Structure-Aware Multiscale Hybrid Network for Change Detection of Remote Sensing Images

Qi Liu; Lin Cao; Shu Tian; Kangning Du; Peiran Song; Yanan Guo

doi:10.3788/LOP240514

Laser & Optoelectronics Progress, Volume. 61, Issue 14, 1428008(2024)

Structure-Aware Multiscale Hybrid Network for Change Detection of Remote Sensing Images

Qi Liu^1,2, Lin Cao^2,3, Shu Tian^3、*, Kangning Du³, Peiran Song³, and Yanan Guo³

¹School of Instrument Science and Opto-Electronics Engineering, Beijing Information Science & Technology University, Beijing 100101, China

²Key Laboratory of Optoelectronic Measurement Technology and Instrument, Ministry of Education, Beijing Information Science & Technology University, Beijing 100101, China

³Key Laboratory of Information and Communication Systems, Ministry of Information Industry, Beijing Information Science & Technology University, Beijing 100101, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(12)

Fig. 1. Overall structure of the proposed change detection model

Download full size

Fig. 2. Structure of global feature extraction module

Download full size

Fig. 3. Structure of local feature extraction module

Download full size

Fig. 4. Structure of feature equalization module

Download full size

Fig. 5. Visualization results on CDD dataset

Download full size

Fig. 6. Visualization results on GZ-CD dataset

Download full size

Fig. 7. Visualization images of CDD dataset ablation experiments

Download full size

Fig. 8. Visualization images of GZ-CD dataset ablation experiments

Download full size

Fig. 9. Comparison of training convergence on different datasets. (a) CDD dataset; (b) GZ-CD dataset

Download full size

Table 1. Compared experimental training parameter setting

View table

Table 1. Compared experimental training parameter setting

Parameter	FC-Siam-diff	FC-Siam-conc	FC-EF	EF-UNet++	Siam-NestedUNet	BIT	ICIFNet
CPU	Intel（R） Xeon（R） CPU E5-2640 v4 @ 2.40 GHz （×40）
GPU	NVIDIA GeForce RTX 2080 Ti （×6）
Memory	12G （×6）
Framework	PyTorch 1.6	PyTorch 1.6	PyTorch 1.6	TensorFlow	PyTorch 1.4	PyTorch 1.6	PyTorch 1.9
l_r	0.001	0.001	0.001	0.0001	0.001	0.0005	0.0005
Epoch	100	100	100	100	100	100	100
Batchsize	16	16	16	32	16	8	8

Table 2. Comparison of experimental quantitative results

View table

Table 2. Comparison of experimental quantitative results

Model	CDD				GZ-CD
Model	Precision	Recall	F1	OA	Precision	Recall	F1	OA
FC-Siam-diff	70.9	60.3	63.7	93.3	81.0	56.7	66.7	87.4
FC-Siam-conc	66.0	60.7	62.8	91.9	80.6	63.8	71.3	88.8
FC-EF	60.9	58.3	59.2	91.1	85.2	73.5	78.9	91.5
EF-UNet++	91.1	88.3	89.6	97.8	85.0	81.0	83.0	91.7
Siam-NestedUNet	92.4	90.8	91.5	98.1	86.1	83.3	84.4	96.8
BIT	97.0	95.2	96.3	98.4	86.5	75.2	80.8	96.7
ICIFNet	96.9	96.1	95.2	97.8	90.2	81.5	85.6	97.1
MSCT	97.1	98.6	97.8	99.1	92.6	82.1	87.1	97.5

Table 3. Results of ablation experiments

View table

Table 3. Results of ablation experiments

Method	GFE	LFE	FEM	CDD				GZ-CD
Method	GFE	LFE	FEM	Precision	Recall	F1	OA	Precision	Recall	F1	OA
MSCT-1	√	×	√	97.2	94.5	95.8	96.4	88.5	82.7	85.5	96.8
MSCT-2	×	√	√	96.5	96.8	96.5	97.0	90.4	81.6	85.8	97.3
MSCT-3	√	√	×	98.3	96.1	97.2	97.9	90.2	82.2	86.1	97.1
MSCT	√	√	√	97.1	98.6	97.8	99.1	92.6	82.1	87.1	97.5

Tools

Get Citation

Copy Citation Text

Qi Liu, Lin Cao, Shu Tian, Kangning Du, Peiran Song, Yanan Guo. Structure-Aware Multiscale Hybrid Network for Change Detection of Remote Sensing Images[J]. Laser & Optoelectronics Progress, 2024, 61(14): 1428008

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Remote Sensing and Sensors

Received: Jan. 12, 2024

Accepted: Mar. 7, 2024

Published Online: Jul. 8, 2024

The Author Email: Shu Tian (shutian@bistu.edu.cn)

DOI:10.3788/LOP240514

CSTR:32186.14.LOP240514

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology