An algorithm for parallel processing of traffic signs video on a graphics processor

Abstract

In the lighting conditions such as snowing, hazing, raining, and weak lighting condition, the accuracy of traffic sign recognition is not very high. It is important to develop an algorithms for real-time fast detection of road signs from video images and use them in driver assistance systems. Firstly, we image preprocessing by using dark channel prior based image dehazing for remove noises from the input image. In this article, we develop a transposed image filtering method combined with other fast parallel filtering algorithms. We propose a high-speed and high-accuracy method by improving image dehazing algorithms and analyzing GPU architecture. In this, we quickly removing fog from HD video images using DCP with image quality and visual effects.