Research On High Speed Railway Snow Monitoring System Based On Laser Ranging Sensor And Video Analysis

In recent years, our country has entered into the phase of rapid construction of high-speed railway and opening. A large number of high-speed railway lines across north and south, from east to west, have been opened. At the same time, our country is a large country with big difference in weather, terrain, landform and geology. Facing the threat of various natural disasters such as strong wind, heavy rain, snowstorm, earthquake, debris flow and rolling stones, high speed railway lines cross from dry and cold northeast to the hot and humid Hainan Island. So collecting different kinds of meteorological and disasters data real-timely and accurately, publishing warning information in time and adjusting train operation by establishing a perfect natural disaster monitoring system is very important and workable.China’s new high-speed railway in northeast China has deployed the snow depth monitoring system by installing laser snow depth measuring device along the railway to the snow depth above the tracks. But this system has relatively low measure accuracy and low reliability due to unsmooth surface, angle error and covers blocking the laser, based on unmovable installation and the distance measurement relying on laser sensor. Besides, it is not enough to estimate the influence of snow on railway operation only by snow depth data. Therefore, this paper is focused on the snow depth monitoring system. Firstly, we give a novel multi-point measurement method by integrating PTZ and angle sensor to the current system. Secondly, we do video capture and intelligent video analysis by adding a high definition webcam. Lastly, we implement a platform software having functions such as data collecting and graphing, synergistic alarming and equipment management.Once the snow depth measurement device being installed, the angle of the sensor has been determined which in fact will change affected by the vibration for a long time. Thus, slight change of angle will cause serious error of snow depth. Integrated with PTZ and angle sensor, the influence of angle error on snow depth will be reduced or avoided through multi-point measurement. At the same time, comparing to single-point snow measuring, multi-point measurement will reduce the error of unsmooth surface by averaging a set of data instead of a single data.The rapid development of computer vision technology and the widespread deployment of railway integrated video monitoring system provide conditions for intelligent video analysis. In snow weather, snow particle density can reflect the real-time snowfall intensity.By real-time video analysis, snow particle density is extracted in foreground to estimate current time snow intensity which provides dynamic snow information.And the real-time snow intensity will be useful for short-time weather forecast of a small area and for analysis of snow disaster trend. In addition, based on contrasting the template image and captured image, the algorithm can detect the existence of a cover in the monitoring scene and estimate the area of it which is capable of eliminating the interference of the cover on snow depth measurement effectively. And cover detection having a certain practical value can also be helpful for railway foreign body invasion monitoring system.Through the configuration software, the monitoring platform can help train dispatching staff and track maintenance personnel have an easier and more direct understanding of the snow stage along the railway by putting real-time snow depth data and video image together in one interface. What is more, the platform will alarm automatically by capturing picture and transmitting it to the monitoring center to notify train dispatching staff and track maintenance personnel to take measures to ensure the safety of the train once the snow depth surpasses the threshold.