Stereo Vision: Depth Perception
Loading...
Authors
Guzelocak, Murat Can
Date of Issue
2025-04-25
Type
Presentation
Language
en_US
Subject Keywords
Other Titles
Stereo Vision and Depth Estimation
Abstract
Snow is arguably one of the biggest challenges of attending college in Montana during the winter. Navigating campus can be particularly difficult due to icy and snow-covered pathways. One solution to this problem is snow removal, which is a labor-intensive task requiring significant time and effort, especially in Montana’s harsh winters. An automated snow shoveling robot has been proposed to assist laborers during this season. Automation requires various algorithms, including object detection, mapping, and distance measurement.
This project presents an algorithm for an automated snow shoveling machine that utilizes stereo vision with two parallel cameras to measure depth accurately. This depth estimation helps the machine determine how close it needs to get to snow for effective removal while also detecting and avoiding obstacles—an essential feature for full automation. The algorithm employs disparity mapping, which calculates depth by leveraging the known distance between the two cameras and comparing pixel differences between the two captured images.