What It Does
Our project utilizes OpenCV to process a live camera feed and applies OCR (Optical Character Recognition) to extract text from the environment. The extracted text undergoes filtering through custom algorithms and a Large Language Model (LLM) before being transmitted to a Raspberry Pi. The Raspberry Pi then converts this data into physical Braille using motors, enabling the user to read their surroundings through a wearable Braille device. Challenges We Faced
This project was an entirely new challenge for us. From working with Python, OpenCV, and OCR, to integrating a Raspberry Pi for controlling a mechanical system, every step was a learning experience.
One of the biggest hurdles we encountered was dealing with faulty servo motors, which took several hours of troubleshooting and developing a workaround. Additionally, we initially attempted to use Python sockets to stream video from the Raspberry Pi to a more powerful computer for processing. However, this approach led to numerous connectivity issues and wasted hours of debugging. Ultimately, we pivoted to a simpler solution—connecting the camera directly to the computer to ensure both our software and hardware functioned smoothly for the demonstration. Future Improvements
With more time, we aim to refine and optimize the mechanical Braille system, ensuring smoother and more reliable operation. Additionally, enhancements to our custom text filtering system will improve the accuracy and readability of detected text, making the device even more effective for users.
Log in or sign up for Devpost to join the conversation.