Revolutionizing Railroad Sleeper Damage Detection with AI/ML

The client is an international company specializing in automating railroad processes. They combine expert knowledge to build, maintain, control, and monitor railroad networks. 

The client needed AI/ML specialists to develop software to identify specific foundation defects.  

1. Software analysis and initial concept. 

We identified weaknesses in the client’s platform and outlined possible solutions for implementation. 

2. Conduct a stability study and present it to the client. 

The client outlined the project objective by identifying specific damage to railroad sleepers. Based on this, we performed a preliminary stability study. 

3. Develop the final software package to solve customer problems, including a data processing library. 

After receiving approval for the initial software concept, we developed the final system and a data processing library to address the client’s problems.  

At the start of our collaboration, we created a software package based on images provided by the client, integrating these with the measurement system. The package processed each image, highlighting areas with potential defects, such as primer cracks or spalling. After processing, we prepared reports with the results and the source code. The client then implemented these packages into their software system. 

The system was well-received by the end user, the National Railway. Additionally, the client presented the results of our work at the Machine Vision Week Conference, where the audience learned about the accuracy and effectiveness of our software library. 

The development and cooperation process with the client proceeded smoothly and was continuously monitored. Our team of experts readily shared their knowledge, which the client appreciated.