Laser & Optoelectronics Progress, Volume. 57, Issue 24, 241504(2020)
Distortion Correction of Single Image Based on Deep Learning
Fig. 1. Relationship between the actual and ideal image points
Fig. 2. Functional relationship between k1 and k2
Fig. 3. Part of the image in the synthetic dataset
Fig. 4. Structure of AlexNet
Fig. 5. Accuracy of the two algorithms on the test set
Fig. 6. Loss rate of the two algorithms on the test set
Fig. 7. Simulation image. (a) Standard image; (b) distorted image
Fig. 8. RMSE of corner coordinates obtained from 50 random experiments
Fig. 9. Experimental results of real checkerboard images. (a) Distorted image; (b) image corrected by Ref.[4]; (c) image corrected by our method
Fig. 10. RMSE of real chessboard image
Fig. 11. Correction effect of wide-angle lens. (a) Distorted image; (b) image corrected by our method
Get Citation
Copy Citation Text
Wenyi Chen, Jie Xu, Hui Yang, Xiaobao Yang, Xiaoqiang Xi. Distortion Correction of Single Image Based on Deep Learning[J]. Laser & Optoelectronics Progress, 2020, 57(24): 241504
Category: Machine Vision
Received: Mar. 16, 2020
Accepted: Jun. 11, 2020
Published Online: Nov. 18, 2020
The Author Email: Xu Jie (1141849828@qq.com)