Smart Vision Assist

• Tech Stack:Machine Learning,Computer Vision with OpenCV,Floor Plan Segmentation,Robotics Navigation and Sensing,Real-Time Data Processing,Algorithm Development (specifically A* Pathfinding)
• Github URL: Project Link

- Executed a comprehensive project comprising four key stages: floor plan segmentation, translation of predictions to OpenCV pixel data, A* algorithm-based path planning, and implementation of obstacle avoidance and re-routing techniques.

- Developed and trained a floor plan segmentation model using custom-generated floor plans and labels, yielding exceptional IoU scores to accurately identify and segment floor plan components.

- Implemented a robust robot navigation system utilizing a scanner for obstacle detection. Upon detecting an obstacle, dynamically generated paths are computed to circumvent the obstruction. This iterative process persists until the robot successfully reaches its intended destination.