Welcome to Francis Academic Press

International Journal of New Developments in Engineering and Society, 2019, 3(2); doi: 10.25236/IJNDES.19238.

Damage detection of quayside crane structure based on improved Faster R-CNN


Dong Han, Gang Tang

Corresponding Author:
Dong Han

Shanghai Maritime University, Shanghai, China


In order to detect multi-damage types of quayside crane simultaneously, a visual detection method based on region proposal deep network is proposed. Aiming at the difference of damage types and size of quayside bridges and the large distribution of small targets, the feature pyramid network (FPN) method is used to increase the resolution of feature mapping and to detect larger cracks and smaller corrosion of quayside bridges by using high-level semantic information after top-down model. By using mixup data enhancement method, the generalization ability of the original Faster R-CNN model structure can be reduced, the memory of error labels can be reduced, and the robustness against samples can be increased. Experiments show that the improved Faster R-CNN improves the detection of mAP by 5.14%.


Faster R-CNN; Damage detection; Characteristic pyramid network; Mixing up.

Cite This Paper

Dong Han, Gang Tang. Damage detection of quayside crane structure based on improved Faster R-CNN. International Journal of New Developments in Engineering and Society (2019) Vol.3, Issue 2: 284-301. https://doi.org/10.25236/IJNDES.19238.


[1] Cha Y J, Choi W, Büyüköztürk O(2017). Deep Learning‐Based Crack Damage Detection Using Convolutional Neural Networks. Computer‐aided Civil & Infrastructure Engineering, vol. 32, no. 5, pp.361-378.
[2] Zhang X, Wang K, Wang Y, et al. (2017). An improved method of rail health monitoring based on CNN and multiple acoustic emission events. 2017 IEEE International Instrumentation and Measurement Technology Conference (I2MTC).
[3] Modarres C, Astorga N, Droguett E L, et al. (2018). Convolutional neural networks for automated damage recognition and damage type identification. Structural Control and Health Monitoring, vol. 25, no.10, pp. e2230.
[4] Wen L, Li X, Gao L, et al. (2018). A new convolutional neural network-based data-driven fault diagnosis method. IEEE Transactions on Industrial Electronics, vol.65, no. 7, pp.5990-5998.
[5] Girshick R , Donahue J , Darrelland T , et al. (2014). Rich feature hierarchies for object detection and semantic segmentation.  2014 IEEE Conference on Computer Vision and Pattern Recognition.
[6] Girshick R(2015). Fast R-CNN. Computer Science.
[7] Ren S, He K, Girshick R, et al. (2015). Faster R-CNN: towards real-time object detection with region proposal networks. International Conference on Neural Information Processing Systems.