The railway is one of the crucial transporters for many countries. Most of the developed countries have massive rail route networks. The running of a railway on track is controlled by a driver called a loco pilot. The signalling system has been developed to give necessary signals to the loco pilot, based on which he controls the train speed. The loco pilot controls the train’s braking system based on the visual colour signals given to him, which are installed near the rail tracks. These railway signals are controlled by the station master of every station via specialised controlling panels provided to them. Reading the aspect of this railway signal becomes a crucial task for the loco pilot. The safety of trains depends totally on his vision capability and his concentration. In order to enhance safety and provide assistance to loco pilots, it is necessary to introduce a real-time railway signal aspect detection system based on computer vision. In this paper, different versions of YOLO algorithms explored real-time signal aspect detection to assist the loco pilots and increase the operational safety of the trains.