Kalman filter-based tracking system for automated inventory of roadway signs

Roadway signs represent a substantial investment of public money in road and highway infrastructure. However, the current level of automation in sign identification and recognition, size dimensioning, and location identification is unsatisfactory. In an effort to improve the automation level of sign inventory, feature extraction and Kalman filter-based tracking techniques for road signs in right-of-way (ROW) images are developed. A framework that combines the conventional image-processing methods with the Kalman filter tracking method is applied to improve the accuracy and efficiency of ROW image processing. With this tracking technique, the candidate region of the road sign in an image can be predicted on the basis of the image in the previous frame. With image processing used near the candidate region of a sign, detection efficiency and accuracy can be improved. The methodologies described fit a dynamic and moving environment, appropriate for a highway survey vehicle.