4.7 Article

Urban Vegetation Extraction from High-Resolution Remote Sensing Imagery on SD-UNet and Vegetation Spectral Features

Journal

REMOTE SENSING
Volume 15, Issue 18, Pages -

Publisher

MDPI
DOI: 10.3390/rs15184488

Keywords

Gaofen-1 imagery; deep learning; dense connection; separable convolution; SD-UNet; urban vegetation extraction; NIR

Ask authors/readers for more resources

Urban vegetation plays a crucial role in the urban ecological system, and efficient extraction of vegetation information is important. Therefore, this study proposes a new model called SD-UNet, which improves the accuracy and generalization ability through the introduction of dense connections and separable convolutions.
Urban vegetation plays a crucial role in the urban ecological system. Efficient and accurate extraction of urban vegetation information has been a pressing task. Although the development of deep learning brings great advantages for vegetation extraction, there are still problems, such as ultra-fine vegetation omissions, heavy computational burden, and unstable model performance. Therefore, a Separable Dense U-Net (SD-UNet) was proposed by introducing dense connections, separable convolutions, batch normalization layers, and Tanh activation function into U-Net. Furthermore, the Fake sample set (NIR-RG), NDVI sample set (NDVI-RG), and True sample set (RGB) were established to train SD-UNet. The obtained models were validated and applied to four scenes (high-density buildings area, cloud and misty conditions area, park, and suburb) and two administrative divisions. The experimental results show that the Fake sample set can effectively improve the model's vegetation extraction accuracy. The SD-UNet achieves the highest accuracy compared to other methods (U-Net, SegNet, NDVI, RF) on the Fake sample set, whose ACC, IOU, and Recall reached 0.9581, 0.8977, and 0.9577, respectively. It can be concluded that the SD-UNet trained on the Fake sample set not only is beneficial for vegetation extraction but also has better generalization ability and transferability.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.7
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available