Asset Detection and Mapping

Knowing the location of assets is extremely important for maintenance and renewals planning. Few rail companies have a detailed list of all their assets with location and visual data on them. Given high costs associated with manual surveys, the cost of update is also very high. A typical approach to surveying the assets on rail tracks is through rail vehicle borne imaging. A number of cameras installed on the moving vehicle capture track images. The process of tagging the images for assets is the same as for aerial imaging and therefore suffers from the usual problem of large resources required to accomplish this task. Surveys take months or even years to complete, and by the time the information is fully available, it is often out of date.

Rail Vision has developed the latest technology for asset surveying based on automated image understanding. Train-borne videos are split into individual images. Each image is processed to recognise assets based on a range of colour, texture, edge, shape and statistical features. Automated approaches are significantly faster and more reliable compared to manual approaches to asset tagging in videos.

The output of the survey is available in the form of a database (SQL/Oracle). Our information management tools allow the user to query and retrieve information from this database as needed. Asset surveys are also presented through a Geographic Information System (GIS) interface (rail route-maps (if available) or superimposed on Google-maps). Once the asset survey is completed, the information can be used on its own (through GIS maps viewed on desktop computers or PDA devices), or exported to other applications that need such information.