Deep convolution neural network-based transfer learning method for civil infrastructure crack detection

Abstract

Crack detection is critical to guaranteeing safety of bridges, highway and other infrastructures. The deep convolution neural network (DCNN) makes it possible to efficiently and accurately implement image classification, and the accumulated knowledge of DCNN in other domains can be reused for crack detection. In this paper, we propose a transfer learning method based on DCNN to detect cracks. The proposed method models the knowledge learned by DCNN and transfers three kinds of knowledge from other research achievements: sample knowledge, model knowledge and parameter knowledge. New fully connected layers have emerged in the Visual Geometry Group (VGG) network as a new learning framework for crack detection. The performance and validity of the proposed method are verified. Compared with other detection methods, the proposed method can detect many kinds of cracks with a high detection accuracy. The detection accuracy for CCIC [24] is 99.83%, that for BCD [25] is 99.72%, and that for SDNET [45] is 97.07%. The accumulated knowledge in this method can also be transferred to other research work.