基于卷积神经网络的混凝土裂缝识别

CONCRETE CRACK DETECTION METHOD BASED ON CNN

针对基于图像处理和边缘检测的混凝土裂缝识别方法易受外部环境干扰,且在图像处理过程中会产生大量噪声,从而导致识别效果不佳的问题,提出一种基于卷积神经网络的裂缝识别方法。通过与当前主流卷积神经网络模型进行对比研究以及对真实的混凝土裂缝图像的识别验证,结果表明,文中所建立的CrackNet模型能够有效识别混凝土图像中的裂缝目标,具有高效性和强鲁棒性。

The crack recognition method based on image processing and edge detection is susceptible to external environment interference, and a lot of noise is generated in the image processing process, which leads to poor recognition effect. To solve this problem, a crack identification method based on convolutional neural network is proposed. By comparing with the current mainstream convolutional neural network model and verifying the real concrete crack image, the results show that the CrackNet model established in this paper can effectively identify the crack target in concrete image, which is highly efficient and robust.