Vehicle LED detection and segmentation recognition based on deep learning for optical camera communication

Abstract

In the vehicle to vehicle (V2V) communication based on optical camera communication (OCC) system, how to achieve high reliability and low latency communication is still a problem. In this paper, we propose a lightweight light-emitting diode (LED) detection algorithm based on deep learning to detect the vehicle LED position at different communication distances, which can improve LED detection accuracy and inference speed. In addition, we design an LED segmentation recognition algorithm to reduce the bit error rate (BER) of the vehicle OCC system. The experimental results demonstrate the effectiveness of the proposed algorithms in real traffic scenes.

References

[1] NGUYEN B L, NGO D T, DAO M N, et al. Scheduling and power control for connectivity enhancement in multi-hop I2V/V2V networks[J]. IEEE transactions on intelligent transportation systems, 2021: 1–11.

[2] TuanN M, PhuongT V, DoT H, et al.. An highly realistic optical camera communication simulation framework for internet of things applications[C], 2021, New York, IEEE: 240-242

[3] MohsanS A H. Optical camera communications: practical constraints, applications, potential challenges, and future directions[J]. Journal of optical technology, 2021, 88(12):729-741

[4] HuangZ, HeJ, YuK, et al.. Efficient demodulation scheme based on adaptive clock extraction and mapping-sampling for a mobile OCC system[J]. Applied optics, 2021, 60(12):3308-3313

[5] ThotaJ, AbdullahN F, DoufexiA, et al.. V2V for vehicular safety applications[J]. IEEE transactions on intelligent transportation systems, 2019, 21(6):2571-2585

[6] MorenoD, RufoJ, GuerraV, et al.. Optical multispectral camera communications using LED spectral emission variations[J]. IEEE photonics technology letters, 2021, 33(12):591-594

[7] IslamA, HossanM T, JangY M. Convolutional neural network scheme-based optical camera communication system for intelligent Internet of vehicles[J]. International journal of distributed sensor networks, 2018, 14(4):155014771877015

[8] ShiJ, HeJ, JiangZ, et al.. Enabling user mobility for optical camera communication using mobile phone[J]. Optics express, 2018, 26(17): 21762-21767

[9] HsuK L, ChowC W, LiuY, et al.. Rolling-shutter-effect camera-based visible light communication using RGB channel separation and an artificial neural network[J]. Optics express, 2020, 28(26):39956-39962

[10] LiuZ, GuanW, WenS. Improved target signal source tracking and extraction method based on outdoor visible light communication using an improved particle filter algorithm based on Cam-Shift algorithm[J]. IEEE photonics journal, 2019, 11(6):1-20

[11] PhamT L, ShahjalalM D, BuiV, et al.. Deep learning for optical vehicular communication[J]. IEEE access, 2020, 8: 102691-102708

[12] SunX, ShiW, ChengQ, et al.. An LED detection and recognition method based on deep learning in vehicle optical camera communication[J]. IEEE access, 2021, 9: 80897-80905

[13] HanK, WangY, TianQ, et al.. Ghostnet: more fea tures from cheap operations, 2020, Piscataway, IEEE: 1577-1586

[14] LiuS, QiL, QinH, et al.. Path aggregation network for instance segmentation, 2018, Piscataway, IEEE: 8759-8768

[15] WangC Y, BochkovskiyA, LiaoH Y M. Scaled-YOLOv4: scaling cross stage partial net-work[C], 2021, Piscataway, IEEE: 13024-13033

[16] ZhengZ, WangP, LiuW, et al.. Distance-IoU loss: faster and better learning for bounding box regression[C], 2020, Menlo Park, AAAI Press: 12993-13000

[17] BOCHKOVSKIY A, WANG C Y, LIAO H Y M. YOLOv4: optimal speed and accuracy of object detection[EB/OL]. (2020-04-23) [2022-04-25]. https://arxiv.org/abs/2004.10934?sid=V22o3y.