APP下载

基于M-Unet的混凝土裂缝实时分割算法

2024-01-03孟庆成李明健万达胡垒吴浩杰齐欣

土木建筑与环境工程 2024年1期
关键词:深度学习

孟庆成 李明健 万达 胡垒 吴浩杰 齐欣

摘要:針对主流深度学习裂缝分割算法消耗大量计算资源、传统图像处理方法检测精度低、丢失裂缝特征等问题,为了实现对混凝土裂缝的实时检测和在像素级水平上分割裂缝,提出一种基于轻量级卷积神经络M-Unet的裂缝语义分割模型,首先对MobileNet_V2轻量网络进行改进,修剪其网络结构并优化激活函数,再用改进的MobileNet_V2替换U-Net参数量巨大的编码器部分,以实现模型的轻量化并提升裂缝的分割效果。构建包含5 160张裂缝图像的SegCracks数据集对提出方法进行验证,试验结果表明:优化后的M-Unet裂缝分割效果优于U-Net、FCN8和SegNet等主流分割网络和传统图像处理技术,获得的IoU_Score为96.10%,F1_Score为97.99%。与改进前U-Net相比,M-Unet权重文件大小减少了7%,迭代一轮时间和预测时间分别缩短了63.3%和68.6%,IoU_Score和F1_Score分别提升了5.79%和3.14%,并且在不同开源数据集上的交叉验证效果良好。表明提出的网络具有精度高、鲁棒性好和泛化能力强等优点。

关键词:混凝土裂缝;卷积神经网络;深度学习;裂缝检测;裂缝分割

中图分类号:TU755.7     文献标志码:A     文章编号:2096-6717(2024)01-0215-08

Real-time segmentation algorithm of concrete cracks based on M-Unet

MENG Qingcheng1, LI Mingjian1, WAN Da1, HU Lei1, Wu Haojie1, QI xin2

(1. School of Civil Engineering and Geomatics, Southwest Petroleum University,Chengdu 610500;

2.School of Civil Engineering, Southwest Jiaotong University, Chengdu 610031)

Abstract: Mainstream deep learning algorithm for crack segmentation consumes a lot of computing resources while the traditional image processing methods are of low detection accuracy and lost crack features. In order to realize the real-time detection of concrete cracks and the segmentation of cracks at the pixel level, a crack semantic segmentation model based on lightweight convolutional neural network M-Unet is proposed. Firstly, the MobileNet_V2 lightweight network is improved, its network structure is trimmed and the activation function is optimized, and then the encoder part with huge parameters of U-Net is replaced by the improved MobileNet_V2 to realize the lightweight of the model and improve the segmentation effect of cracks. The SegCracks data set containing 5 160 crack images is constructed to verify the proposed method. The experimental results show that the crack segmentation effect of the optimized M-Unet is better than the mainstream segmentation networks of U-Net, FCN8 and SegNet and the traditional image processing techniques, the obtained IoU_Score is 96.10%, F1_Score is 97.99%. Compared with the original U-Net, the weight file size M-Unet is reduced by 7 %, the iteration time and prediction time are reduced by 63.3% and 68.6% respectively, and the IoU_Score and F1_Score are increased by 5.79 % and 3.14 % respectively. The cross validation results on different open source data sets are good, which shows that the proposed network has the advantages of high accuracy, good robustness and strong generalization ability.

Keywords: concrete cracks; convolutional neural network; deep learning; crack detection; crack segmentation

在混凝土自身材料特性影响及多重因素作用下,混凝土结构服役期间会不可避免地产生裂缝。裂缝是混凝土结构病害的一种初期表现,若初期裂缝得不到即时修补,在外力的作用下,裂缝的长度、宽度、深度会不断地增加,使得混凝土构件的有效截面面积减小,承载力降低,继而可能会导致事故的发生[1]。早期的裂缝检测主要是人工检测方法,人工检测可以直接对裂缝的特征进行判断,较为简便灵活,但人工检测的结果受限于主观判断,依赖检测人员的经验储备,且人力物力消耗大,容易发生安全事故,导致人工检测难以完成大量的裂缝检测任务[2]。……

登录APP查看全文

猜你喜欢

深度学习
从合坐走向合学:浅议新学习模式的构建
面向大数据远程开放实验平台构建研究
基于自动智能分类器的图书馆乱架图书检测
搭建深度学习的三级阶梯
有体验的学习才是有意义的学习
电子商务中基于深度学习的虚假交易识别研究
利用网络技术促进学生深度学习的几大策略
MOOC与翻转课堂融合的深度学习场域建构
大数据技术在反恐怖主义中的应用展望
深度学习算法应用于岩石图像处理的可行性研究