Advanced Driver Assistance System

Home Services Case studies Advanced Driver Assistance System
PDF Print
Verticals: Traffic and Transportation
Technologies: Object Detection & Recognition,Optical Character Recognition
Tags: Advanced Driver assistance, OCR, Image Segmentation, Adaptive Thresholding, shape detection,

CLIENT

Client is a multi-national corporation which provides Geospacial, Engineering and Information Technology Solutions to the Utility, Telecom, Energy, Government, Oil & gas and Petrochemical sectors.

OBJECTIVE

Automatic identification and classification of road signs to direct driver accordingly and hence achieve a vigilent drive. three basic signs were employed including 'No-Right Turn', 'No-left Turn and Stop. 

 KRITIKAL'S ROLE

  • Image segmentation based on color slicing, adaptive thresholding and shape detection.
  • Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) for classified sign detection.
  • Color notations used to display the signs, in particular Green for 'No-Right Turn, Blue for 'No-Left' Turn and Red for 'Stop'.

The process can be represented by following flow chart:

 


 

 

PROJECT SCOPE

Design a PC based software to

  • Identify basic road signs in various light conditions.
  • Detect road signs of various shapes and sizes.
  • Distinguish and identify road signs with various colors.

CHALLENGES

  •  Various lighting condition:
    • Very Dark conditions
    • Very Bright Conditions
  • Variable sizes of Road signs: It ranges from 35x35 to 210x210.
  • Variable shapes of road signs: Road signs may not necessarily be circular. The shape of the signs depend on the angle at which images of the signs have been taken.
  • Variable Colors: Road signs like 'No-Right' turn,'No-Left' Turn and 'STOP' invariably appear in different shades of Red color. At a particular instance the shade depends on the lighting conditions, distance of the object from camera etc.
  • Low contrast of Input Images: It may be due to motion blur, lighting conditions etc.
< return to the case studies