A Traffic Speed Limit Signs Detection and Recognition System for iOS Devices

• Main Authors: JIANG, CHENG-ZHE, 江丞哲
• Other Authors: CHEN, CHING-YUNG
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/u74775

碩士 === 國立高雄第一科技大學 === 電腦與通訊工程系碩士班 === 105 === Mobile devices, such as smart phones and tablet PCs, have recently become popular in our daily life. Various kinds of sensors embedded in the mobile devices make it possible and straightforward to develop Apps that can perceive the environment and (help users) take actions appropriately. Specifically, more and more Apps that run on mobile devices equip with abilities of multimedia processing and artificial intelligence (e.g., some Apps are developed for detecting and/or recognizing targets of interest either in a photo or in the real world). Aiming on intelligently reminding drivers the speed limits of the road via an App, this paper proposes a traffic speed limit signs detection and recognition system based on image processing. The system consists of two main modules: traffic speed limit signs detection and recognition. The first module, to detect the traffic speed limit signs as regions of interest (ROIs) in an image with complex background, cascades several image processing techniques such as color space conversion, color segmentation and shape detection. It has been shown that both color space and shape features are adopted can achieve better detection result than that with only color space or shape feature. In the second module, the detected ROIs are further processed with character segmentation and classification techniques to recognize the speed limit information. Some experimental results have been obtained to support the efficacy of the proposed system. Moreover, the proposed system has also been implemented as an iOS App which takes images continuously from the embedded rear camera as inputs to the proposed system. In the App implementation, the recognized speed limit information is presented through both UI and voice messages to remind the drivers.