Forest Change Detection based on Siamese Neural Network with GF-2 Image: A Case of Jiande Forest Farm, Zhejiang

Table 1. GF-2 load parameters
View table
View in Article
Table 1. GF-2 load parameters
载荷谱段号谱段范围/μm 空间分辨率/m
全色多光谱相机 1 0.45~0.90 1
2 0.45~0.52 4
3 0.52~0.59
4 0.63~0.69
5 0.77~0.89

Table 2. ResNet50 network structure

View table

View in Article

Table 2. ResNet50 network structure

分层名称	输出尺寸	50-layer
conv1	112×112	7×7,64,stride 2
conv2-v	56×56	3×3 max pool,stride 2
conv2-v	56×56	$[\begin{matrix} 1 \times 1,64 \\ 3 \times 3,64 \\ 1 \times 1,256 \end{matrix}] \times 3$
conv3-x	28×28	$[\begin{matrix} 1 \times 1,128 \\ 3 \times 3,128 \\ 1 \times 1,512 \end{matrix}] \times 4$
conv4-x	14×14	$[\begin{matrix} 1 \times 1,256 \\ 3 \times 3,256 \\ 1 \times 1,1024 \end{matrix}] \times 6$
conv5-x	7×7	$[\begin{matrix} 1 \times 1,512 \\ 3 \times 3,512 \\ 1 \times 1,2048 \end{matrix}] \times 3$
	1×1	average pool 1 000-d fc,softmax
FLOPs		3.8×10⁹

Table 3. Experimental software and hardware configuration
View table
View in Article
Table 3. Experimental software and hardware configuration
名称试验软硬件配置
操作系统 Windows10 (64 bit)
编程语言 Python 3.7
深度学习框架 PyTorch 1.10.2
GPU NVIDIA GeForce RTX3070
CUDA Cuda 11.3
环境管理 PyCharm 2021.3

Table 4. Forest land change detection confusion matrix
View table
View in Article
Table 4. Forest land change detection confusion matrix
预测值
真实值未变化-0 林地变化非林地-1 非林地变化林地-2
未变化-0 N₀₀ N₀₁ N₀₂
林地变化非林地-1 N₁₀ N₁₁ N₁₂
非林地变化林地-2 N₂₀ N₂₁ N₂₂

Table 5. Comparison of change detection indicators of different methods

View table

View in Article

Table 5. Comparison of change detection indicators of different methods

编号	方法	主干提取网络	变化类别	精确率	召回率	F1分数
1	Ours	ResNet50	Class0	96.48	94.59	95.53
			Class1	76.25	75.52	75.88
			Class2	54.06	78.49	64.02
			Macro avg	75.60	82.87	78.48
2	Ours	ResNet50+CBAM	Class0	96.60	95.71	96.15
			Class1	75.41	81.63	78.40
			Class2	66.23	67.33	66.78
			Macro avg	79.41	81.59	80.44
3	Ours	ResNet50+SE	Class0	95.81	98.57	97.17
			Class1	87.12	77.12	81.81
			Class2	90.86	56.80	69.90
			Macro avg	91.26	77.50	82.96
4	FC-Siam-conc	—	Class0	96.66	89.50	92.94
			Class1	75.97	87.61	81.37
			Class2	21.71	49.50	30.18
			Macro avg	64.78	75.53	68.16
5	FC-Siam-diff	—	Class0	95.90	95.20	95.55
			Class1	83.84	83.64	83.74
			Class2	37.77	44.10	40.67
			Macro avg	72.50	74.30	73.32

Tools

Get Citation

Copy Citation Text

Qiuyi AI, Huaguo HUANG, Ying GUO, Bingjie LIU, Shuxin CHEN, Xin TIAN. Forest Change Detection based on Siamese Neural Network with GF-2 Image: A Case of Jiande Forest Farm, Zhejiang[J]. Remote Sensing Technology and Application, 2024, 39(1): 24

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites