注意力引导的三维卷积网络用于遥感场景变化检测

doi:10.3969/j.issn.0255-8297.2021.02.009

摘要/Abstract

摘要： 场景级变化检测策略可以容忍高分遥感影像的大量噪声，进而从语义层级更准确地描述遥感图像在前后时相的变化，为高分辨率影像变化检测提供了可能。本文提出了一种注意力引导的三维卷积神经网络用于高分遥感影像场景变化检测的方法。首先构建一个在AlexNet基础上进行简化的三维卷积网络，然后加入一个语义注意力模块来进一步提取地表覆盖变化显著的候选判别区域；最后输入分类层得到分类结果，整个框架以端对端、可训练的方式进行组织，直接由双时相场景切片通过卷积网络得到变化检测结果。为评估场景级变化检测方法性能，本文制作了一个语义级高分遥感影像场景变化检测数据集，在该数据集上的实验结果显示本文方法变化检测的准确率高于相关方法，验证了方法的有效性，初步展示了基于深度学习的场景级遥感变化检测的发展前景。

关键词: 场景级变化检测, 语义注意力模块, 三维卷积神经网络, 高分遥感解译, 场景变化检测数据集

Abstract: With high tolerance to the great amount of noise and precise depiction of image changes in high resolution remote sensing images (HRRSI), scene-level change detection strategy makes it possible to detect changes in HRRSI. In this paper, we propose an attention guided 3D ConvNet for HRRSI change detection. Firstly, we develop a simplified 3D AlexNet to extract convolutional features. Then we add a semantic attention module (SAM) to further extract the discriminative regions which strongly relate to land-cover changes. Finally, the refined features are fed into a classification layer to organize the whole framework in an end-to-end trainable manner. Scenes in different phases are put into the convolutional neural network (CNN) with the result of change detection. In order to evaluate the performance of scene level change detection methods, we create a public semantic level high resolution remote sensing images change detection benchmark. Experimental results on this dataset are obviously better than other related methods, demonstrate the effectiveness of our method, and show the prospect of scene level remote sensing change detection based on deep learning.

Key words: scene-level change detection, semantic attention module, 3D ConvNet, high resolution remote sensing interpretation, scene-level change detection benchmark

中图分类号:

TP751.1

张涵, 秦昆, 毕奇, 张晔, 许凯. 注意力引导的三维卷积网络用于遥感场景变化检测[J]. 应用科学学报, 2021, 39(2): 272-280.

ZHANG Han, QIN Kun, BI Qi, ZHANG Ye, XU Kai. Attention Guided 3D ConvNet for Aerial Scene Change Detection[J]. Journal of Applied Sciences, 2021, 39(2): 272-280.

参考文献

[1] Xia G S, Bai X, Ding J, et al. DOTA:a large-scale dataset for object detection in aerial images[C]//IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2018.
[2] 黄昕, 张良培, 李平湘. 融合形状和光谱的高空间分辨率遥感影像分类[J]. 遥感学报, 2007, 11(2):193-200. Huang X, Zhang L P, Li P X. Classification of high spatial resolution remotely sensed imagery based on the fusion of spectral and shape features[J]. Journal of Remote Sensing, 2007, 11(2):193-200. (in Chinese)
[3] 胡荣明, 黄小兵, 黄远程. 增强形态学建筑物指数应用于高分辨率遥感影像中建筑物提取[J]. 测绘学报, 2014, 43(5):514-520. Hu R M, Huang X B, Huang Y C. An enhanced morpho logical building index for building extraction from high-resolution images[J]. Acta Geodaetica et Cartographica Sinica, 2014, 43(5):514-520. (in Chinese)
[4] 冯发杰, 吏军平, 丁亚洲, 等. 基于多尺度视觉特征组合的高分遥感影像目标检测[J]. 应用科学学报, 2018, 36(3):471-484. Feng F J, Li J P, Ding Y Z, et al. Target detection from high resolution remote sensing images based on combination of multi-scale visual features[J]. Journal of Applied Sciences, 2018, 36(3):471-484. (in Chinese)
[5] 毕奇, 童心, 张济勇, 等. 基于PLSA和BoW的高分遥感影像小型港口检测[J]. 应用科学学报, 2019, 37(3):301-312. Bi Q, Tong X, Zhang J Y, et al. Small harbor detection based on PLSA and BoW in high resolution remotely sensed imagery[J]. Journal of Applied Sciences, 2019, 37(3):301-312. (in Chinese)
[6] 郑卓, 方芳, 刘袁缘, 等. 高分辨率遥感影像场景的多尺度神经网络分类法[J]. 测绘学报, 2018, 47(5):620-630. Zheng Z, Fang F, Liu Y Y, et al. Joint multi-scale convolution neural network for scene classification of high resolution remote sensing imagery[J]. Acta Geodaetica et Cartographica Sinica, 2018, 47(5):620-630. (in Chinese)
[7] 李欣, 唐文莉, 杨博. 利用深度残差网络的高分遥感影像语义分割[J]. 应用科学学报, 2019, 37(2):282-290. Li X, Tang W L, Yang B. Semantic segmentation of high-resolution remote sensing image based on deep residual network[J]. Journal of Applied Sciences, 2019, 37(2):282-290. (in Chinese)
[8] Bi Q, Qin K, Zhang H, et al. A multi-scale filtering building index for building extraction in very high-resolution satellite imagery[J]. Remote Sensing, 2019, 11(5):482.
[9] Liu Y F, Zhong Y F, Qin Q Q. Scene classification based on multiscale convolutional neural network[J]. IEEE Transactions on Geoscience and Remote Sensing, 2018, 56(12):7109-7121.
[10] He D, Zhong Y, Zhang L. Spatiotemporal subpixel geographical evolution mapping[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(4):2198-2220.
[11] Du P, Liu S, Gamba P, et al. Fusion of difference images for change detection over urban areas[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2012, 5(4):1076-1086.
[12] Liu S, Bruzzone L, Bovolo F, et al. Unsupervised multitemporal spectral unmixing for detecting multiple changes in hyperspectral images[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(5):2733-2748.
[13] Wang X, Du P, Liu S, et al. Unsupervised change detection in VHR images based on morphological profiles and automated training sample extraction[C]//International Workshop on the Analysis of Multitemporal Remote Sensing Images. 2019.
[14] Lü Z Y, Zhang P L, Atli B J. Automatic object-oriented, spectral-spatial feature extraction driven by Tobler's first law of geography for very high resolution aerial imagery classification[J]. Remote Sensing, 2017, 9(3):285.
[15] Lü Z Y, Shi W, Zhang X, et al. Landslide inventory mapping from bitemporal high-resolution remote sensing images using change detection and multiscale segmentation[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2018, 11(5):1520-1532.
[16] Du B, Wang Y, Wu C, et al. Unsupervised scene change detection via latent dirichlet allocation and multivariate alteration detection[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2018, 11(12):4676-4689.
[17] Wen D W, Huang X, Zhang L P, et al. A novel automatic change detection method for urban high-resolution remotely sensed imagery based on multiindex scene representation[J]. IEEE Transactions on Geoscience and Remote Sensing, 2016, 54(1):609-625.
[18] Bi Q, Qin K, Li Z L, et al. Multiple instance dense connected convolution neural network for aerial image scene classification[C]//IEEE International Conference on Image Processing (ICIP). 2019.
[19] Bi Q, Qin K, Zhang H, et al. APDCNet:attention pooling-based convolutional neural network for aerial scene classification[J]. IEEE Geoscience and Remote Sensing Letters, 2019(99):1-5.
[20] Bao S Q, Wang P, Mok T C W, et al. 3D randomized connection network with graph-based label inference[J]. IEEE Transactions on Image Processing, 2018, 27(8):3883-3892.
[21] Song H, Tian L, Li C. 3D convolutional network based foreground feature fusion[C]//IEEE International Symposium on Multimedia (ISM). 2018.
[22] Seo P H, Lin Z, Cohen S, et al. Hierarchical attention networks[DB/OL]. 2016[2020-07-27]. https://arxiv.org/abs/1606.02393v1.
[23] Xu H J, Saenko K. Ask, attend and answer:exploring question-guided spatial attention for visual question answering[C]//European Conference on Computer Vision. 2016.
[24] Sun W Y, Zhao H T, Zhong J. A visual attention based ROI detection method for facial expression recognition[J]. Neurocomputing, 2018, 296(28):12-22.
[25] Nardelli P, Jimenez-Carretero D, Bermejo-Pelaez D, et al. Pulmonary artery-vein classification in CT images using deep learning[J]. IEEE Transactions on Medical Imaging, 2018, 37(11):2428-2440.