• Laser & Optoelectronics Progress
  • Vol. 56, Issue 7, 072801 (2019)
Wenxiu Teng1、2, Xiaorong Wen1、2、*, Ni Wang3、4, and Huihui Shi3
Author Affiliations
  • 1 Co-Innovation Center for the Sustainable Forestry in Southern China, Nanjing Forestry University, Nanjing, Jiangsu 210037, China
  • 2 College of Forest, Nanjing Forestry University, Nanjing, Jiangsu 210037, China
  • 3 School of Geographic Information and Tourism, Chuzhou University, Chuzhou, Anhui 239000, China
  • 4 Anhui Engineering Laboratory of Geographical Information Intelligent Sensor and Service, Chuzhou, Anhui 239000, China
  • show less
    DOI: 10.3788/LOP56.072801 Cite this Article Set citation alerts
    Wenxiu Teng, Xiaorong Wen, Ni Wang, Huihui Shi. Tree Species Classification and Mapping Based on Deep Transfer Learning with Unmanned Aerial Vehicle High Resolution Images[J]. Laser & Optoelectronics Progress, 2019, 56(7): 072801 Copy Citation Text show less
    Model structure of convolutional neural network
    Fig. 1. Model structure of convolutional neural network
    Comparison between traditional machine learning and transfer learning. (a) Traditional machine learning; (b) transfer learning
    Fig. 2. Comparison between traditional machine learning and transfer learning. (a) Traditional machine learning; (b) transfer learning
    Flowchart of tree species classification and mapping of UAV high resolution image based on deep migration learning
    Fig. 3. Flowchart of tree species classification and mapping of UAV high resolution image based on deep migration learning
    Tree species classification model
    Fig. 4. Tree species classification model
    Flowchart of tree species mapping
    Fig. 5. Flowchart of tree species mapping
    Tree species samples. (a) Metasequoia; (b) populus; (c) bamboo; (d) ginkgo
    Fig. 6. Tree species samples. (a) Metasequoia; (b) populus; (c) bamboo; (d) ginkgo
    Training accuracy and loss curves of different models. (a) Training accuracy curves; (b) training loss curves
    Fig. 7. Training accuracy and loss curves of different models. (a) Training accuracy curves; (b) training loss curves
    Confusion matrices of classification with different models. (a) VGG16; (b) ResNet50; (c) Inecption-v3; (d) Inception-ResNet-v2
    Fig. 8. Confusion matrices of classification with different models. (a) VGG16; (b) ResNet50; (c) Inecption-v3; (d) Inception-ResNet-v2
    Training accuracy and loss curves of different methods. (a) Training accuracy curves; (b) training loss curves
    Fig. 9. Training accuracy and loss curves of different methods. (a) Training accuracy curves; (b) training loss curves
    Confusion matrices of classification with different methods. (a) Small CNN; (b) proposed method
    Fig. 10. Confusion matrices of classification with different methods. (a) Small CNN; (b) proposed method
    Tree species maps with different methods. (a) Original image; (b) actual map; (c) uniform decomposition method; (d) proposed method
    Fig. 11. Tree species maps with different methods. (a) Original image; (b) actual map; (c) uniform decomposition method; (d) proposed method
    Wenxiu Teng, Xiaorong Wen, Ni Wang, Huihui Shi. Tree Species Classification and Mapping Based on Deep Transfer Learning with Unmanned Aerial Vehicle High Resolution Images[J]. Laser & Optoelectronics Progress, 2019, 56(7): 072801
    Download Citation