16. Core image analysis technology for utilizing train-front images

In railways, inspections are conducted by staff on-board the train to check the conditions of railway track side equipment and the surrounding environment. In recent years, an increasing number of railway operators have been using video cameras to film from the front of the train so that the conditions during these inspections can be preserved as images. By utilizing these images to understand the location and condition of equipment, we can expect improvements in the efficiency of construction design and a reduction in the labor required for inspection and examination tasks.

To effectively use train-front images across various technical fields such as track maintenance and electrical systems, and to apply them to tasks like construction design and maintenance, the following methods are useful: (1) A method for accurately calculating the position on the railway track corresponding to the distance from train-front images, (2) A method for converting train-front images into overhead view images, (3) A method for generating long-length railway track surface images from multiple overhead view images (Figure 1), (4) A method for recognizing equipment within the images (Figure 2), and (5) A method for assessing the degradation of recognized equipment. Therefore, we developed and modularized methods (1) through (5) as core technologies that can be commonly used when applying train-front images to various purposes.

When each railway operator develops applications to improve operations and safety using train-front images, the development process can be streamlined by combining these groups of modules to form the basic part (Figure 3).