Yukio Fukuzawa (2013). Traffic Sign Recognition system on Android devices (Unpublished bachelor thesis). Massey University, Auckland, New Zealand.
You are welcome to download and read this document. I especially welcome feedback on it. As it is not yet published in final form, if you want to cite the paper, please check with me first. Thanks.
This project implements on an Android device a Traﬃc Sign Recognition(TSR) system capable of recognising 85 New Zealand traﬃc signs with variouscolours and shapes. I ﬁnd it possible to have such system running almost realtime on medium-class Android devices by having the core processing modulewritten in native code, interacting with the user interface via JNI calls. Dif-ferent techniques are used to extract sign candidates from raw images beforethey are classiﬁed to the correct class. The ﬁrst technique uses a colour ﬁlteroperating on either HSV or RGB colour space to ﬁlter out non-sign pixelsand the sign detector is applied on the remaining part of the images. Thesecond technique applies the detector on two consecutive raw images withoutcolour-ﬁltering. Sign detector is a AdaBoost classiﬁer using Linear BinaryPattern (LBP) to extract features. Products of sign detector are called ”signcandidates” and recognised by a back propagating neural network. This reportﬁnds the colour-ﬁltering technique with the ﬁltering threshold carefully tunedoverperforms the cascade detecting technique, but the latter is less likely tomiss a sign in diﬀerent lighting conditions.