Safeguarding Metro Tunnels

Elevating Tunnel Safety: Real-time Anomaly Detection with Thermal Imaging

Comment

Project Story

About the Project

The inspiration behind our project stemmed from the critical need to enhance safety measures in metro tunnel construction, particularly in densely populated urban areas like Chennai. As the demand for faster tunneling increased, so did the challenges associated with ensuring worker safety and infrastructure integrity. Our goal was to develop a solution that not only addressed these challenges but also leveraged cutting-edge technology to make metro tunneling safer and more efficient.

What We Learned

Throughout the project journey, we learned invaluable lessons about the intricacies of metro tunnel construction and the complexities involved in implementing real-time safety measures. We gained a deeper understanding of thermal imaging technology, deep learning algorithms, and integration techniques. Additionally, we honed our skills in project management, collaboration, and problem-solving.

How We Built Our Project

Our project was built upon a foundation of thorough research, innovative technology, and collaborative teamwork. We meticulously designed and recreated tunnel structures using GI metal sheets to provide a realistic testing environment. The Grid Eye Evaluation Gen 2.0 thermal camera was strategically installed for monitoring purposes, ensuring comprehensive coverage of the tunnel interior.

To ensure seamless integration, we developed a protocol to deactivate the thermal camera during train passages, preventing unnecessary image capture and optimizing system efficiency. A robust detection protocol was implemented to activate image capture upon detecting cracks, anomalies, or intruders, enabling swift response to potential safety threats.

Captured images were promptly transmitted to the Jenkins pipeline for further processing. Within a Docker environment, we employed a machine learning model, trained with 3000 labeled thermogram images using YoloVNAS, to analyze captured images for verification. The model's accuracy was continuously monitored and refined to ensure reliable anomaly detection.

To facilitate real-time alerts, we developed a user-friendly Flutter application that generated and transmitted alerts promptly when anomalies occurred. Optimal camera positioning was achieved through reinforcement learning algorithms, simulating the environment to identify blind spots and optimize coverage.

Solution and Benefits

Our solution revolutionizes metro tunnel safety by providing real-time anomaly detection and alerts, thereby significantly reducing the risk of accidents and ensuring the well-being of tunnel workers. By leveraging thermal imaging technology and deep learning algorithms, we offer a proactive approach to safety management, enabling early detection of potential hazards such as cracks and unauthorized access.

Key benefits of our solution include:

  • Enhanced Safety: By continuously monitoring tunnel conditions and detecting anomalies in real-time, our solution minimizes the risk of accidents and injuries to tunnel workers.
  • Improved Efficiency: The integration of advanced technology streamlines safety protocols, allowing for faster response times and more efficient management of safety incidents.
  • Cost-effectiveness: Our solution utilizes existing infrastructure and affordable components, making it a cost-effective option for metro tunnel operators.
  • Compliance: By implementing rigorous testing and adherence to industry standards, our solution ensures compliance with safety regulations and protocols.
  • Peace of Mind: With real-time alerts and comprehensive monitoring capabilities, our solution provides peace of mind to tunnel operators and stakeholders, knowing that safety is prioritized at all times.

POC Cost

The proof of concept (POC) for our project incurred the following costs:

  • Grid Eye Evaluation Kit: ₹7,000
  • Ultra Sonic Sensor: ₹350
  • Docker: ₹0
  • Flutter Application: ₹0

Total POC Cost: ₹7,350

Challenges Faced

  • Technical Challenges: Developing and fine-tuning the deep learning model to accurately detect anomalies posed a significant technical challenge. Additionally, ensuring seamless integration and compatibility between various components required meticulous planning and execution.
  • Cost Considerations: Balancing the project's technological requirements with budget constraints was a challenge, particularly when selecting equipment and tools for the proof of concept.
  • Operational Constraints: Adapting the solution to real-world operational conditions, such as train passages and varying lighting conditions, required careful consideration and testing.

Despite these challenges, our team remained dedicated and resilient, ultimately delivering a comprehensive solution that prioritized safety and efficiency in metro tunnel construction.

Built With

Share this project:

Updates