多任务学习孪生网络的遥感影像多类变化检测

doi:10.6046/zrzyyg.2022446

摘要
图/表
参考文献
相关文章
Metrics

全文: PDF(8701 KB)

HTML
输出: BibTeX | EndNote (RIS)

摘要

精确掌握土地覆盖/利用的变化及变化类型对国土空间规划、生态环境监测、灾害评估等有着重要意义,然而现有大部分变化检测研究主要关注二值变化检测。为此,该文首先提出了一种多任务学习深度孪生网络用于遥感影像的多类变化检测。首先提出面向对象的无监督变化检测方法,选择出新、旧时相影像中最有可能发生变化和最不可能发生变化的区域,并作为多任务学习深度孪生网络的样本; 其次,采用多任务学习深度孪生网络模型同时对新、旧时相的土地利用图以及新、旧时相的二值变化图这3个任务模型进行学习和预测; 最后,基于模型预测的新、旧时相土地利用图及新、旧时相的二值变化图获取最终的多类变化检测结果。采用第三次全国国土调查的影像数据和相应的土地利用图斑数据对多任务学习深度孪生网络模型进行了测试,结果表明所提出的方法适用于这种在没有变化、未变化样本而有历史专题图的变化检测场景中。

	服务

	把本文推荐给朋友
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	马惠
	刘波
	杜世宏

关键词 ：多任务学习, 孪生网络, 多类变化检测, 第三次全国国土调查

Abstract：

The accurate acquisition of land cover/use changes and their types is critical to territorial space planning, ecological environment monitoring, and disaster assessment. However, most current studies on the change detection focus on binary change detection. This study proposed a multi-class change detection method using a multi-task Siamese network of remote sensing images. First, an object-oriented unsupervised change detection method was employed to select areas that were most/least prone to change in the new and old temporal images. These areas were used as samples for the multi-task Siamese network. Subsequently, the multi-task Siamese network model was used to learn and predict the new and old temporal land-use maps and binary change maps. Finally, the final multi-class change detection results were derived from these maps. The multi-task Siamese network was tested based on the images from the Third National Land Survey and corresponding land-use maps. The results demonstrate that the method proposed in this study is applicable to the change detection cases where changed and unchanged samples lack but there are available historical thematic maps.

Key words： multi-task learning Siamese network multi-class change detection the third national land resource survey

收稿日期: 2022-11-17 出版日期: 2024-03-13

ZTFLH:

TP751

基金资助:国家重点研发计划-政府间国际创新合作“时空大数据驱动的可持续发展城市人居环境监测评估与应用示范”(2021YFE0117100)

通讯作者: 刘波(1995-),男,研究生,主要从事遥感信息智能解译方法研究。Email: liubo_rs@pku.edu.cn。

作者简介: 马惠(1980-),女,高级工程师,主要从事地籍调查、土地资源调查遥感监测技术方法研究。Email: 42879575@qq.com。

引用本文:

马惠, 刘波, 杜世宏. 多任务学习孪生网络的遥感影像多类变化检测[J]. 自然资源遥感, 2024, 36(1): 77-85.
MA Hui, LIU Bo, DU Shihong. Multi-class change detection using a multi-task Siamese network of remote sensing images. Remote Sensing for Natural Resources, 2024, 36(1): 77-85.

链接本文:

https://www.gtzyyg.com/CN/10.6046/zrzyyg.2022446 或 https://www.gtzyyg.com/CN/Y2024/V36/I1/77

Fig.1 本文所用数据

Fig.2 遥感影像变化检测流程

Fig.3 面向对象无监督变化检测流程

Fig.4 多任务学习深度孪生网络结构图

Fig.5 Xception网络结构图

变化轨迹		样本/未变化					样本/变化
变化轨迹		$L 11$ → $L 12$	…	$L 41$ → $L 42$	…	$L m 1$ → $L m 2$	$L 11$ → $L 22$	…	$L 41$ → $L 52$	…	$L m - 1 1$ → $L m 2$
检测结果/未变化	$L 11$ → $L 12$	S₁	S₃	S₃	S₃	S₃	S₅	S₅	S₅	S₅	S₅
	…	S₃	S₁	S₃	S₃	S₃	S₅	S₅	S₅	S₅	S₅
	$L 41$ → $L 42$	S₃	S₃	S₁	S₃	S₃	S₅	S₅	S₅	S₅	S₅
	…	S₃	S₃	S₃	S₁	S₃	S₅	S₅	S₅	S₅	S₅
	$L m 1$ → $L m 2$	S₃	S₃	S₃	S₃	S₁	S₅	S₅	S₅	S₅	S₅
检测结果/变化	$L 11$ → $L 22$	S₄	S₄	S₄	S₄	S₄	S₂	S₆	S₆	S₆	S₆
	…	S₄	S₄	S₄	S₄	S₄	S₆	S₂	S₆	S₆	S₆
	$L 41$ → $L 52$	S₄	S₄	S₄	S₄	S₄	S₆	S₆	S₂	S₆	S₆
	…	S₄	S₄	S₄	S₄	S₄	S₆	S₆	S₆	S₂	S₆
	$L m - 1 1$ → $L m 2$	S₄	S₄	S₄	S₄	S₄	S₆	S₆	S₆	S₆	S₂

Tab.1 土地利用变化轨迹矩阵及其分组

Tab.2 深度学习平台硬件配置

Tab.3 无监督变化检测结果

Tab.4 无监督变化检测精度评价结果

Tab.5 多类变化检测结果

[1]	史文中, 张鹏林. 光学遥感影像变化检测研究的回顾与展望[J]. 武汉大学学报(信息科学版), 2018, 43(12):1832-1837.
	Shi W Z, Zhang P L. State-of-the-art remotely sensed images-based change detection methods[J]. Geomatics and Information Science of Wuhan University, 2018, 43(12):1832-1837.
[2]	李德仁. 利用遥感影像进行变化检测[J]. 武汉大学学报(信息科学版), 2003, 28(s1):7-12.
	Li D R. Change detection from remote sensing images[J]. Geomatics and Information Science of Wuhan University, 2003, 28(s1):7-12.
[3]	任秋如, 杨文忠, 汪传建, 等. 遥感影像变化检测综述[J]. 计算机应用, 2021, 41(8):2294-2305. doi: 10.11772/j.issn.1001-9081.2020101632
	Ren Q R, Yang W Z, Wang C J, et al. Review of remote sensing image change detection[J]. Journal of Computer Applications, 2021, 41(8):2294-2305. doi: 10.11772/j.issn.1001-9081.2020101632
[4]	杨存建, 欧晓昆, 党承林, 等. 森林植被动态变化信息的遥感检测[J]. 地球信息科学, 2000, 2(4):71-74.
	Yang C J, Ou X K, Dang C L, et al. Detecting the change information of forest vegetation in Lushui County of Yunnan Province[J]. Geo-Information Science, 2000, 2(4):71-74.
[5]	Li P J, Xu H Q. Land-cover change detection using one-class support vector machine[J]. Photogrammetric Engineering and Remote Sensing, 2010, 76(3):255-263. doi: 10.14358/PERS.76.3.255
[6]	Cao G, Li Y P, Liu Y Z, et al. Automatic change detection in high-resolution remote-sensing images by means of level set evolution and support vector machine classification[J]. International Journal of Remote Sensing, 2014, 35(16):6255-6270. doi: 10.1080/01431161.2014.951740
[7]	Zhang Z J, Li A N, Lei G B, et al. Change detection of remote sensing images based on multiscale segmentation and decision tree algorithm over mountainous area:A case study in Panxi region,Sichuan Province[J]. Acta Ecologica Sinica, 2014, 34(24):7222-7232.
[8]	黄晓君, 颉耀文, 卫娇娇, 等. 基于变化检测-CART决策树模式自动识别沙漠化信息[J]. 灾害学, 2017, 32(1):36-42.
	Huang X J, Xie Y W, Wei J J, et al. Automatic recognition of desertification information based on the pattern of change detection-CART decision tree[J]. Journal of Catastrophology, 2017, 32(1):36-42.
[9]	Seo D K, Kim Y H, Eo Y D, et al. Generation of radiometric,phenological normalized image based on random forest regression for change detection[J]. Remote Sensing, 2017, 9(11):1163. doi: 10.3390/rs9111163
[10]	冯文卿, 眭海刚, 涂继辉, 等. 高分辨率遥感影像的随机森林变化检测方法[J]. 测绘学报, 2017, 46(11):1880-1890. doi: 10.11947/j.AGCS.2017.20170074
	Feng W Q, Sui H G, Tu J H, et al. Change detection method for high resolution remote sensing images using random forest[J]. Acta Geodaetica et Cartographica Sinica, 2017, 46(11):1880-1890. doi: 10.11947/j.AGCS.2017.20170074
[11]	Wang Q M, Shi W Z, Atkinson P M, et al. Land cover change detection at subpixel resolution with a hopfield neural network[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2015, 8(3):1339-1352. doi: 10.1109/JSTARS.4609443
[12]	Cao G, Liu Y Z, Shang Y F. Automatic change detection in remote sensing images using level set method with neighborhood constraints[J]. Journal of Applied Remote Sensing, 2014, 8(1):083678. doi: 10.1117/1.JRS.8.083678
[13]	严明, 曹国, 夏梦. 基于水平集演化和支持向量机分类的高分辨率遥感图像自动变化检测[J]. 哈尔滨理工大学学报, 2019, 24(1):78-84.
	Yan M, Cao G, Xia M. Automatic change detection of high resolution remote sensing images based on level set evolution and support vector machine classification[J]. Journal of Harbin University of Science and Technology, 2019, 24(1):78-84.
[14]	Bruzzone L, Prieto D F. Automatic analysis of the difference image for unsupervised change detection[J]. IEEE Transactions on Geoscience and Remote Sensing, 2000, 38(3):1171-1182. doi: 10.1109/36.843009
[15]	Zhou L C, Cao G, Li Y P, et al. Change detection based on conditional random field with region connection constraints in high-resolution remote sensing images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2016, 9(8):3478-3488. doi: 10.1109/JSTARS.4609443
[16]	魏立飞, 牟紫微, 王晓燕, 等. 基于CRF模型的高分辨率遥感影像变化检测[J]. 测绘通报, 2017, 9:28-31.
	Wei L F, Mou Z W, Wang X Y, et al. Change detection of high resolution remote sensing image based on CRF model[J]. Bulletin of Surveying and Mapping, 2017, 9:28-31.
[17]	Weismiller R A, Kristof S J, Scholz D K, et al. Change detection in coastal zone environments[J]. Photogrammetric Engineering and Remote Sensing, 1977, 42(12):1533-1539.
[18]	Malila W A. Change vector analysis:An approach for detecting forest changes with Landsat[J]. Proceedings of the Society of Photo-Optical Instrumentation Engineers, 1980:326-336.
[19]	Jensen J R, Toll D L. Detecting residential land use development at the urban fringe[J]. Photogrammetric Engineering and Remote Sensing, 1982, 48:629-643.
[20]	Byrne G F, Crapper P F, Mayo K K. Monitoring land cover change by principal component analysis of multitemporal Landsat data[J]. Remote Sensing of Environment, 1980, 10(3):175-184. doi: 10.1016/0034-4257(80)90021-8
[21]	Ingebritsen S E, Lyon R J P. Principal components analysis of multitemporal image pairs[J]. International Journal of Remote Sensing, 1985, 6(5):687-696. doi: 10.1080/01431168508948491
[22]	吴海平, 温礼, 邓凯, 等. 基于深度学习的高分辨率遥感影像自动变化检测[J]. 测绘与空间地理信息, 2021, 44(7):102-106.
	Wu H P, Wen L, Deng K, et al. Deep learning based change detection for high-resolution remote sensing images[J]. Geomatics & Spatial Information Technology, 2021, 44(7):102-106.
[23]	Su L Z, Gong M G, Zhang P Z, et al. Deep learning and mapping based ternary change detection for information unbalanced images[J]. Pattern Recognition, 2017, 66:213-228. doi: 10.1016/j.patcog.2017.01.002
[24]	Gong M G, Zhan T, Zhang P Z, et al. Superpixel-based difference representation learning for change detection in multispectral remote sensing images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2017, 55(5):2658-2673. doi: 10.1109/TGRS.2017.2650198
[25]	Zhang H, Gong M G, Zhang P Z, et al. Feature-level change detection using deep representation and feature change analysis for multispectral imagery[J]. IEEE Geoscience and Remote Sensing Letters, 2016, 13(11):1666-1670. doi: 10.1109/LGRS.2016.2601930
[26]	Mou L C, Bruzzone L, Zhu X X. Learning spectral-spatial-temporal features via a recurrent convolutional neural network for change detection in multispectral imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(2):924-935. doi: 10.1109/TGRS.2018.2863224
[27]	Li S C, Han P C, Bu S H, et al. Change detection in images using shape-aware Siamese convolutional network[J]. Engineering Applications of Artificial Intelligence, 2020, 94:103819. doi: 10.1016/j.engappai.2020.103819
[28]	Chen H R X, Wu C, Du B, et al. Change detection in multisource VHR images via deep Siamese convolutional multiple-layers recurrent neural network[J]. IEEE Transactions on Geoscience and Remote Sensing, 2020, 58(4):2848-2864. doi: 10.1109/TGRS.36
[29]	Gong M G, Niu X D, Zhang P Z, et al. Generative adversarial networks for change detection in multispectral imagery[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(12):2310-2314. doi: 10.1109/LGRS.2017.2762694
[30]	Li B L, Zhou Q M. Accuracy assessment on multi-temporal land-cover change detection using a trajectory error matrix[J]. International Journal of Remote Sensing, 2009, 30(5):1283-1296. doi: 10.1080/01431160802474022

[1]	伍江, 刘春, 应申, 余婷. 城区范围空间划定方法研究[J]. 自然资源遥感, 2021, 33(4): 89-97.
[2]	刘尚旺, 崔智勇, 李道义. 基于Unet网络多任务学习的遥感图像建筑地物语义分割[J]. 国土资源遥感, 2020, 32(4): 74-83.

Viewed

Full text

Abstract

Cited

Shared

Discussed