Traffic Sign Detection and Recognition (TSDR) is a critical component of many automotive applications, such as life-saving driver assistance. Accurate 3D recognition of traffic signs is indispensable for a correct interpretation of a sign’s content. Despite its importance, the topic of traffic signs’ 3D recognition has not received adequate attention from the community. In addition to their scarcity, existing solutions present limitations that make them inappropriate for real-time fully automatic applications. This paper proposes a new approach for estimating the 3D pose of a traffic sign from a single monocular view. The proposed solution exploits the planarity of the signs, reducing their mapping to the image space to a planar homography. The estimation of the projective homography is achieved through the detection and matching of keypoint features. A thorough evaluation of 15 detectors and 19 detector/descriptor pairs is conducted to extract the best combination for each class of traffic signs. Obtained results, on challenging data generated from the German Traffic Sign Recognition benchmark (GTSRB), show the capacity of the proposed solution in dealing with the majority of traffic sign categories. We also propose amelioration to the evaluation metrics used in previous works namely the repeatability and matching scores.
AlamAJafferyZA.Indian traffic sign detection and recognition. Int J Intell Transp Syst Res2020; 18: 98–112.
2.
BaradeAPoornachandranHHarshithaKM.Automatic traffic sign recognition system using cnn. Int J Inf Retr Res2022; 12(1): 1–14.
3.
AlsibaiMHHiraiY.Real-time recognition of blue traffic signs designating directions. Int J Intell Transp Syst Res2010; 8: 96–105.
4.
YangYLuoHXuH, et al. Towards real-time traffic sign detection and classification. IEEE Trans Intell Transp Syst2016; 17(7): 2022–2031.
5.
FanBBYangH.Multi-scale traffic sign detection model with attention. Proc IMechE, Part D: J Automobile Engineering2021; 235(2–3): 708–720.
6.
YaoZSongXZhaoL, et al. Real-time method for traffic sign detection and recognition based on yolov3-tiny with multiscale feature extraction. Proc IMechE, Part D: J Automobile Engineering2021; 235(7): 1978–1991.
7.
QuXSoheilianBPaparoditisN.Vehicle localization using mono-camera and geo-referenced traffic signs. In: IEEE Intelligent Vehicles Symposium, Proceedings, Seoul, South Korea, 2015 August, pp.605–610, 2015.
8.
SoheilianBPaparoditisNValletB.Detection and 3D reconstruction of traffic signs from multiple view color images. ISPRS J Photogramm Remote Sens2013; 77: 1–20.
9.
TimofteRZimmermannKVan GoolL.Multi-view traffic sign detection, recognition, and 3d localisation. Mach Vision Appl2014; 25(3): 633–647.
10.
WangKCPHouZGongW.Automated road sign inventory system based on stereo vision and tracking. Comput Civ Infrastruct Eng2010; 25(6): 468–477.
11.
WuZ. Computer vision-based traffic sign detection and extraction: A hybrid approach using GIS and machine learning, Master’s thesis, Statesboro (GA), Georgia Southern University, 2019.
12.
LeeHSKimK.Simultaneous traffic sign detection and boundary estimation using convolutional neural network. IEEE Trans Intell Transp Syst2018; 19(5): 1652–1663.
Hemang ChawlaTJukolaM.Crowdsourced 3D mapping: a combined multi-view geometry and self-supervised learning approach. In: IEEE international conference on intelligent robots and systems, Las Vegas, NV, USA, pp.4750–4757, 2020.
15.
PedersenKFTorpK.Geolocating traffic signs using large imagery datasets. Proceedings of the 17th international symposium on spatial and temporal databases, USA, New York (NY), pp.34–43, 2021.
16.
CuiZLiuYRenF.Homography-based traffic sign localisation and pose estimation from image sequence. IET Image Process2019; 13: 2829–2839.
17.
MikolajczykKTuytelaarsTSchmidC, et al. A comparison of affine region detectors. Int J Comput Vis2005; 65: 43–72.
18.
MouatsTAoufNNamD, et al. Performance evaluation of feature detectors and descriptors beyond the visible. J Intell Robot Syst2018; 92: 33–63.
19.
BalaliVJahangiriAMachianiSG.Multi-class us traffic signs 3d recognition and localization via image-based point cloud model using color candidate extraction and texture-based recognition. Adv Eng Inform2017; 32: 263–274.
20.
WelzelAAuerswaldAWanielikG.Accurate camera-based traffic sign localization. In: 17th IEEE international conference on intelligent transportation systems, ITSCQingdao, China, pp.445–450, 2014.
21.
ZhangZ.A flexible new technique for camera calibration. IEEE Trans Pattern Anal Mach Intell2000; 22(11): 1330–1334.
22.
LiuZShenCFanX, et al. Scale-aware limited deformable convolutional neural networks for traffic sign detection and classification. IET Intell Transp Syst2020; 14(12): 1712–1722.
23.
HechriAMtibaaA.Two-stage traffic sign detection and recognition based on SVM and convolutional neural networks. IET Image Process2020; 14(5): 939–946.
24.
TorrPHSZissermanA. MLESAC: a new robust estimator with application to estimating image geometry. Comput Vis Image Underst2000; 78(1): 138–156.
25.
MikolajczykKSchmidC.An affine invariant interest point detector. In: Computer Vision—ECCV 2002: 7th European conference on computer vision Copenhagen, Denmark, May 28–31, 2002 Proceedings, Part I 7, pp.128–142, Springer, Berlin, Heidelberg, 2002.
26.
LeuteneggerSChliMSiegwartRY. Brisk: binary robust invariant scalable keypoints. In: International conference on computer vision, Barcelona, Spain, pp.2548–2555, IEEE, 2011.
27.
RostenEDrummondT. Fusing points and lines for high performance tracking. In: Tenth IEEE international conference on computer vision (ICCV’05), Beijing, China, volume 1, vol. 2, pp.1508–1515, IEEE, 2005.
28.
BayHEssATuytelaarsT, et al. Speeded-up robust features (surf). Comput Vis Image Underst2008; 110(3): 346–359.
29.
LoweDG.Distinctive image features from scale-invariant keypoints. Int J Comput Vis2004; 60: 91–110.
30.
RubleeERabaudVKonoligeK, et al. Orb: an efficient alternative to sift or surf. In: International conference on computer vision, Barcelona, Spain, pp.2564–2571, IEEE, 2011.
31.
AlcantarillaPFBartoliADavisonAJ.Kaze features. In: Computer Vision–ECCV 2012: 12th European conference on computer vision, Florence, Italy, October 7–13, 2012, Proceedings, Part VI 12, pp.214–227, Springer, Berlin, Heidelberg, 2012.
32.
ShiJTomasiC. Good features to track. In: Proceedings of IEEE conference on computer vision and pattern recognition, Seattle, WA, USA, pp.593–600, IEEE, 1994.
33.
MikolajczykKSchmidC.Performance evaluation of local descriptors. IEEE Trans Pattern Anal Mach Intell2005; 27(10): 1615–1630.
34.
AraarOAmamraAAbdeldaimA, et al. Traffic sign recognition using a synthetic data training approach. Int J Artif Intell Tools2020; 29(05): 2050013.
