The coal flow measurement method based on TOF feature images can provide coal flow data support for intelligent production,transportation,and dispatching management,laying a foundation for the construction of smart mines.