Projects

Automatic Object Detection with Rear Cameras


Source Organization:
University:
Principal Investigator:
PI Contact Information:HH B205, Dept. of ECE, CMU, Pittsburgh, PA 15213
phone: (412) 2683026
email: kumar@ece.cmu.edu
Project Manager:
Funding Source(s) and Amounts Provided (by each agency or organization):
Total Dollars:$50,000 (UTC) + $25,000 (Discretionary)
Agency ID/Contract/Grant Number:
Start and End Dates:
Project Status:Complete
Subject Categories:In-vehicle safety technologies, rear cameras, object detection
Abstract:The project goal is to develop automatic object detection in images/video provided by the rear cameras. Such automatic object detection can provide an alert to the driver. According to Kidsandcars.org, the number of backing-up fatalities increased five-fold from 88 (1996-2000) to 448 (2006-10). Our object detection method will consist of 3 steps. (1) Normalization: Images acquired change in appearance due to weather and illumination conditions, lens distortion and other factors. (2) Feature Computation: Systematically evaluate and select features. (3) Classification: Determine the presence of objects.
Describe Implementation of Research Outcomes (or why not implemented):
Impacts/Benefits of Implementation (actual, not anticipated):
Project URL:http://users.ece.cmu.edu/~kumar/