Real-time detection of Sri Lankan road signs with intelligent speed assistant

Abstract

Road Sign detection and recognition systems provide an additional level of driver assistance, leading to improved safety for passengers, road users and vehicles. It can be used to benefit drivers by alerting them about the presence of road signs to reduce risks in situations of driving distraction, fatigue, poor sight and weather conditions. Although a number of systems have been proposed in literature; the design of a robust algorithm still remains an open research problem. This thesis aims to resolve some of the outstanding research challenges in road sign recognition systems, while considering variations in colour illumination, scale, rotation, translation, occlusion, computational complexity and functional limitations. Real-time detection of Sri Lankan road signs with intelligent speed assistant system pipeline is divided into few main parts namely; Image pre-processing, Segmentation, Classification and Speed assistance. This thesis presents a detailed investigation on above mentioned areas. A colour segmentation algorithm is developed by combining HSV and YCrCb colour spaces. Colour segmentation algorithm is robust as it was tested in different lighting conditions, weather conditions, and times of the day. Approximately 96% successful segmentation rate was achieved using this algorithm. Based on two shape measures - the circle and octagon, rules were developed to determine the shape of the sign. Among these shape measures octagon detection algorithm has been introduced in this research. Circle detection has been conducted by improved the Hough Transform algorithm. The final decision of the road sign detection is based on the combination of both the colour and shape of the sign. The recogniser was tested in a variety of testing conditions giving an overall performance of approximately 78%. Classification was undertaken using a neural network. The classification is carried out in two parts: stop sign classification and speed sign classification. The average classification rate achieved is about 89%. Presented work has achieved 88% of an accuracy rate along with an average efficiency rate of 0.0467 seconds for recognizing six different speed signs and stop sign according to the Sri Lankan road sign context. Key words: Road sign, segmentation, classification, neural network, colour space