Abstract

Crack detection on dam surfaces is an important task for safe inspection of hydropower stations. More and more object detection methods based on deep learning are being applied to crack detection. However, most of the methods can only achieve the classification and rough location of cracks. Pixel-level crack detection can provide more intuitive and accurate detection results for dam health assessment. To realize pixel-level crack detection, a method of crack detection on dam surface (CDDS) using deep convolution network is proposed. First, we use an unmanned aerial vehicle (UAV) to collect dam surface images along a predetermined trajectory. Second, raw images are cropped. Then crack regions are manually labelled on cropped images to create the crack dataset, and the architecture of CDDS network is designed. Finally, the CDDS network is trained, validated and tested using the crack dataset. To validate the performance of the CDDS network, the predicted results are compared with ResNet152-based, SegNet, UNet and fully convolutional network (FCN). In terms of crack segmentation, the recall, precision, F-measure and IoU are 80.45%, 80.31%, 79.16%, and 66.76%. The results on test dataset show that the CDDS network has better performance for crack detection of dam surfaces.

Highlights

  • With the rapid development of water conservancy projects, to meet the needs of power generation, shipping and irrigation, lots of hydropower stations are built

  • Regular crack detection plays a crucial part in the maintenance and operation of existing dams

  • Dorafshan et al [31] investigated the feasibility of using a Deep Learning Convolutional Neural Network (DLCNN) in inspection of concrete decks and buildings using small Unmanned Aerial Systems

Read more

Summary

Introduction

With the rapid development of water conservancy projects, to meet the needs of power generation, shipping and irrigation, lots of hydropower stations are built. Dorafshan et al [31] investigated the feasibility of using a Deep Learning Convolutional Neural Network (DLCNN) in inspection of concrete decks and buildings using small Unmanned Aerial Systems These methods achieve excellent accuracy of crack classification, but crack locating is highly necessary for crack detection. Kim et al [32] suggested a bridge crack detection method combining unmanned aerial vehicles and region with convolutional neural networks (R-CNN)-based transfer learning. Bang et al [40] proposed a pixel-level detection method for identifying road cracks in black-box images using a deep convolutional encoder–decoder network. Problems, we propose pixel-levelwe dam surfacea background texture, random of cracks, To overcome these aproblems, propose crack detection usingdetection a deep convolutional to extract features.

Methodology
The Architecture of CDDS Network
Convolutional
Pooling
Deconvolutional
Crack Database of the Dam Surface
Experimental
Evaluation Metrics of the Network
Results of CDDS Training
Calculation of Crack Size on the Dam Surface
Comparative
Methods
Conclusions
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call