Application and evaluation about obstacle edge extraction technology in the parking assistant system

Abstract

The obstacle edge detection technology used in the parking assistant system helps drivers avoid obstacles, especially for inexperienced drivers. Reversing image processing, as part of smart driving, must meet the parking requirements of real-time and accuracy. Processing system should convert the color images from CCD camera to grayscale, abate noise and immunity with non-linear median filtering. And the Sobel operator that of real-time and accuracy is used in edge detection; adaptive valve segmentation technique separates these points in time for improving Hough transform identification. In order to identify the obstacles for drivers in reversing at once, it also need to be given the edge lines thicker, warning color and superimposed displayed with the original image on the terminal screen, so that the whole process can make the parking assistant system more accurate, consistent and quickly to show information.

Keywords:

• Parking assistant system
• Obstacles edge detection
• Hough transform
• Sobel operator
• Subjective evaluation