College Senior Project: Visual Assistance Cane
Spring 2021
On a team with 4 other engineering students, we designed, programmed and built a device intended to aid those with visual impairment.
High-Powered Computing for Real-Time Visual Assistance
The VAC was designed to be a sensory aid for the visually impaired. An Intel RealSense D435 interfaced with an Intel UP Squared Board model were used to generate and evaluate a depth map and visual stream of the users immediate surroundings. The physical cane, featuring a compact, yet similar design to that of a collapsible white mobility cane. The end product was capable of Bluetooth audio cues and/ or vibrations cues to alert the user of objects on the verge of entering their space. At the time of completion, some early stage machine learning had been incorporated for future object recognition capabilities.
Design
Throughout the course of this project, the team was responsible for sourcing materials, hardware, and programming. The cane housed the D435 Camera/Up Squared Board interface near the handle of the cane for ergonomic purposes and to broaden the camera's field of view. Carbon fiber tube segments made up the length of the cane, allowing for durability and easy stowability.
Programming
While all members of the team took part in all aspects of the device's design, my role oversaw all programming and machine learning components. The program was built with Python, and heavily utilized the OpenCV library for image acquisition and evaluation.
