Helping People with Visual Impairment Navigate Buses Through AI and Computer Vision

Background

People with visual impairment face numerous daily challenges, particularly in orientation and mobility—the ability to travel safely and accurately from one location to another. Public transportation, like buses, is a critical means for these individuals to navigate their surroundings and maintain independence. However, using buses often presents significant obstacles, such as identifying the correct bus, locating its entrance, and finding an empty seat.

To address this issue, Omdena collaborated with RenewSenses, an Israeli company that develops assistive technologies for people who are blind. Together, they worked on a solution using AI for visual impairment, aiming to make bus navigation more accessible and empowering for visually impaired individuals.

Objective

The primary goal of this project was to design and implement a Computer Vision-based system capable of helping visually impaired individuals:

• Detect buses and identify bus lines.
• Locate bus entrances.
• Find available empty seats inside buses.
  This system aims to enhance daily mobility and foster greater independence for individuals with visual impairment.

Approach

To tackle the challenge, the team adopted a structured and innovative approach:

1. Ethical Data Collection: Using advanced techniques to gather relevant images and videos while respecting privacy.
2. Data Annotation and Preprocessing: Leveraging tools like CVAT to annotate data and refine it for model training.
3. Computer Vision Model Development: Implementing state-of-the-art algorithms for object detection, focusing on buses, bus lines, and seats.
4. Mobile and Real-World Testing: Testing the system on mobile devices to ensure practical usability.
5. Deployment and Integration: Using GPU-based Docker deployment to ensure efficient, real-time performance in real-world conditions.

Over 35 experts and volunteers contributed to this effort, utilizing a collaborative and iterative process to refine the solution.

Results and Impact

Within just two months, the team successfully developed a fully functional Computer Vision pipeline capable of:

• Detecting buses and identifying their line numbers.
• Recognizing empty seats within buses.
• Operating in real-world environments with high reliability.

The solution is being tested directly with visually impaired users through pilot programs conducted by RenewSenses. This direct feedback loop ensures the technology meets real-life needs and has a significant positive impact on mobility and independence.

The project has the potential to transform how visually impaired individuals interact with public transportation, enabling them to travel more confidently and safely.

Future Implications

The success of this project demonstrates the potential of AI for visual impairment in solving real-world challenges. In the future, this technology could:

• Expand to other forms of public transportation, like trains and subways.
• Be integrated into broader mobility solutions for visually impaired individuals.
• Influence policies and encourage further investment in assistive technologies.
• Serve as a foundation for advancing AI solutions tailored to accessibility and inclusivity.

By addressing the critical mobility challenges faced by individuals with visual impairment, this project represents a significant step toward creating a more inclusive and equitable society.