Laser & Optoelectronics Progress, Volume. 59, Issue 18, 1811001(2022)
Hyperspectral Image Classification Based on Convolution Neural Network with Attention Mechanism
Fig. 1. Basic structure of convolutional neural network
Fig. 2. Residual structure
Fig. 3. SE module
Fig. 4. SE-Res block
Fig. 5. SE-ResNet structure
Fig. 6. Flase color images. (a) Indian Pines; (b) Pavia University
Fig. 7. Overall accuracy of different reduction ratio
Fig. 8. Overall accuracy of different space sizes
Fig. 9. Loss and accuracy of training and validation sets of Indian Pines with different sample ratios. (a) 1∶1∶8; (b) 2∶1∶7; (c) 3∶1∶6; (d) 4∶1∶5
Fig. 10. Classification maps for Indian Pines. (a) Ground truth; (b) SVM; (c) 2D-CNN; (d) 3D-CNN; (e) ResNet; (f) SE-ResNet
Fig. 11. Classification maps for Pavia University. (a) Ground truth; (b) SVM; (c) 2D-CNN; (d) 3D-CNN; (e) ResNet; (f) SE-ResNet
|
|
Get Citation
Copy Citation Text
Wenhao Chen, Jing He, Gang Liu. Hyperspectral Image Classification Based on Convolution Neural Network with Attention Mechanism[J]. Laser & Optoelectronics Progress, 2022, 59(18): 1811001
Category: Imaging Systems
Received: May. 31, 2021
Accepted: Jul. 20, 2021
Published Online: Aug. 22, 2022
The Author Email: He Jing (xiao00yao@163.com)