{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,24]],"date-time":"2026-02-24T19:08:05Z","timestamp":1771960085309,"version":"3.50.1"},"reference-count":37,"publisher":"MDPI AG","issue":"20","license":[{"start":{"date-parts":[[2022,10,13]],"date-time":"2022-10-13T00:00:00Z","timestamp":1665619200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Young Elite Scientists Sponsorship Program by CAST","award":["(2021QNRC001)"],"award-info":[{"award-number":["(2021QNRC001)"]}]},{"name":"Young Elite Scientists Sponsorship Program by CAST","award":["(Grant No. 52202451)"],"award-info":[{"award-number":["(Grant No. 52202451)"]}]},{"name":"National Natural Science Foundation of China","award":["(2021QNRC001)"],"award-info":[{"award-number":["(2021QNRC001)"]}]},{"name":"National Natural Science Foundation of China","award":["(Grant No. 52202451)"],"award-info":[{"award-number":["(Grant No. 52202451)"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Sensors"],"abstract":"<jats:p>To resolve the issues of a deep backbone network, a large model, slow reasoning speed on a mobile terminal, low detection accuracy for small targets and difficulties detecting and recognizing traffic lights in real time and accurately with YOLOv4, a traffic lights recognition method based on improved YOLOv4 is proposed. The lightweight ShuffleNetv2 network is utilized to replace the CSPDarkNet53 network of YOLOv4 to satisfy the requirements of a mobile terminal. The reformed k-means clustering algorithm is applied to generate anchor boxes for avoiding the sensitivity issue of outliers and initial values. A novel attention mechanism named CS2A is added to enhance the extraction capability of effective features. Multiple data augmentation methods are combined to improve the generalization ability of the model. Ultimately, the detection and recognition of traffic lights can be realized. The S2TLD dataset is selected for training and testing, and it can be proved that the recognition accuracy and model size are greatly optimized. Meanwhile, a self-made dataset is selected for training and testing. Compared with the conventional YOLOv4, the recognition accuracy of the proposed algorithm for traffic lights\u2019 state information increases by 1.79%, and the model size decreases by 81.97%. Appropriate scenes are selected for real-vehicle testing and the results demonstrate that the detection speed of the presented algorithm increases by 16%, and the recognition effect for small targets increases by 37% in comparison with conventional YOLOv4.<\/jats:p>","DOI":"10.3390\/s22207787","type":"journal-article","created":{"date-parts":[[2022,10,14]],"date-time":"2022-10-14T01:44:13Z","timestamp":1665711853000},"page":"7787","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":10,"title":["Study on Detection and Recognition of Traffic Lights Based on Improved YOLOv4"],"prefix":"10.3390","volume":"22","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-0096-5614","authenticated-orcid":false,"given":"Ying","family":"Zhao","sequence":"first","affiliation":[{"name":"College of Engineering and Technology, Southwest University, Chongqing 400715, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yiyuan","family":"Feng","sequence":"additional","affiliation":[{"name":"College of Engineering and Technology, Southwest University, Chongqing 400715, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yueqiang","family":"Wang","sequence":"additional","affiliation":[{"name":"Department of Autonomous Driving, Changan Research Institute of Automotive Engineering, Chongqing 400023, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhihan","family":"Zhang","sequence":"additional","affiliation":[{"name":"College of Engineering and Technology, Southwest University, Chongqing 400715, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhihao","family":"Zhang","sequence":"additional","affiliation":[{"name":"College of Engineering and Technology, Southwest University, Chongqing 400715, China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1968","published-online":{"date-parts":[[2022,10,13]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","unstructured":"Kai, Y. (2020, January 23\u201325). Intelligent Traffic Control System Based on Visible Light Communication. Proceedings of the 2020 IEEE Eurasia Conference on IOT, Communication and Engineering (ECICE), Yunlin, Taiwan.","DOI":"10.1109\/ECICE50847.2020.9301946"},{"key":"ref_2","doi-asserted-by":"crossref","unstructured":"Behrendt, K., Novak, L., and Botros, R. (June, January 29). A deep learning approach to traffic lights: Detection, tracking, and classification. Proceedings of the 2017 IEEE International Conference on Robotics and Automation (ICRA), Singapore.","DOI":"10.1109\/ICRA.2017.7989163"},{"key":"ref_3","first-page":"143","article-title":"Traffic Light Detection Based on YOLOv3 Optimization Algorithm","volume":"40","author":"Sun","year":"2020","journal-title":"Acta Opt. Sin."},{"key":"ref_4","doi-asserted-by":"crossref","unstructured":"Wonghabut, P., Kumphong, J., Ung-arunyawee, R., Leelapatra, W., and Satiennam, T. (2018, January 16). Traffic Light Color Identification for Automatic Traffic Light Violation Detection System. Proceedings of the 2018 International Conference on Engineering, Applied Sciences, and Technology (ICEAST), Phuket, Thailand.","DOI":"10.1109\/ICEAST.2018.8434400"},{"key":"ref_5","first-page":"34","article-title":"Traffic Light Detection and Recognition Based on Improved Lightweight YOLOv3","volume":"20","author":"Shao","year":"2021","journal-title":"J. Nantong Univ. (Nat. Sci. Ed.)"},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"159","DOI":"10.1109\/TCI.2015.2480006","article-title":"Saliency Map Generation by the Convolutional Neural Network for Real-Time Traffic Light Detection Using Template Matching","volume":"1","author":"John","year":"2015","journal-title":"IEEE Trans. Comput. Imaging"},{"key":"ref_7","first-page":"137","article-title":"A Traffic Light Recognition Method Based on Image Enhancement","volume":"45","author":"Liu","year":"2022","journal-title":"Electron. Meas. Technol."},{"key":"ref_8","first-page":"1526","article-title":"Nuclear power plant thermocouple sensor-fault detection and classification using deep learning and generalized likelihood ratio test","volume":"6","author":"Mandal","year":"2017","journal-title":"IEEE Trans. Nucl. Sci."},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"4827","DOI":"10.1109\/JSEN.2020.3029459","article-title":"Sensor-Fault Detection, Isolation and Accommodation for Digital Twins via Modular Data-Driven Architecture","volume":"21","author":"Darvishi","year":"2021","journal-title":"IEEE Sens. J."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"400","DOI":"10.1007\/s42154-021-00157-x","article-title":"VCANet: Vanishing-Point-Guided Context-Aware Network for Small Road Object Detection","volume":"4","author":"Chen","year":"2021","journal-title":"Automot. Innov."},{"key":"ref_11","first-page":"960","article-title":"Traffic Light Detection Based on Genetic Optimization and Deep Learning","volume":"41","author":"Xiong","year":"2019","journal-title":"Automot. Eng."},{"key":"ref_12","first-page":"272","article-title":"Fast Detection and Identification of Traffic Lights Based on Deep Learning","volume":"46","author":"Qian","year":"2019","journal-title":"Comput. Sci."},{"key":"ref_13","first-page":"45","article-title":"Object Detection Method Based on Improved YOLO Lightweight Network","volume":"57","author":"Li","year":"2020","journal-title":"Laser Optoelectron. Prog."},{"key":"ref_14","first-page":"109","article-title":"Research on Automatic Classification of Optical Tomography Retina Image Based on Multi-Channel","volume":"48","author":"Chen","year":"2021","journal-title":"Chin. J. Lasers"},{"key":"ref_15","first-page":"149","article-title":"Detection and recognition of small scale traffic lights","volume":"41","author":"Wang","year":"2022","journal-title":"Transducer Microsyst. Technol."},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Wang, Q., Zhang, Q., Liang, X., Wang, Y., Zhou, C., and Mikulovich, V.I. (2022). Traffic Lights Detection and Recognition Method Based on the Improved YOLOv4 Algorithm. Sensors, 22.","DOI":"10.3390\/s22010200"},{"key":"ref_17","first-page":"148","article-title":"Digital recognition of LED lights based on convolutional neural networks","volume":"34","author":"Wang","year":"2020","journal-title":"J. Electron. Meas. Instrum."},{"key":"ref_18","first-page":"265","article-title":"Light-weight Object Detection Network Optimized Based on YOLO Family","volume":"48","author":"Xu","year":"2021","journal-title":"Sch. Electron. Sci. Eng."},{"key":"ref_19","first-page":"1620","article-title":"A Real-Time Traffic Light Detection Algorithm Based on Adaptive Edge Information","volume":"1","author":"Yu","year":"2018","journal-title":"SAE Tech. Pap."},{"key":"ref_20","first-page":"25","article-title":"Lung Nodule Detection System Based on Data Augmentation and Attention Mechanism","volume":"7","author":"Li","year":"2022","journal-title":"J. Beijing Univ. Posts Telecommun."},{"key":"ref_21","first-page":"313","article-title":"Target Detection Algorithm Based on Improved YOLO v3","volume":"57","author":"Zhao","year":"2020","journal-title":"Laser Optoelectron. Prog."},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Wang, C.Y., Liao, H.Y., Wu, Y.H., Chen, P.Y., Hsieh, W.J., and Yeh, I.H. (2020, January 28). CSPNet: A New Backbone that can Enhance Learning Capability of CNN. Proceedings of the 2020 IEEE\/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Seattle, WA, USA.","DOI":"10.1109\/CVPRW50498.2020.00203"},{"key":"ref_23","unstructured":"Misra, D. (2019). Mish: A Self Regularized Non-Monotonic Neural Activation Function. arXiv."},{"key":"ref_24","first-page":"179","article-title":"Application of improved YOLOv4 in aircraft skin damage detection","volume":"37","author":"Nong","year":"2022","journal-title":"J. Navy Aviat. Univ."},{"key":"ref_25","doi-asserted-by":"crossref","unstructured":"Liu, S., Qi, L., Qin, H., Shi, J.P., and Jia, J.Y. (2018, January 16). Path Aggregation Network for Instance Segmentation. Proceedings of the 2018 IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA.","DOI":"10.1109\/CVPR.2018.00913"},{"key":"ref_26","doi-asserted-by":"crossref","unstructured":"Zhang, X., Zhou, X., Lin, M., and Sun, J. (2018, January 16). Shufflenet: An extremely efficient convolutional neural network for mobile devices. Proceedings of the 2018 IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA.","DOI":"10.1109\/CVPR.2018.00716"},{"key":"ref_27","doi-asserted-by":"crossref","unstructured":"Ma, N., Zhang, X., Zheng, H.T., and Sun, J. (2018). ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design. arXiv.","DOI":"10.1007\/978-3-030-01264-9_8"},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"396","DOI":"10.1007\/s11431-021-1950-9","article-title":"Real-time detection network for tiny traffic sign using multi-scale attention module","volume":"65","author":"Yang","year":"2022","journal-title":"Sci. China (Technol. Sci.)"},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Hou, Q.B., Zhou, D.Q., and Feng, J.S. (2021, January 2). Coordinate Attention for Efficient Mobile Network Design. Proceedings of the 2021 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Nashville, TN, USA.","DOI":"10.1109\/CVPR46437.2021.01350"},{"key":"ref_30","doi-asserted-by":"crossref","unstructured":"Woo, S., Park, J., Lee, J.Y., and Kweon, I.S. (2018, January 6). CBAM: Convolutional Block Attention Module. Proceedings of the European Conference on Computer Vision, Munich, Germany.","DOI":"10.1007\/978-3-030-01234-2_1"},{"key":"ref_31","first-page":"120","article-title":"Improved YOLOv4\u2019s algorithm for detecting defects on the sealing surface of inner wire joints","volume":"36","author":"Dan","year":"2022","journal-title":"J. Electron. Meas. Instrum."},{"key":"ref_32","doi-asserted-by":"crossref","unstructured":"Ghiasi, G., Cui, Y., Srinivas, A., Qian, R., Lin, T.Y., Cubuk, E.D., Le, Q.V., and Zoph, B. (2021, January 2). Simple Copy-Paste is a Strong Data Augmentation Method for Instance Segmentation. Proceedings of the 2021 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Nashville, TN, USA.","DOI":"10.1109\/CVPR46437.2021.00294"},{"key":"ref_33","doi-asserted-by":"crossref","unstructured":"Zhang, H., Cisse, M., Dauphin, Y.N., and Lopez-Paz, D. (2017). mixup: Beyond Empirical Risk Minimization. arXiv.","DOI":"10.1007\/978-1-4899-7687-1_79"},{"key":"ref_34","doi-asserted-by":"crossref","unstructured":"Wang, C.Y., Bochkovskiy, A., and Liao, H. (2021, January 2). Scaled-YOLOv4: Scaling Cross Stage Partial Network. Proceedings of the 2021 IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Nashville, TN, USA.","DOI":"10.1109\/CVPR46437.2021.01283"},{"key":"ref_35","doi-asserted-by":"crossref","unstructured":"Howard, A., Sandler, M., Chen, B., Wang, W.J., Chen, L.C., Tan, M.X., Chu, G., Vasudevan, V., Zhu, Y.K., and Pang, R.M. (2020, January 27). Searching for MobileNetV3. Proceedings of the 2019 IEEE\/CVF International Conference on Computer Vision (ICCV), Seoul, Korea.","DOI":"10.1109\/ICCV.2019.00140"},{"key":"ref_36","doi-asserted-by":"crossref","first-page":"52138","DOI":"10.1109\/ACCESS.2018.2870052","article-title":"Peeking inside the black-box: A survey on explainable artificial intelligence (XAI)","volume":"6","author":"Adadi","year":"2018","journal-title":"IEEE Access"},{"key":"ref_37","doi-asserted-by":"crossref","first-page":"328","DOI":"10.1007\/s42154-021-00151-3","article-title":"End-to-End Autonomous Driving Through Dueling Double Deep Q-Network","volume":"4","author":"Peng","year":"2021","journal-title":"Automot. Innov."}],"container-title":["Sensors"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1424-8220\/22\/20\/7787\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,11]],"date-time":"2025-10-11T00:53:42Z","timestamp":1760144022000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1424-8220\/22\/20\/7787"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,10,13]]},"references-count":37,"journal-issue":{"issue":"20","published-online":{"date-parts":[[2022,10]]}},"alternative-id":["s22207787"],"URL":"https:\/\/doi.org\/10.3390\/s22207787","relation":{},"ISSN":["1424-8220"],"issn-type":[{"value":"1424-8220","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,10,13]]}}}