计算机科学 ›› 2026, Vol. 53 ›› Issue (6): 270-280.doi: 10.11896/jsjkx.250400015
马宁1,3, 常霞2,3, 苑玲玉2,3
MA Ning1,3, CHANG Xia2,3, YUAN Lingyu2,3
摘要： 遥感卫星成像过程中,通常会获取分辨率高的全色图像和分辨率低但包含丰富光谱信息的多光谱图像。为了充分利用这两类图像的优势,提出了一种基于双边引导滤波和多特征重校准的全色锐化方法。该方法设计了双流U-Net锐化网络,通过并行分支分别提取多光谱和全色图像的多尺度特征,避免了直接进行特征融合易导致信息损失的问题。此外,提出了双边引导滤波模块,该模块能够通过并行的特征引导路径和自适应权重融合机制增强特征交互。同时,设计了多特征重校准模块,该模块结合多方向边缘检测器和多头特征提取机制,通过动态特征重标定策略增强空间细节的重建能力,有效避免了过度增强导致的伪影问题。实验表明,该方法具有良好的波段适应性,不仅能有效处理4波段的GeoEye1和PLeiades1数据,在8波段的WorldView2和WorldView3数据集上同样表现出色。定量评价结果显示,所提方法在WorldView3数据集上CC值达到0.957 3;在GeoEye1数据集上PSNR值达到38.093 8 dB,ERGAS值较最优对比方法降低94.9%,Q4值提高了2.3%。该方法在主观视觉效果和客观评价指标上均优于现有方法,为遥感图像全色锐化任务提供了一种有效的解决方案。
中图分类号:
| [1]SEELAN S K,LAGUETTE S,CASADY G M,et al.Remote sensing applications for precision agriculture:A learning community approach[J].Remote Sensing of Environment,2003,88(1/2):157-169. [2]HUANG G,WAN Z,LIU X,et al.Ship detection based onsqueeze excitation skip-connection path networks for optical remote sensing images[J].Neurocomputing,2019,332:215-223. [3]BAKKEN S,LANGER D D,GARRETT J L,et al.Spatial resolution enhancement of hyperspectral images from miniaturized sensors[C]//Image and Signal Processing for Remote Sensing XXIX.SPIE,2023:24-38. [4]LIU Z,SONG W,DENG A,et al.HyperESSANet:An Efficient Spatial-Spectral Attention Network for Hyperspectral Pansharpening[J].IEEE Sensors Journal,2025,25(13):25211-25266. [5]OTAZU X,GONZÁLEZ-AUDÍCANA M,FORS O,et al.Introduction of sensor spectral responseinto image fusion methods.Application to wavelet-based methods[J].IEEE Transactions on Geoscience and Remote Sensing,2005,43(10):2376-2385. [6]VIVONE G,DALLA M M,GARZELLI A,et al.A new benchmark based on recent advances in multispectral panshar-pening:Revisiting pansharpening with classical and emerging panshar-pening methods[J].IEEE Geoscience and Remote Sensing Maga-zine,2020,9(1):53-81. [7]XU S,ZHANG J,ZHAO Z,et al.Deep gradient projection networks for pan-sharpening[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2021:1366-1375. [8]ALCARAS E,PARENTE C,VALLARIO A.Automation ofpan-sharpening methods for pléiades images using GIS basic functions[J].Remote Sensing,2021,13(8):1550. [9]WANG M,HE L X,CHENG Y F,et al.A pans-harpening me-thod based on adaptive Gaussian filter and SFIM model [J].Acta Geodaetica et Cartographica Sinica,2018,47(1):82-90. [10]LI H,NIE R C,YANG X F,et al.Remote Sensing Image Fusion with Dual-branch Attention Network[J].Computer Science,2023,50(S2):230200072-7. [11]MASI G,COZZOLINO D,VERDOLIVA L,et al.Pansharpening by convolutional neural networks[J].Remote Sensing,2016,8(7):594. [12]RONNEBERGER O,FISCHER P,BROX T.Unet:Convolutio-nal networks for biomedical image segmentation[C]//18th International Conference Medical Image Computing and Compu-ter-assisted Intervention(MICCAI 2015).Springer,2015:234-241. [13]YAN C,FAN X,FAN J,et al.Improved U-Net remote sensing classification algorithm based on Multi-Feature Fusion Perception[J].Remote Sensing,2022,14(5):1118. [14]ZHANG C,DENG X,LING S H.Next-Gen medical imaging:U-Net evolution and the rise of transformers[J].Sensors,2024,24(14):4668. [15]ADEEL H,ALI S S,RIAZ M M,et al.Learning Deep Pyramid-based Representations for Pan-sharpening[J].Arabian Journal for Science and Engineering,2022,47(8):10655-10666. [16]LI J,FAN W,LIAN T,et al.Cross-Attention-Based Common and Unique Feature Extraction for Pansharpening[J].IEEE Geoscience and Remote Sensing Letters,2023,20:5002905. [17]GUO H,JIN X,JIANG Q,et al.DMF-net:a dual remote sensing image fusion network based on multiscale convolutional dense connectivity with performance measure[J].IEEE Transactions on Instrumentation and Measurement,2024,73:1-15. [18]WU H,ZHENG S,ZHANG J,et al.Fast end-to-end trainable guided filter[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2018:1838-1847. [19]CHEN Y,LIU H,FANG F.A novel pansharpening methodbased on cross stage partial network and transformer[J].Scientific Reports,2024,14(1):12631. [20]DE BORBA A A,MARENGONI M,FRERY A C.Fusion ofevidences in intensity channels for edge detection in PolSAR images[J].IEEE Geoscience and Remote Sensing Letters,2020,19:1-5. [21]TANG L,DENG Y,YI X,et al.DRMF:Degradation-RobustMulti-Modal Image Fusion via Composable Diffusion Prior[C]//Proceedings of the 32nd ACM International Conference on Multimedia.2024:8546-8555. [22]LI T,LIU L,LI M.Multi-Scale Residual Depthwise Separable Convolution for Metro Passenger Flow Prediction[J].Applied Sciences,2023,13(20):11272. [23]YADAV O,GHOSAL K,LUTZ S,et al.Frequency-domain loss function for deep exposure correction of dark images[J].Signal,Image and Video Processing,2021,15(8):1829-1836. [24]AIAZZI B,ALPARONE L,BARONTI S,et al.MTF-tailored multiscale fusion of high-resolution MS and Pan imagery[J].Photogrammetric Engineering & Remote Sensing,2006,72(5):591-596. [25]ZHANG H,WANG H,TIAN X,et al.P2Sharpen:A progressive pansharpening network with deep spectral transformation[J].Information Fusion,2023,91:103-122. [26]GARDNER B,HASKELL J,MATOUSEK P,et al.Guidedprincipal component analysis(GPCA):a simple method for improving detection of a known analyte[J].Analyst,2024,149(1):205-211. [27]JAVAN F D,SAMADZADEGAN F,MEHRAVAR S,et al.A review of image fusion techniques for pansharpeningof high-reso-lution satellite imagery[J].ISPRS Journal of Photogrammetry and Remote Sensing,2021,171:101-117. |
|
||
首页