Vision Processing Methods for Advanced Driver Assistance Systems

Paper #:
  • 2017-01-1729

Published:
  • 2017-03-28
Author(s):
Abstract:
In the increasing development of autonomous vehicles, advanced driver assistance systems play a vital role in the safety of the driver, surrounding vehicles, and pedestrians. The scope of this research is to explore the methods of recognition used to detect obstacles that a vehicle encounters. This includes, but is not limited to, road lines, stop signs, pedestrians, other vehicles, speed bumps, etc. Many challenges are presented as the importance of the visual identification becomes more ubiquitous. For example, when conditions are less than ideal, such as heavy rain, snow, or fog, the approach to ground truth recognition becomes much more difficult. This can be achieved by creating a dynamic system that evaluates the change in luminance and/or ground truth and determines the vanishing point of the current ground truth recognized. These methods cannot be achieved without the fundamental techniques of visual processing. These techniques include using thresholds, convolution, edge detection, and multiple morphological operations. Using a histogram of oriented gradients (HOG) is a useful approach to determine the count of individual pixel values and that information can be used to filter out areas of a frame that are unneeded. This information is especially useful when determining thresholding values or applying filters. Object recognition is not the only important factor in advanced driver assistance systems. Using a stereo vision camera, a 3D point cloud can be made to determine the distance from the obstacles seen in the vehicle's path. For example, this is a technique commonly used in today’s vehicle emergency braking systems.
Access
Now
SAE MOBILUS Subscriber? You may already have access.
Buy
Attention: This item is not yet published. Pre-Order to be notified, via email, when it becomes available.
Select
Price
List
Download
$22.00
Mail
$22.00
Members save up to 36% off list price.
Share
HTML for Linking to Page
Page URL

Related Items

Standard
1988-01-01