Bangladesh has been experiencing frequent deadly accidents at rail crossings all over the country. Recently, (on 23 July 2022) a train ran over a newly-married couple at a rail crossing in Sirajganj. After a couple of weeks, another tragic accident occurred. A microbus carrying 11 students, who were returning from an excursion, got through an unmanned level crossing and was hit by a speedy train. According to Bangladesh Railway (BR) statistics, almost 90% of rail accidents happen at rail crossings. In the last one decade, a total of 936 accidents happened at the country's rail crossings, which took away a staggering 516 lives. According to data from the Railways Ministry, among the 2822 rail crossings of BR, 2258 are without gatemen. Among thousands of rail crossings spread all over the country, only 564 have gatemen, majority of whom is hired temporarily under different projects. Among these gatemen, only 331 are holding permanent positions. At least six persons are needed to operate a single rail crossing properly, which the BR lacks. To fill up the gap, BR has to recruit around 13,548 people, which is very difficult and time consuming.

As an alternative to the gateman, we propose “Safe Rail Gate”, a machine learning based smart rail gate management system, to automatically control the gates on a rail crossing before a train approaches. The proposed system will use a combination of laser and camera sensors to identify the arriving of a train. Mainly, the system will continuously capture images which will be fed to a Convolutional Neural Network (CNN) to decide whether there is a train arriving towards the crossing. If a train is detected, the system will automatically sound an alarm and close the gates without any human intervention. Also, the system will open the gates when the train leaves the crossing.