Remote sensing images change detection based on MCRASN

Table 1. Confusion matrix
View table
View in Article
Table 1. Confusion matrix
Predicted to be change (1) Predicted to be non-change (0)
Truth change (1) TP FN
Truth non-change (0) FP TN

Table 2. Different sizes of MCRASN model

View table

View in Article

Table 2. Different sizes of MCRASN model

Stages	Feature size	MCRASN-0	MCRASN-1	MCRASN-2
S0-Stem	1/2	N=2 D=64	N=2 D=64	N=2 D=128
Sl-MbConv	1/4	N=2 D=96	N=2 D=96	N=2 D=128
	1/8	N=3 D=192	N=4 D=192	N=6 D=256
S2-TFM_Rel	1/16	N=5 D=384	N=7 D=384	N=7 D=512
	1/32	N=2 D=768	N=2 D=768	N=2 D=1024

Table 3. Evaluation of prediction results of each model

View table

View in Article

Table 3. Evaluation of prediction results of each model

Methods	Precision	Recall	F1	IoU	OA
FC-EF	86.91	80.17	83.40	71.53	98.39
DTCDSCN	88.53	86.83	87.67	78.05	98.77
BIT	89.24	89.37	89.31	80.68	98.92
ChangeFormer	92.05	88.80	90.40	82.48	99.04
MCRASN-0	91.45	87.60	89.49	81.17	98.97
MCRASN-1	92.85	89.12	90.94	83.57	99.07
MCRASN-2	93.94	89.26	91.54	84.45	99.18

Table 4. Ablation experiment on each improved part
View table
View in Article
Table 4. Ablation experiment on each improved part
BASE MC-RA DM EFL LEVIR-CD
Precision IoU
√ - - - 92.05 82.48
√ - - 92.41 82.86
√ √ - 92.56 83.23
√ - √ 92.52 83.26
√ √ √ 92.85 83.57

Tools

Get Citation

Copy Citation Text

Guobo XIE, Wenkang LIAO, Zhiyi LIN, Jiayuan ZHANG. Remote sensing images change detection based on MCRASN[J]. Journal of Applied Optics, 2024, 45(2): 430

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites