Automating the Railway Track Inspection Process through Drone Technology using state-of-the-art Deep Learning Techniques

Abstract

Rail transport is said to be the second-largest transportation means in India. This transportation mode is said to be the fastest way within a large network of commercial nodes in a country to transport various goods, products and passengers every day. Hence, the quality and maintenance of railway tracks in good condition are highly important for safe and secure transportation. Cracks existing in the rail tracks may lead to train accidents. As the railway tracks pass through rural areas and non-residential places, it is practically difficult for human beings to check for any cracks in the tracks daily. Even if it is observed, the distance between two stations will be large and hence it takes a long time to repair the cracks in the track. Timely detection is therefore, critical for the warranty of uninterrupted continuous transportation.   Hence, automating the railway track inspection process through Drone technology is proposed in this paper as an efficient methodology to manage such problems. A drone is designed and trained to inspect the railway tracks automatically and frequently in a periodical manner. The drone has a high-resolution camera fixed which captures the images of the track and checks and detects cracks in them. The drone is equipped with a Global Positioning System and a Wi-Fi module and interface with the base station drone unit located in the nearby railway station. It intimates or sends an alert message to the station master when a crack is detected. The exact location with latitude and longitude parameters are captured and sent to the station master. Thus, it saves manpower and minimize the time in the inspecting process. Above all, it assures a safe journey through the rail tracks.