Enhancement of Road Defect Detection Using First Level Algorithm Based on Channel Shuffling and Multi Scale Feature Fusion

Authors: Yifan Hou, Haibo Liu, Le Jiang, Wandong Su, Binqing Wang

Abstract:

Road defect detection is crucial for modern urban management and infrastructure maintenance. Traditional road defect detection methods mostly rely on manual labor, which is not only inefficient but also difficult to ensure their reliability. However, existing deep learning-based road defect detection models have poor detection performance in complex environments and lack robustness to multi-scale targets. To address this challenge, this paper proposes an innovative detection framework based on the One Stage algorithm network structure. This article designs a deep feature extraction network based on RCSDarknet, which applies channel shuffling to enhance information fusion between tensors. Through repeated stacking of RCS (Reparametrized Convolution based on channel Shuffle) modules, the information flow between different channels of adjacent layer features is enhanced to improve the model's ability to capture target spatial features. In addition, a multi-scale feature fusion mechanism with weighted dual flow paths was adopted to fuse spatial features of different scales, thereby further improving the detection performance of the model at different scales. To validate the performance of the proposed algorithm, we tested it using the RDD2022 dataset. The experimental results show that the enhancement algorithm achieved 84.14% mAP, which is 1.06% higher than the currently advanced YOLOv8 algorithm. Through visualization analysis of the results, it can also be seen that our proposed algorithm has good performance in detecting targets of different scales in complex scenes. The above experimental results demonstrate the effectiveness and superiority of the proposed algorithm, providing valuable insights for advancing real-time road defect detection methods.

Keywords: Road defect detection, defect detection, visualization, deep learning.

References:

[1] J. C. Pais, S. I. R. Amorim, and M. J. C. Minhoto, "Impact of traffic overload on road pavement performance," J. Transp. Eng., vol. 139, no. 9, pp. 873 - 879, 2013.
[2] R. Madli, S. Hebbar, P. Pattar, et al., "Automatic detection and notification of potholes and humps on roads to aid drivers," IEEE Sens. J., vol. 15, no. 8, pp. 4313 - 4318, 2015.
[3] Y. Gao, H. Cao, W. Cai, et al., "Pixel-level road crack detection in UAV remote sensing images based on ARD - Unet," Measurement, vol. 219, 2023.
[4] M. - T. Cao, Q. - V. Tran, N. - M. Nguyen, and K. - T. Chang, "Survey on performance of deep learning models for detecting road damages using multiple dashcam image resources," Adv. Eng. Informat., vol. 46, Oct. 2020.
[5] C. Zhang, E. Nateghinia, L. F. Miranda - Moreno, and L. Sun, "Pavement distress detection using convolutional neural network (CNN): A case study in Montreal Canada," Int. J. Transp. Sci. Technol., vol. 11, no. 2, pp. 298 - 309, Jun. 2022.
[6] S. Huang, W. Tang, G. Huang, L. Huangfu, and D. Yang, "Weakly supervised patch label inference networks for efficient pavement distress detection and recognition in the wild," IEEE Trans. Intell. Transp. Syst., vol. 24, no. 5, pp. 5216 - 5228, May 2023.
[7] W. Tang, S. Huang, Q. Zhao, R. Li, and L. Huangfu, "An iteratively optimized patch label inference network for Automatic pavement distress detection," IEEE Trans. Intell. Transp. Syst., vol. 23, no. 7, pp. 8652 - 8661, Jul. 2022.
[8] B. Wan et al., "LFRNet: Localizing focus and refinement network for salient object detection of surface defects," IEEE Trans. Instrum. Meas., vol. 72, pp. 1 - 12, 2023.
[9] Y. Wang, M. Niu, K. Song, P. Jiang, and Y. Yan, "Normal - knowledge - based pavement defect segmentation using relevance - aware and cross - reasoning mechanisms," IEEE Trans. Intell. Transp. Syst., vol. 24, no. 4, pp. 4413 - 4427, Apr. 2023.
[10] Z. Qu, C. Cao, L. Liu, and D. - Y. Zhou, "A deeply supervised convolutional neural network for pavement crack detection with multiscale feature fusion," IEEE Trans. Neural Netw. Learn. Syst., vol. 33, no. 9, pp. 4890 - 4899, Sep. 2022.
[11] Q. Zhou, Z. Qu, Y. Li, and F. Ju, "Tunnel crack detection with linear seam based on mixed attention and multiscale feature fusion," IEEE Trans. Instrum. Meas., vol. 71, pp. 1 - 11, 2022.
[12] Y. Li, C. Liu, Y. Shen, J. Cao, S. Yu, and Y. Du, "RoadID: A dedicated deep convolutional neural network for multipavement distress detection," J. Transp. Eng. B Pavements, vol. 147, no. 4, Dec. 2021.
[13] B. Su, H. Zhang, Z. Wu, and Z. Zhou, "FSRDD: An efficient few - shot detector for rare city road damage detection," IEEE Trans. Intell. Transp. Syst., vol. 23, no. 12, pp. 24379 - 24388, Dec. 2022.
[14] C. Lin, D. Tian, X. Duan, J. Zhou, D. Zhao, and D. Cao, "DA - RDD: Toward domain adaptive road damage detection across different countries," IEEE Trans. Intell. Transp. Syst., vol. 24, no. 3, pp. 3091 - 3103, Mar. 2023.
[15] D. Arya et al., "Deep learning - based road damage detection and classification for multiple countries," Autom. Construct., vol. 132, Dec. 2021.
[16] L. Wang, J. Duan, and L. Xin, "YOLOv5 helmet weardetection method with introduction of attention mechanism," Computer Engineering and Applications, vol. 58, no. 9, pp. 303 - 312, 2022.
[17] C. Cheng, W. Chen, and J. Chen, "YOLOimprove detection method of real - time mask wearing," Electronic Science and Technology, vol. 36, no. 2, pp. 73 - 80, 2023.
[18] M. Rojo, H. Gonzalo - Orden, A. Linares, et al., "Impact of a lower conservation budget on road safety indices," J. Adv. Transp., pp. 1 - 9, 2018.
[19] J. Zhang, Y. Wang, and Y. Wang, "Defect detection of aluminum alloy wheels in radiography images using adaptive threshold and morphological reconstruction," Appl Sci., vol. 8, no. 12, 2018.
[20] C. Spinola, S. Battiato, and G. Gallo, "Real - time image processing for edge inspection and defect detection in stainless steel production lines," Proc IEEE Int Conf Imaging Syst Tech (IST), pp. 170 - 175, 2011.
[21] R. Borwankar and R. Ludwig, "An optical surface inspection and Automatic classification technique using the rotated wavelet transform," IEEE Trans. on Instrum Meas., pp. 692 - 697, 2018.
[22] M. Ke, V. T. F. Yago, D. Samaras, M. Petrucci, and D. L. Magnus, "Texture classification for rail surface condition evaluation," in 2016 IEEE Winter Conference on Applications of Computer Vision, WACV, IEEE, 2016, pp. 1 - 9.
[23] Q. Li and S. Ren, "A real - time visual inspection system for discrete surface defects of rail heads," IEEE Trans. Instrum. Meas., vol. 61, no. 8, pp. 2189 - 2199, 2012.
[24] S. Zhai, D. Shang, S. Wang, et al., "DF - SSD: an improved SSD object detection algorithm based on DenseNet and feature fusion," IEEE Access, vol. 8, pp. 24344 - 24357, 2020.
[25] J. Cao, H. Cholakkal, R. Anwer, et al., "D2det: towards high quality object detection and instance segmentation," IEEE/CVF CVPR, pp. 11482 - 11491, 2020.
[26] A. Bochkovskiy, C. Wang, and H. Liao, "YOLOv4: Optimal speed and accuracy of object detection," arXiv: 2004.10934, 2020.
[27] Z. Ge, S. Liu, F Wang, Z. Li, and J Sun, "YOLOx: Exceeding yolo series in 2021," arXiv: 2107.08430, 2021.
[28] C. Wang, I. Yeh, and H. Liao, "You only learn one representation: Unified network for multiple tasks," arXiv: 2105.04206, 2021.
[29] L. Zhang, F. Yang, Y. D. Zhang, and Y. J. Zhu, "Road crack detection using deep convolutional neural network," Proc. IEEE Int. Conf. Image Process. (ICIP), pp. 3708 - 3712, Sep. 2016.
[30] S. Zhong, C. Chen, W. Luo, and S. Chen, "DDVC - YOLOv5: An Improved YOLOv5 Model for Road Defect Detection," in IEEE Access, vol. 12, pp. 134008 - 134019, 2024, doi: 10.1109/ACCESS.2024.3453914.
[31] W. Wang, B. Wu, S. Yang, et al., "Road damage detection and classification with faster R - CNN," in Proceedings of the 2018 IEEE International Conference on Big Data (Big data), 2018. IEEE.
[32] Q. Chen, X. Gan, W. Huang, et al., "Road damage detection and classification using mask R - CNN with DenseNet backbone," Comput. Mater. Continua, vol. 65, no. 3, pp. 2201 - 2215, 2020.
[33] F. Wan, C. Sun, H. He, G. Lei, L. Xu, and T. Xiao, "YOLO - LRDD: A lightweight method for road damage detection based on improved YOLOv5s," EURASIP J. Adv. Signal Process., vol. 2022, no. 1, pp. 98, Oct. 2022.
[34] L. Shen, H. Tao, Y. Ni, Y. Wang, and V. Stojanovic, "Improved YOLOv3 model with feature map cropping for multi - scale road object detection," Meas. Sci. Technol., vol. 34, no. 4, Jan. 2023.
[35] X. Zhu, S. Lyu, X. Wang, and Q. Zhao, "TPH - YOLOv5: Improved YOLOv5 based on transformer prediction head for object detection on drone - captured scenarios," Proc. IEEE/CVF Int. Conf. Comput. Vis. Workshops (ICCVW), pp. 2778 - 2788, Oct. 2021.
[36] D. Arya, H. Maeda, S. K. Ghosh, D. Toshniwal, H. Omata, T. Kashiyama, et al., "Global road damage detection: State - of - the - art solutions," Proc. IEEE Int. Conf. Big Data, pp. 5533 - 5539, Dec. 2020.