基于多尺度融合的遥感图像变化检测模型

doi:10.13229/j.cnki.jdxbgxb.20221340

摘要/Abstract

摘要：

为精准识别双时相遥感图像的变化区域，提出了一种基于多尺度融合的遥感图像变化检测模型。该模型在源图像特征提取阶段构造多尺度输入金字塔，接受多层次的感受野，增强对特征信息的感知；并通过对深层差异特征进行多尺度计算，实现精准定位变化区域与充分挖掘细节信息间的平衡；同时融合网络不同层级的差异特征检测结果，极大程度识别并保留语义变化信息。实验结果表明：本文模型在主观评价与客观指标上都具有良好的表现效果。

关键词: 变化检测, 多尺度融合, 遥感图像, 深度学习

Abstract:

The remote sensing image change detection model which is based on multi-scale fusion is proposed to accurately identify the change region of the bi-temporal remote sensing images. First， a multi-scale input pyramid is constructed in the feature extraction stage to receive multi-layer receptive fields and enhance the perception of all information. Then， in order to make a tradeoff between locating the changing area and mining details， the multi-scale calculation is carried out for deep difference features. Finally， the semantic change information can be identified and retained to a great extent by integrating the different feature results of the network. The experimental results show that the proposed model has good performance in both subjective evaluation and objective indexes.

Key words: change detection, multi-scale fusion, remote sensing image, deep learning

中图分类号:

TP391

李雄飞,宋紫萱,朱芮,张小利. 基于多尺度融合的遥感图像变化检测模型[J]. 吉林大学学报(工学版), 2024, 54(2): 516-523.

Xiong-fei LI,Zi-xuan SONG,Rui ZHU,Xiao-li ZHANG. Remote sensing change detection model based on multi⁃scale fusion[J]. Journal of Jilin University(Engineering and Technology Edition), 2024, 54(2): 516-523.

图/表 5

图1

图2

图3

表1

表2

参考文献 22

1	Luo H, Liu C, Wu C, et al. Urban change detection based on Dempster-Shafer theory for multitemporal very high-resolution imagery[J]. Remote Sensing, 2018, 10(7): 10070980.
2	Lv Z, Liu T, Wan Y, et al. Post-processing approach for refining raw land cover change detection of very high-resolution remote sensing images[J]. Remote Sensing, 2018, 10(3): 10030472.
3	Willis K S. Remote sensing change detection for ecological monitoring in United States protected areas[J]. Biological Conservation, 2015, 182: 233-242.
4	Hussain M, Chen D, Cheng A, et al. Change detection from remotely sensed images: from pixel-based to object-based approaches[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2013, 80: 91-106.
5	Deng J S, Wang K, Deng Y H, et al. PCA‐based land‐use change detection and analysis using multitemporal and multisensor satellite data[J]. International Journal of Remote Sensing, 2008, 29(16): 4823-4838.
6	Nielsen A A, Conradsen K, Simpson J J. Multivariate alteration detection (MAD) and MAF postprocessing in multispectral, bitemporal image data: new approaches to change detection studies[J]. Remote Sensing of Environment, 1998, 64(1): 1-19.
7	Malila W A. Change vector analysis: an approach for detecting forest changes with Landsat[C]∥6th Proceedings of the Society of Photo-Optical Instrumentation Engineers, Lafeyette, USA, 1980: 326-336.
8	慕彩红, 霍利利, 刘逸, 等. 基于小波融合和PCA-核模糊聚类的遥感图像变化检测[J]. 电子学报, 2015, 43(7): 1375-1381.
	Mu Cai-hong, Huo Li-li, Liu Yi, et al. Change detection for remote sensing images based on wavelet fusion and PCA-Kernel fuzzy clustering[J]. Acta Electronica Sinica, 2015, 43(7): 1375-1381.
9	Canty M J, Nielsen A A. Automatic radiometric normalization of multitemporal satellite imagery with the iteratively re-weighted MAD transformation[J]. Remote Sensing of Environment, 2008, 112(3): 1025-1036.
10	Shi W, Zhang M, Zhang R, et al. Change detection based on artificial intelligence: State-of-the-art and challenges[J]. Remote Sensing, 2020, 12(10): 12101688.
11	Daudt R C, Le Saux B, Boulch A. Fully convolutional siamese networks for change detection[C]∥ 25th IEEE International Conference on Image Processing, Athens, Greece, 2018: 4063-4067.
12	Fang S, Li K, Shao J, et al. SNUNet-CD: a densely connected siamese network for change detection of VHR images[J]. IEEE Geoscience and Remote Sensing Letters, 2021, 19: 1-5.
13	Peng D, Zhang Y, Guan H. End-to-end change detection for high resolution satellite images using improved UNet++[J]. Remote Sensing, 2019, 11(11): 11111382.
14	Chen H, Shi Z. A spatial-temporal attention-based method and a new dataset for remote sensing image change detection[J]. Remote Sensing, 2020, 12(10): 12101662.
15	Peng X, Zhong R, Li Z, et al. Optical remote sensing image change detection based on attention mechanism and image difference[J]. IEEE Transactions on Geoscience and Remote Sensing, 2020, 59(9): 7296-7307.
16	赵亚慧, 杨飞扬, 张振国, 等. 基于强化学习和注意力机制的朝鲜语文本结构发现[J]. 吉林大学学报: 工学版, 2021, 51(4): 1387-1395.
	Zhao Ya-hui, Yang Fei-yang, Zhang Zhen-guo, et al. Korean text structure discovery based on reinforcement learning and attention mechanism[J]. Journal of Jilin University (Engineering and Technology Edition), 2021, 51(4): 1387-1395.
17	周大可, 张超, 杨欣. 基于多尺度特征融合及双重注意力机制的自监督三维人脸重建[J]. 吉林大学学报: 工学版, 2022, 52(10): 2428-2437.
	Zhou Da-ke, Zhang Chao, Yang Xin. Self-supervised 3D face reconstruction based on multi-scale feature fusion and dual attention mechanism[J]. Journal of Jilin University (Engineering and Technology Edition), 2022, 52(10): 2428-2437.
18	Jiang H, Hu X, Li K, et al. PGA-SiamNet: pyramid feature-based attention-guided Siamese network for remote sensing orthoimagery building change detection[J]. Remote Sensing, 2020, 12(3): 12030484.
19	Song K, Jiang J. AGCDetNet: an attention-guided network for building change detection in high-resolution remote sensing images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2021, 14: 4816-4831.
20	Ding Q, Shao Z, Huang X, et al. DSA-Net: a novel deeply supervised attention-guided network for building change detection in high-resolution remote sensing images[J]. International Journal of Applied Earth Observation and Geoinformation, 2021, 105: 102591.
21	Woo S, Park J, Lee J Y, et al. Cbam: convolutional block attention module[C]∥Proceedings of the European Conference on Computer Vision, Munich, Germany, 2018: 3-19.
22	Ji S, Wei S, Lu M. Fully convolutional networks for multisource building extraction from an open aerial and satellite imagery data set[J]. IEEE Transactions on Geoscience and Remote Sensing, 2018, 57(1): 574-586.

相关文章 15

[1]	杨国俊,齐亚辉,石秀名. 基于数字图像技术的桥梁裂缝检测综述[J]. 吉林大学学报(工学版), 2024, 54(2): 313-332.
[2]	王春华,李恩泽,肖敏. 多特征融合和孪生注意力网络的高分辨率遥感图像目标检测[J]. 吉林大学学报(工学版), 2024, 54(1): 240-250.
[3]	陈岳林,高铸成,蔡晓东. 基于BERT与密集复合网络的长文本语义匹配模型[J]. 吉林大学学报(工学版), 2024, 54(1): 232-239.
[4]	霍光,林大为,刘元宁,朱晓冬,袁梦,盖迪. 基于多尺度特征和注意力机制的轻量级虹膜分割模型[J]. 吉林大学学报(工学版), 2023, 53(9): 2591-2600.
[5]	蔡志丹,方明,李喆,许佳路. 基于高斯曲率和加权图总变分正则化的遥感图像盲去模糊算法[J]. 吉林大学学报(工学版), 2023, 53(9): 2649-2658.
[6]	金小俊,孙艳霞,于佳琳,陈勇. 基于深度学习与图像处理的蔬菜苗期杂草识别方法[J]. 吉林大学学报(工学版), 2023, 53(8): 2421-2429.
[7]	耿庆田,刘植,李清亮,于繁华,李晓宁. 基于一种深度学习模型的土壤湿度预测[J]. 吉林大学学报(工学版), 2023, 53(8): 2430-2436.
[8]	巫威眺,曾坤,周伟,李鹏,靳文舟. 基于多源数据和响应面优化的公交客流预测深度学习方法[J]. 吉林大学学报(工学版), 2023, 53(7): 2001-2015.
[9]	张振海,季坤,党建武. 基于桥梁裂缝识别模型的桥梁裂缝病害识别方法[J]. 吉林大学学报(工学版), 2023, 53(5): 1418-1426.
[10]	吴飞,农皓业,马晨浩. 基于粒子群优化算法⁃长短时记忆模型的刀具磨损预测方法[J]. 吉林大学学报(工学版), 2023, 53(4): 989-997.
[11]	何科,丁海涛,赖宣淇,许男,郭孔辉. 基于Transformer的轮式里程计误差预测模型[J]. 吉林大学学报(工学版), 2023, 53(3): 653-662.
[12]	刘春晖,王思长,郑策,陈秀连,郝春蕾. 基于深度学习的室内导航机器人避障规划算法[J]. 吉林大学学报(工学版), 2023, 53(12): 3558-3564.
[13]	吕晓琪,李浩,谷宇. 基于深度学习算法的人脸图像活体特征变换尺度提取[J]. 吉林大学学报(工学版), 2023, 53(11): 3201-3206.
[14]	王军,王华琳,黄博文,付强,刘俊. 基于联邦学习和自注意力的工业物联网入侵检测[J]. 吉林大学学报(工学版), 2023, 53(11): 3229-3237.
[15]	成丽波,李新月,李喆,贾小宁. 基于曲波变换与拟合优度检验的遥感图像去噪方法[J]. 吉林大学学报(工学版), 2023, 53(11): 3207-3213.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

模型	LEVIR-CD					WHU
模型	Pre	Rec	F₁	OA	KC	Pre	Rec	F₁	OA	KC
FC-EF	0.8509	0.7401	0.7916	0.9802	0.7813	0.7673	0.8936	0.8256	0.9838	0.8172
FC-Siam-diff	0.8500	0.8233	0.8365	0.9836	0.8278	0.8134	0.8931	0.8514	0.9866	0.8444
FC-Siam-conc	0.8363	0.8239	0.8300	0.9828	0.8210	0.8142	0.8987	0.8544	0.9869	0.8475
SNUNet-CD	0.9049	0.8498	0.8765	0.9878	0.8701	0.8251	0.8898	0.8562	0.9872	0.8495
MSOF	0.8784	0.8680	0.8732	0.9872	0.8664	0.9471	0.8895	0.9174	0.9931	0.9138
STANet	0.8501	0.9138	0.8808	0.9874	0.8741	0.7936	0.9120	0.8487	0.9861	0.8414
本文	0.8852	0.9090	0.8969	0.9894	0.8913	0.9176	0.9202	0.9189	0.9931	0.9153

模型	Pre	Rec	F₁	OA	KC
去除输入金字塔	0.8872	0.8993	0.8932	0.9890	0.8874
去除深度差异特征多尺度计算	0.8817	0.9071	0.8942	0.9891	0.8885
去除多级差异特征融合	0.8860	0.8986	0.8923	0.9889	0.8865
本文	0.8852	0.9090	0.8969	0.9894	0.8913