








The Safe Rail academic research and development project, sponsored by H.E.C., aims to enhance rail track safety through innovative technological solutions.
Railway accidents persist as a pressing concern within Pakistan’s transportation landscape, resulting in the tragic loss of hundreds of lives. This critical issue demands immediate attention and comprehensive measures to ensure the safety and security of passengers, thereby preventing further casualties and enhancing the reliability of the railway system.
As a part of this, DAANISH Technologies took the initiative to help the researchers. We provide solutions through our partnership with Luxonis, a USA-based AI Robotics Vision Company. Using advanced imaging technology and algorithmic analysis, Daanish Technologies collaborated with NUST to address track inspection challenges.
Luxonis is a spatial AI and CV platform that focuses on embedded machine learning and computer vision technology. Luxonis, our technology partner, provided invaluable support in algorithm development and optimization for real-time defect detection.
National University of Sciences and Technology (NUST) Robotics Research Center is a hub for cutting-edge research and innovation in robotics. With the emphasis on science and technology innovation, the Robotics Research Center focuses on advancing robotics technologies, fostering interdisciplinary collaborations, and addressing real-world challenges through robotics solutions. The center conducts Robotics research in various domains such as autonomous systems, human-robot interaction, machine learning, computer vision, and more.
When NUST presented the challenge of enhancing rail track inspection, we recognized the need for a comprehensive solution to detect track defects effectively. The project aims to conduct research into the classification of the different modes of failure in Rail track components and to perform a statistical analysis to converge the data to determine the most important failures that occur in turnouts.
Understanding the limitations of traditional inspection methods, such as human error due to visual inspection. Mistakes made by railway personnel, whether in operations, maintenance, or management, can cause failures. We identified the necessity for a robust track inspection system to capture high-resolution images and detect various rail defects in real-time. AI cameras can automatically adjust camera settings to improve the image quality such as facial recognition, object detection, and scene recognition, even in challenging conditions that are impossible with traditional cameras.
We delved into the requirements and challenges associated with track inspection, considering the complexities of identifying rail gauge issues, fish plate conditions, and sleeper defects. The defects could be the result of fatigue crack failures, wear failure, shear failure, material deformation failure, and rolling contact fatigue cracks.
The cause of the majority of cracked fishplates lies in inadequate support to the sleepers. Cracks may originate at either the upper or lower edge of the fishplate. A transverse crack emanates from behind the fishplate and extends to a position.
The sharp corner which has formed on the switch rail occurs most commonly. It causes the wheel to derail subsequently. In addition, as a result of a wheel strike, a deformed nose might form, which leads to a change in the check rail gauge dimension, especially when there is a loose bolt. Sharp corners should be removed by grinding.
Sleepers or ties receive the load from the rail and distribute it over the supporting ballast. They also hold the fastening system to maintain the proper track gauge and restrain the rail movement. Failure mechanism in sleeper has been followed in this section. Bending cracks are often detected at the bottom of mid mid-span of the sleepers by allowing a sleeper to settle on the ballast packed under the center of the sleeper. It also sometimes comes from the soffit underneath the rail seat of the sleepers. It eventually decreases the flexural stiffness of the sleeper. Likewise, flexural failure appears as top surface cracks located near the rail seat.
Leveraging the Luxonis OAK-D-POE Cameras and collaborating with an industrial partner, we designed specialized track inspection equipment that is integrated into the train. This equipment housed the selected sensors and integrated developed algorithms for real-time defect detection and analysis. The OAK-D-POE Cameras utilized advanced imaging capabilities and Power over Ethernet (PoE) functionality, enabling high-resolution image capture and seamless integration into the inspection vehicle. The original OAK-D PoE offers an IP67 rated enclosure for harsher environments and RJ45 Ethernet connector. See Series 2 PoE models for more CCM options including wide field of view, industrial M12X connector, night vision or improved depth perception.
Creating customized camera mounts and ensuring seamless integration with the inspection vehicle posed initial challenges. Also, crafting algorithms capable of accurately identifying various track defects required extensive testing and optimization.
DAANISH Technologies played a pivotal role in revolutionizing rail track safety through their collaboration with Luxonis and NUST. Their innovative approach, leveraging Luxonis OAK-D-POE Cameras, resulted in a specialized track inspection system integrated into trains. Despite initial challenges, DAANISH Technologies showcased dedication and expertise, providing a comprehensive solution for real-time defect detection. The success of the project, highlighted by the development of the OAK-D Pro W PoE, underscores their commitment to advancing railway safety. We highly commend DAANISH Technologies for their invaluable contributions to the Safe Rail project.
The project was led by Mr. Ahmed Said, DIRECTOR DAANISH Technologies, Lahore Pakistan. He collaborated with the Robotics Research Center at NUST, Islamabad Pakistan to make the project successful on route.