NOVATEUR PUBLICATIONS INTERNATIONAL JOURNAL OF RESEARCH PUBLICATIONS IN ENGINEERING AND TECHNOLOGY [IJRPET] NATIONAL CONFERENCE ON INNOVATIVE TRENDS IN ENGINEERING & TECHNOLOGY-2017 15TH & 16TH MARCH 2017 CONFERENCE PROCEEDINGS ISSN NO - 2454-7875 _______________________________________________________________________________________________________________________________________________________________________________ 60 | Page www.ijrpet.org PAPER ID: NITET14 AUTOMATIC RECOGNITION OF TRAFFIC SIGNS USING FANN AND OPENCV ANIL KANNUR Department of Computer Science & Engineeering, A.G.Patil Institute of Technology, Solapur archanapatil708@gmail.com ARCHANA S. PATIL Department of Computer Science & Engineeering, A.G.Patil Institute of Technology, Solapur archanapatil708@gmail.com NAYANA N. ARADHYE Department of Computer Science & Engineeering, A.G.Patil Institute of Technology, Solapur archanapatil708@gmail.com SUPRIYA S. MALGE Department of Computer Science & Engineeering, A.G.Patil Institute of Technology, Solapur ABSTRACT Automation Recognition of Traffic Signs is integrated and automation software for Traffic Symbol Recognition. The proposed system detects candidate regions as Maximally Stable Extremely Region (MSERs), which offers robustness to variations in lighting conditions. Recognition is based on Artificial Neural Network (ANN) classifiers. The training data are generated from real footage road signs which will be fetched using camera board and by applying threshold values we get proper training data for each frame. By applying thinning mechanism like erode and corrode and segmentation we can recognize proper shape and symbol. The proposed system is accurate at high vehicle speeds, operates under a range of weather conditions, runs at an average speed of 10 frames per second, and recognizes all classes of ideogram-based (non-text) traffic symbols from real footage road signs. Comprehensive comparative results to illustrate the performance of the system are presented. KEYWORDS: Detection, Recognition, Segmentation, FANN, OpenCV, Traffic Symbol Analysis, Thinning- erode and corrode 1. INTRODUCTION Automatic traffic sign detection and recognition is an important part of an advanced driver assistance system. Traffic symbols have several distinguishing features that may be used for their detection and identification. They are designed in specific colors and shapes, with the text or symbol in high contrast to the background. Because traffic signs are generally oriented upright and facing the camera, the amount of rotational and geometric distortion is limited. Information about traffic symbols, such as shape and color, can be used to place traffic symbols into specific groups; however, there are several factors that can hinder effective detection and recognition of traffic signs. These factors include variations in perspective, variations in illumination, occlusion of signs and deterioration of signs. Road scenes are also generally much cluttered and contain many strong geometric shapes that could easily be misclassified as road signs. Accuracy is a key consideration, because even onemisclassified or undetected sign could have an adverse impact on the driver. The proposed method consists of the following two stages: 1) Detection 2) Recognition The most common approach, quite sensibly, consists of two main stages: Detection and recognition. The detection stage identifies the regions of interest and is mostly performed using color segmentation, followed by some form of shape recognition. Detected candidates are then either identified or rejected during the recognition stage. 1) Detection is performed using a novel application ofmaximally stable extremely regions (MSERs) for this purpose color space and flood fill algorithm of segmentation is used. Before MSERs system applies thinning mechanism with erode and corrode methods for thinning captured frame. 2) Recognition is performed with the help of artificial neural network (ANN). By applying threshold value and frames per second system generate proper training data which is then passed to ANN classifier, it recognizes the symbol type and feeds it as audio notification. At the time of recognition each frame builds neural network, and training data builds train network. Frames neural network and train network is passed to ANN classifier, then it will classify images and from that we get different symbol type. In addition with ANN classifier output image segments are also used to classify different symbol type. In this paper firstly we outline the methodology used, which includes detection, recognition and the generation of synthetic data then we describe comparative results