Blav

Blav is a tool for the blind to navigate around. It uses a Qualcomm Dragonboard 410c for sensors and other data.

How it was made

• The dragonboard collects data like webcam, distance, etc. and determines which direction to walk in using Machine Learning and Image Processing.
• The hasura backend is used to communicate between the dragonboard and the computer. It has a flask-restful api deployed.

How does it detect

• Using image processing, we detect certain points on the 'sidewalks'.
• Then, we run regression over it to get a polynomial.
• The angle of intersection between these two polynomials determines which direction to turn in.

Difficulties we ran into

• The dragonboard was a huge pain to get working as there isn't much documentation on it. The Hasura APIs had little to none documentation, which made it very hard to set up.
• The biggest challenge we ran into was getting the dragonboard to output current through its GPIO pins. It sounds easy, but with almost no documentation about a certain include error, we took hours to resolve that. ## What's next We would love to get this out in the real world and make our algorithm much more reliable.
  

Built With

• arduino
• c++
• dragon-board
• flask
• hasura
• opencv
• python