Multi-sensor optical remote sensing images change detection based on global differential enhancement module and balance penalty loss

doi:10.11947/j.AGCS.2025.20240300

Abstract

Abstract:

High-resolution optical imagery, characterized by its rich spatial details and high interpretability, is one of the primary data sources for remote sensing change detection tasks. In terms of practical applications, single-source optical imagery is often constrained by revisit periods and the availability of archived data, which may not fully meet the requirements. In contrast, the integration of multi-source optical remote sensing imagery offers greater flexibility and applicability. Nonetheless, the substantial spatiotemporal heterogeneity among optical images acquired by different sensors leads to significant “pseudo-invariance” and “pseudo-change” phenomenon, posing severe challenges to accurately extracting genuine change information. In order to address these challenges, this paper aims to develop a model with strong discriminative capability for addressing both types of false detection. In this paper, a change detection model for multi-sensor optical remote sensing images is proposed, based on global differential enhancement module (GDEM) and balance penalty loss (BP Loss), named GB-UNet++. The GDEM introduces Transformer to facilitate the interaction of global information across multi-temporal images, thereby enhancing the model's ability to capture cross-pixel/region change information. Additionally, the proposed BP Loss adaptively adjusts the weights to enhance the model's capacity to learn from both types of misclassified samples. Extensive experiments on six datasets demonstrate that the proposed method achieves an overall accuracy (OA) of 99.02% and an F₁ score of 84.86%, significantly outperforming five state-of-the-art methods.

Key words: multi-sensor optical remote sensing images, change detection, Transformer, deep learning

CLC Number:

P237

Chao WANG, Tianyu CHEN, Tong ZHANG, Tanvir AHMED, Liqiang JI, Tao XIE, Jiajun YANG, Shuai WANG. Multi-sensor optical remote sensing images change detection based on global differential enhancement module and balance penalty loss[J]. Acta Geodaetica et Cartographica Sinica, 2025, 54(5): 873-887.

Figures/Tables 13

Fig. 1

Fig. 2

Fig. 3

Tab. 1

Fig. 4

Fig. 5

Tab. 2

Tab. 3

Fig. 6

Fig. 7

Fig. 8

Fig. 9

Fig. 10

References 40

[1]	黄亮. 多时相遥感影像变化检测技术研究[J]. 测绘学报, 2020, 49(6): 801. DOI: . doi: 10.11947/j.AGCS.2020.20190236
	HUANG Liang. Research on change detection technology in multi-temporal remote sensing images[J]. Acta Geodaetica et Cartographica Sinica, 2020, 49(6): 801. DOI: . doi: 10.11947/j.AGCS.2020.20190236
[2]	宫金杞. 复杂城市场景的高分辨率遥感影像建筑物变化检测[J]. 测绘学报, 2023, 52(7): 1233. DOI: . doi: 10.11947/j.AGCS.2023.20210728
	GONG Jinqi. Building change detection in high-resolution remote sensing images of complex urban scenes[J]. Acta Geodaetica et Cartographica Sinica, 2023, 52(7): 1233. DOI: . doi: 10.11947/j.AGCS.2023.20210728
[3]	姜明, 张新长, 孙颖, 等. 全尺度特征聚合的高分辨率遥感影像变化检测网络[J]. 测绘学报, 2023, 52(10): 1738-1748. DOI: . doi: 10.11947/j.AGCS.2023.20220505
	JIANG Ming, ZHANG Xinchang, SUN Ying, et al. Full-scale feature aggregation network for high-resolution remote sensing image change detection[J]. Acta Geodaetica et Cartographica Sinica, 2023, 52(10): 1738-1748. DOI: . doi: 10.11947/j.AGCS.2023.20220505
[4]	WU Haoran, GENG Jie, JIANG Wen. Multidomain constrained translation network for change detection in heterogeneous remote sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2024, 62: 3381196.
[5]	许淑淑. 基于对象的多源数据变化检测的方法[J]. 测绘通报, 2020(): 122-126.
	XU Shushu. Research on object based multisource data change detection method[J]. Bulletin of Surveying and Mapping, 2020 (): 122-126.
[6]	CHEN Dong, WANG Yafei, SHEN Zhenyu, et al. Long time-series mapping and change detection of coastal zone land use based on Google Earth Engine and multi-source data fusion[J]. Remote Sensing, 2022, 14(1): 1.
[7]	施向丰, 帅梅琴, 申劲松. 基于多时相遥感图像智能变化检测方法的研究[J]. 测绘通报, 2012(9): 23-25.
	SHI Xiangfeng, SHUAI Meiqin, SHEN Jinsong. Study on the intelligent change detection methods on the basis of multi-temporal remotely sensed images[J]. Bulletin of Surveying and Mapping, 2012(9): 23-25.
[8]	韩星, 韩玲, 李良志, 等. 基于深度学习的高分辨率遥感图像建筑物变化检测[J]. 激光与光电子学进展, 2022, 59(10): 1001003.
	HAN Xing, HAN Ling, LI Liangzhi, et al. Building change detection in high-resolution remote-sensing images based on deep learning[J]. Laser & Optoelectronics Progress, 2022, 59(10): 1001003.
[9]	ZHANG Hongyan, LIN Manhui, YANG Guanyi, et al. ESCNet: an end-to-end superpixel-enhanced change detection network for very-high-resolution remote sensing images[J]. IEEE Transactions on Neural Networks and Learning Systems, 2021, 34(1): 28-42.
[10]	DONG Sijun, WANG Libo, DU Bo, et al. ChangeCLIP: remote sensing change detection with multimodal vision-language representation learning[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2024, 208: 53-69.
[11]	ZHENG Zhi, WAN Yi, ZHANG Yongjun, et al. CLNet: cross-layer convolutional neural network for change detection in optical remote sensing imagery[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2021, 175: 247-267.
[12]	孙雯婷, 施文灶, 王磊, 等. 改进UNet++遥感影像建筑物变化检测[J]. 电脑知识与技术, 2022, 18(25): 20-25.
	SUN Wenting, SHI Wenzao, WANG Lei, et al. Improved UNet++ for building change detection in remote sensing images[J]. Computer Knowledge and Technology, 2022, 18(25): 20-25.
[13]	SU Linzhi, XIE Qiaoyun, ZHAO Fengjun, et al. Change detection for multispectral images using modified semantic segmentation network[J]. Journal of Applied Remote Sensing, 2022, 16: 014518.
[14]	ZHANG Xiuwei, YUE Yuanzeng, GAO Wenxiang, et al. DifUnet++: a satellite images change detection network based on UNet++ and differential pyramid[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19: 3049370.
[15]	FANG Sheng, LI Kaiyu, SHAO Jinyuan, et al. SNUNet-CD: a densely connected Siamese network for change detection of VHR images[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19: 3056416.
[16]	BAO Yintu, LIU Wei, GAO Ouyang, et al. E-Unet++: a semantic segmentation method for remote sensing images[C]//Proceedings of 2021 Advanced Information Management, Communicates, Electronic and Automation Control Conference. Chongqing: IEEE, 2021: 1858-1862.
[17]	LI Jian, LIU Kongyu, HU Yating, et al. Eres-UNet++: liver CT image segmentation based on high-efficiency channel attention and Res-UNet++[J]. Computers in Biology and Medicine, 2023, 158: 106501.
[18]	LI Zan, ZHANG Hong, LI Zhengzhen, et al. Residual-attention UNet++: a nested residual-attention U-Net for medical image segmentation[J]. Applied Sciences, 2022, 12(14): 7149.
[19]	袁洲, 郭海涛, 卢俊, 等. 融合UNet++网络和注意力机制的高分辨率遥感影像变化检测算法[J]. 测绘科学技术学报, 2021, 38(2): 155-159.
	YUAN Zhou, GUO Haitao, LU Jun, et al. Change detection algorithm for high-resolution remote sensing images by integrating UNet++ network and attention mechanism[J]. Journal of Geomatics Science and Technology, 2021, 38(2): 155-159.
[20]	PENG Xueli, ZHONG Ruofei, LI Zhen, et al. Optical remote sensing image change detection based on attention mechanism and image difference[J]. IEEE Transactions on Geoscience and Remote Sensing, 2021, 59(9): 7296-7307.
[21]	王超, 王帅, 陈晓, 等. 联合UNet++和多级差分模块的多源光学遥感影像对象级变化检测[J]. 测绘学报, 2023, 52(2): 283-296. DOI: . doi: 10.11947/j.AGCS.2023.20220202
	WANG Chao, WANG Shuai, CHEN Xiao, et al. Object-level change detection of multi-sensor optical remote sensing images combined with UNet++ and multi-level difference module[J]. Acta Geodaetica et Cartographica Sinica, 2023, 52(2): 283-296. DOI: . doi: 10.11947/j.AGCS.2023.20220202
[22]	SUN Bangyong, LIU Qinsen, YUAN Nianzeng, et al. Spectral token guidance transformer for multisource images change detection[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2023, 16: 2559-2572.
[23]	FU Yanping, LIU Yun. Contrastive Transformer based domain adaptation for multi-source cross-domain sentiment classification[J]. Knowledge-Based Systems, 2022, 245: 108649.
[24]	VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[J]. Advances in Neural Information Processing Systems, 2017, 30: 5998-6008.
[25]	YAN Tianyu, WAN Zifu, ZHANG Pingping. Fully Transformer network for change detection of remote sensing images[C]//Proceedings of 2022 Asian Conference on Computer Vision. Cham: Springer-Verlag, 2022: 75-92.
[26]	BANDARA W G C, PATEL V M. A Transformer-based siamese network for change detection[C]//Proceedings of 2022 IEEE International Geoscience and Remote Sensing Symposium. Kuala Lumpur: IEEE, 2022: 207-210.
[27]	JIANG Bo, WANG Zitian, WANG Xixi, et al. VcT: visual change Transformer for remote sensing image change detection[J]. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61: 1-14.
[28]	YAN Tianyu, WAN Zifu, ZHANG Pingping, et al. TransY-Net: learning fully Transformer networks for change detection of remote sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61: 1-12.
[29]	CHEN Hao, QI Zipeng, SHI Zhenwei. Remote sensing image change detection with Transformers[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: 1-14.
[30]	MAO Zan, TONG Xinyu, LUO Ze, et al. MFATNet: multi-scale feature aggregation via transformer for remote sensing image change detection[J]. Remote Sensing, 2022, 14(21): 5379.
[31]	JAFFARI R, HASHMANI M A, REYES-ALDASORO C C. A novel focal phi loss for power line segmentation with auxiliary classifier U-Net[J]. Sensors, 2021, 21(8): 2803.
[32]	JI Shunping, WEI Shiqing, LU Meng. Fully convolutional networks for multisource building extraction from an open aerial and satellite imagery data set[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(1): 574-586.
[33]	ZHANG Chenxiao, YUE Peng, TAPETE D, et al. A deeply supervised image fusion network for change detection in high resolution bi-temporal remote sensing images[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2020, 166: 183-200.
[34]	PENG Daifeng, BRUZZONE L, ZHANG Yongjun, et al. SemiCDNet: a semisupervised convolutional neural network for change detection in high resolution remote-sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2021, 59(7): 5891-5906.
[35]	PENG Daifeng, ZHANG Yongjun, GUAN Haiyan. End-to-end change detection for high resolution satellite images using improved UNet++[J]. Remote Sensing, 2019, 11(11): 1382-1405.
[36]	MILLETARI F, NAVAB N, AHMADI S A. V-net: fully convolutional neural networks for volumetric medical image segmentation[C]//Proceedings of 2016 International Conference on 3D Vision. Stanford: IEEE, 2016: 565-571.
[37]	CHEN Jieneng, LU Yongyi, YU Qihang, et al. TransUNet: Transformers make strong encoders for medical image segmentation[EB/OL]. [2023-12-10]. https://arxiv.org/abs/2102.04306v1.
[38]	LI Qingyang, ZHONG Ruofei, DU Xin, et al. TransUNetCD: a hybrid Transformer network for change detection in optical remote-sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: 3169479.
[39]	TIAN Juan, PENG Daifeng, GUAN Haiyan, et al. RACDNet: resolution- and alignment-aware change detection network for optical remote sensing imagery[J]. Remote Sensing, 2022, 14(18): 4527.
[40]	RAZA A, HUO Hong, FANG Tao. EUNet-CD: efficient UNet++ for change detection of very high-resolution remote sensing images[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19: 3144304.

数据集	数据来源	空间分辨率/m	样本尺寸/像素	时间跨度
BCDD	航拍图像	0.2	256×256	2012—2016年
DSIFN	在线地图服务平台	2	256×256	2001—2019年
GZCD	在线地图服务平台	0.55	256×256	2010—2019年
YCCD	GF-1、GF-2	2	128×128	2018—2020年
NJCD	World View-Ⅱ、航拍图像	2	128×128	2011—2013年
GYCD	GF-1、GF-2	2	128×128	2015—2020年

数据集	方法	P/（%）	R/（%）	PRD/（%）	F₁值/（%）	OA/（%）	Kappa系数
BCDD	Vc T＿CD	79.55	86.42	6.87	82.46	98.86	0.817 2
	BIT＿CD	76.55	85.39	8.84	80.74	98.73	0.806 1
	ChangeFormer	64.97	77.29	12.32	70.60	98.23	0.696 9
	CLNet	74.95	80.61	5.66	77.67	98.59	0.769 4
	MFED-UNet++	81.01	84.63	3.62	82.77	98.89	0.822 1
	GB-UNet++	83.24	86.54	3.30	84.86	99.02	0.843 5
	Vc T＿CD	68.15	67.54	0.61	67.84	88.18	0.638 8
	BIT＿CD	64.05	69.69	5.64	66.75	87.88	0.608 5
DSIFN	ChangeFormer	63.27	65.18	1.91	64.21	86.89	0.609 3
DSIFN	CLNet	75.13	55.62	19.51	63.92	80.05	0.547 9
	MFED-UNet++	55.39	78.28	22.89	64.88	88.15	0.612 2
	GB-UNet++	68.83	69.15	0.32	68.99	88.24	0.666 2
	Vc T＿CD	70.10	74.53	4.43	72.24	93.94	0.688 5
	BIT＿CD	47.58	84.02	36.44	60.75	93.08	0.572 8
GZCD	ChangeFormer	46.51	87.71	41.20	60.78	93.25	0.574 7
GZCD	CLNet	42.70	95.60	52.90	59.03	93.33	0.559 8
	MFED-UNet++	73.10	77.47	4.37	75.22	94.10	0.721 9
	GB-UNet++	75.41	86.12	10.71	80.41	95.87	0.781 2
	Vc T＿CD	52.95	63.63	10.68	57.80	95.60	0.554 6
	BIT＿CD	51.22	63.81	12.59	56.82	95.59	0.555 5
YCCD	ChangeFormer	53.07	52.80	0.27	52.94	94.54	0.500 4
YCCD	CLNet	49.81	63.76	13.95	55.93	95.46	0.535 7
	MFED-UNet++	55.51	65.41	9.90	60.06	95.73	0.578 2
	GB-UNet++	57.31	65.75	8.44	61.24	95.77	0.585 4
	Vc T＿CD	64.48	79.59	15.11	71.21	95.92	0.673 3
	BIT＿CD	65.40	79.87	14.47	71.92	96.37	0.700 1
NJCD	ChangeFormer	62.21	77.86	15.65	69.16	96.06	0.670 9
NJCD	CLNet	64.74	78.53	13.79	70.97	96.24	0.689 8
	MFED-UNet++	66.69	81.33	14.64	73.28	96.55	0.714 6
	GB-UNet++	67.18	81.88	14.70	73.81	96.59	0.720 2
	Vc T＿CD	65.47	72.39	6.92	68.79	91.74	0.658 6
	BIT＿CD	54.18	66.66	12.48	59.78	89.85	0.540 5
GYCD	ChangeFormer	57.51	65.98	8.47	61.46	89.96	0.557 2
GYCD	CLNet	53.95	65.93	11.98	59.34	89.71	0.535 2
	MFED-UNet++	68.66	71.60	2.94	70.09	91.85	0.653 8
	GB-UNet++	68.85	72.53	3.68	70.62	92.04	0.661 1

数据集	基础网络	GDEM	P/（%）	R/（%）	PRD/（%）	F₁值/（%）	OA/（%）	Kappa系数
BCDD	UNet++	—	65.91	63.89	2.02	64.89	97.66	0.636 1
BCDD	UNet++	√	83.24	86.54	3.30	84.86	99.02	0.843 5
DSIFN	UNet++	—	70.57	56.71	13.86	62.89	85.47	0.539 9
DSIFN	UNet++	√	68.83	69.15	0.32	68.99	88.24	0.666 2
GZCD	UNet++	—	56.06	71.56	15.50	62.87	92.55	0.588 1
GZCD	UNet++	√	75.41	86.12	10.71	80.41	95.87	0.781 2
YCCD	UNet++	—	50.77	51.93	1.16	51.35	94.43	0.483 9
YCCD	UNet++	√	57.31	65.75	8.44	61.24	95.77	0.585 4
NJCD	UNet++	—	62.67	80.48	17.81	70.47	96.27	0.685 1
NJCD	UNet++	√	67.18	81.88	14.70	73.81	96.59	0.720 2
GYCD	UNet++	—	57.89	66.99	9.10	62.11	90.16	0.564 9
GYCD	UNet++	√	68.85	72.53	3.68	70.62	92.04	0.661 1