Remote Sensing for Natural Resources >
Change detection network for dual-temporal optical remote sensing images integrating fast Fourier transform and efficient multi-head self-attention
Received date: 2024-07-31
Revised date: 2025-04-02
Online published: 2026-06-03
The deep learning-based change detection of remote sensing images has seen rapid advances in the past few years. However,it still faces challenges for change detection in complex scenes,such as incomplete recognition and high false detection rates. In response to these challenges,this paper proposed the FTUNet,a network based on SNUnet that integrates the fast Fourier transform (FFT) and efficient multi-head self-attention (EMHSA). Specifically,the FFT module in the network enabled style unification of dual-temporal images,reducing false detection caused by “pseudo changes” due to external factors such as light variations. Additionally,the EMHSA was introduced in the feature extraction stage to fully extract the contextual information from the feature maps,thereby enhancing the segmentation integrity of target changes. Experiments on the LEVIR-CD and SYSU-CD public datasets showed that the FTUNet exhibited increases of 1.42 and 1.53 percentage points in F1 score,as well as increases of 2.31 and 2.07 percentage points in intersection over union (IoU),compared to the SNUNet.
WANG Xingwei , TANG Kangqi , LIU Yan , LIU Huan . Change detection network for dual-temporal optical remote sensing images integrating fast Fourier transform and efficient multi-head self-attention[J]. Remote Sensing for Natural Resources, 2025 , 37(5) : 113 -121 . DOI: 10.6046/zrzyyg.2024261
表1 消融实验结果Tab.1 Results of ablation experiment (%) |
| 网络模型 | P | R | IoU | F1 |
|---|---|---|---|---|
| SNUNet(基准网络) | 89.91 | 87.21 | 79.44 | 88.54 |
| SNUNet+FFT | 90.81 | 87.86 | 80.69 | 89.31 |
| SNUNet+LRCFEM | 90.46 | 89.15 | 81.50 | 89.81 |
| FTUNet(本文网络) | 90.72 | 89.22 | 81.75 | 89.96 |
表3 LEVIR-CD数据集对比实验结果Tab.3 Comparison experiment results of LEVIR-CD dataset (%) |
| 网络模型 | P | R | IoU | F1 |
|---|---|---|---|---|
| FC-EF | 86.69 | 79.87 | 71.15 | 83.14 |
| FC-Siam-conc | 89.41 | 85.02 | 77.24 | 87.16 |
| FC-Siam-diff | 90.12 | 83.01 | 76.09 | 86.42 |
| SNUNet | 89.91 | 87.21 | 79.44 | 88.54 |
| BIT | 92.01 | 86.72 | 80.65 | 89.28 |
| FTUNet(本文网络) | 90.72 | 89.22 | 81.75 | 89.96 |
表4 LEVIR-CD数据集对比实验可视化结果Tab.4 Visualization results of LEVIR-CD dataset comparison experiment |
| 序号 | 影像A | 影像B | 标签 | FE-EF | FC-Siam- conc | FC-Siam- diff | BIT | SNUNet | FTUNet |
|---|---|---|---|---|---|---|---|---|---|
| a | ![]() | ||||||||
| b | ![]() | ||||||||
| c | ![]() | ||||||||
| d | ![]() | ||||||||
| e | ![]() | ||||||||
| 图例 | ![]() | ||||||||
表5 SYSU-CD数据集对比实验结果Tab.5 Comparison experiment results of SYSU-CD dataset (%) |
| 网络模型 | P | R | IoU | F1 |
|---|---|---|---|---|
| FC-EF | 78.60 | 69.55 | 58.47 | 73.80 |
| FC-Siam-conc | 81.11 | 70.76 | 60.75 | 75.59 |
| FC-Siam-diff | 90.35 | 49.33 | 46.86 | 63.81 |
| SNUNet | 81.90 | 74.37 | 63.87 | 77.95 |
| BIT | 81.71 | 73.49 | 63.11 | 77.38 |
| FTUNet(本文网络) | 79.95 | 79.01 | 65.94 | 79.48 |
表6 SYSU-CD数据集对比实验可视化结果Tab.6 Visualization results of SYSU-CD dataset comparison experiment |
| 序号 | 影像A | 影像B | 标签 | FE-EF | FC-Siam- conc | FC-Siam- diff | BIT | SNUNet | FTUNet |
|---|---|---|---|---|---|---|---|---|---|
| a | ![]() | ||||||||
| b | ![]() | ||||||||
| c | ![]() | ||||||||
| d | ![]() | ||||||||
| e | ![]() | ||||||||
| 图例 | ![]() | ||||||||
| [1] |
|
| [2] |
|
| [3] |
|
| [4] |
|
| [5] |
|
| [6] |
|
| [7] |
|
| [8] |
|
| [9] |
|
| [10] |
|
| [11] |
|
| [12] |
|
| [13] |
|
| [14] |
|
| [15] |
|
| [16] |
|
| [17] |
宋嘉鑫, 李轶鲲, 杨树文, 等. 基于后验概率空间变化向量分析的NSCT高分辨率遥感影像变化检测[J]. 自然资源遥感, 2024, 36(3):128-136.doi:10.6046/zrzyyg.2023079.
|
| [18] |
柴佳兴, 张云生, 杨振, 等. 联合无人机影像生成DSM和DOM的多层次建筑物变化检测[J]. 自然资源遥感, 2024, 36(2):80-88.doi:10.6046/zrzyyg.2023001.
|
| [19] |
|
| [20] |
|
| [21] |
|
| [22] |
|
/
| 〈 |
|
〉 |