Anterior|Siguiente

Stripe noise removal in high resolution satellite remote sensing images based on deep learning

  • role： First author第一作者
  • Affiliation：

    School of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430079, China

  • Email：934741099@qq.com
  • Introduction：高浩博，研究方向为摄影测量与遥感。E-mail： 934741099@qq.com
GAO Haobo1,  
  • Affiliation：

    China Centre for Resources Satellite Data and Application, Beijing 100094, China

BU Tong3,  
  • Affiliation：

    School of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430079, China

LI Xin1,  
  • role： Corresponding author通信作者
  • Affiliation：

    Hubei Provincial Research Institute of Land and Resources, Wuhan 430071, China

  • Email：76235431@qq.com
  • Introduction：陆世东，研究方向为自然资源大数据。E-mail：76235431@qq.com
LU Shidong2*,  
  • Affiliation：

    China Centre for Resources Satellite Data and Application, Beijing 100094, China

ZHONG Huimin3,  
  • Affiliation：

    China Centre for Resources Satellite Data and Application, Beijing 100094, China

CUI Lin3

resumen

Affected by imaging conditions, data transmission, and other factors, stripe noise is common in satellite remote sensing images. It seriously restricts the quality and further use of images. In early studies, various denoising methods, such as statistics-based methods, filtering-based methods, and optimization-based methods, have been proposed to overcome the above problems. These proposed methods have achieved inspiring results in some aspects. However, they still suffer from poor adaptability, low denoising efficiency, and the need for prior knowledge. Therefore, stripe noise removal remains a challenging task.In this study, we take advantage of the convolutional deep network while considering the characteristics of the stripe noise image itself. A deep-learning-based method is proposed, which includes three parts: a feature extraction module, a feature fusion module, and a stripe denoising module. The feature extraction module uses the convolutional layer of the same channel with different strides to extract features. As a result, different-scale feature maps of the noisy image are obtained for the following feature fusion module. The feature fusion module upsamples different-scale feature maps. It fuses these upsampled feature maps through the element-wise addition method. Finally, a denoising network is used to predict the components of stripe noise. The stripe component is subtracted from the noise image based on predictions. Given the difficulties in obtaining real noise samples, the network is trained by simulation samples. Then, it is extended to denoise real images.Experiments on simulation and real images show the excellent performance of our network. In the quantitative assessment, the PSNR and the SSIM of our network when simulated images are used are higher than those of the four methods. In the visual assessment, our network performs well on homogeneous and nonhomogeneous objects. Our network denoises more efficiently and retains more details of ground features than traditional methods and other denoising networks. In real noise images, our method achieves the best denoising performance with the highest ICV and the lowest MRD. Compared with traditional methods, our network has a fast denoising speed, approximately 100 times faster than the denoising speed of the optimization-based denoising method. The above experimental results demonstrate that our network has the best denoising performance in simulated and real images.In this study, a convolutional neural network denoising method based on multiscale feature fusion is proposed based on the fully convolutional neural network. The method uses residual learning to predict the strip-noise components on images. It achieves clean images by subtracting the strip components from noisy images. Experiments demonstrate that compared with traditional methods, deep learning denoising methods are adaptive for removing stripe noise of different intensities without losing image details. The strategy of feature fusion and residual learning can effectively improve the training speed and denoising accuracy of the network. In the future, skip-connected and batch normalization layers will be included to optimize training speed and improve denoising performance. Further studies will be conducted in terms of the transfer ability of the network and the extension of its application in other types of remote sensing images, such as aerial images and hyperspectral images.

palabra clave

high-resolution image;deep learning;stripe noise;convolutional neural network;Feature fusion

References

  1. 1.
    Bouali M and Ladjal S. 2011. Toward optimal destriping of MODIS data using a unidirectional variational model. IEEE Transactions on Geoscience and Remote Sensing, 49(8): 2924-2935
  2. 2.
    Carfantan H and Idier J. 2010. Statistical linear destriping of satellite-based pushbroom-type images. IEEE Transactions on Geoscience and Remote Sensing, 48(4): 1860-1871
  3. 3.
    Chang Y, Fang H Z, Yan L X and Liu H. 2013. Robust destriping method with unidirectional total variation and framelet regularization. Optics Express, 21(20): 23307-23323
  4. 4.
    Chang Y, Yan L X, Fang H Z, Zhong S and Liao W S. 2019. HSI-DeNet: hyperspectral image restoration via convolutional neural network. IEEE Transactions on Geoscience and Remote Sensing, 57(2): 667-682
  5. 5.
    Chang Y, Yan L X, Tao W and Zhong S. 2016. Remote sensing image stripe noise removal: from image decomposition perspective. IEEE Transactions on Geoscience and Remote Sensing, 54(12): 7018-7031
  6. 6.
    Chen J S, Shao Y, Guo H D, Wang W M and Zhu B Q. 2003. Destriping CMODIS data by power filtering. IEEE Transactions on Geoscience and Remote Sensing, 41(9): 2119-2124
  7. 7.
    Fischel D. 1984. Validation of the thematic mapper radiometric and geometric correction algorithms. IEEE Transactions on Geoscience and Remote Sensing, GE-22(3): 237-242
  8. 8.
    He K M, Zhang X Y, Ren S Q and Sun J. 2016. Deep residual learning for image recognition//2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, USA: IEEE: 770-778
  9. 9.
    Horn B K P and Woodham R J. 1979. Destriping LANDSAT MSS images by histogram modification. Computer Graphics and Image Processing, 10(1): 69-83
  10. 10.
    Kim J, Lee J K and Lee K M. 2016. Accurate image super-resolution using very deep convolutional networks//2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, USA: IEEE: 1646-1654
  11. 11.
    Kuang X D, Sui X B, Chen Q and Gu G H. 2017. Single infrared image stripe noise removal using deep convolutional networks. IEEE Photonics Journal, 9(4): 3900913
  12. 12.
    Liu P and Fang R G. 2017. Wide inference network for image denoising. CoRR, abs/1707.05414.
  13. 13.
    Pande-Chhetri R and Abd-Elrahman A. 2011. De-striping hyperspectral imagery using wavelet transform and adaptive frequency domain filtering. ISPRS Journal of Photogrammetry and Remote Sensing, 66(5): 620-636
  14. 14.
    Rakwatin P, Takeuchi W and Yasuoka Y. 2007. Stripe noise reduction in MODIS data by combining histogram matching with facet filter. IEEE Transactions on Geoscience and Remote Sensing, 45(6): 1844-1856
  15. 15.
    Shen H F and Zhang L P. 2009. A MAP-based algorithm for destriping and inpainting of remotely sensed images. IEEE Transactions on Geoscience and Remote Sensing, 47(5): 1492-1502
  16. 16.
    Song Q, Wang Y H, Yan X Y and Gu H G. 2018. Remote sensing images stripe noise removal by double sparse regulation and region separation. Remote Sensing, 10(7): 998
  17. 17.
    Sun B, Li J L, Zhang X X and Ren J Y. 2014. Interleaving assembly of TDICCDs on 600 mm focal plane. Optics and Precision Engineering, 22(11): 2908-2913
  18. 18.
    Wang C, Wang X and Ji S. 2019. Stripe noise removal of remote images based on variation. Journal of Xi'an Jiaotong University, 53(3): 143-149
  19. 19.
    Wang M, Zheng X H, Pan J and Wang B. 2016. Unidirectional total variation destriping using difference curvature in MODIS emissive bands. Infrared Physics and Technology, 75: 1-11
  20. 20.
    Wang Z, Bovik A C, Sheikh H R and Simoncelli E P. 2004. Image quality assessment: from error visibility to structural similarity. IEEE Transactions on Image Processing, 13(4): 600-612
  21. 21.
    Xiao P F, Guo Y C and Zhuang P X. 2018. Removing stripe noise from infrared cloud images via deep convolutional networks. IEEE Photonics Journal, 10(4): 7801114
  22. 22.
    Zhang K, Zuo W M, Chen Y J, Meng D Y and Zhang L. 2017. Beyond a gaussian denoiser: residual learning of deep CNN for image denoising. IEEE Transactions on Image Processing, 26(7): 3142-3155
  23. 23.
    Zhao B H, He B, Yang L H, Tao M H and Ren J Y. 2012. Destriping for TDI-CCD remote sensing images. Chinese Journal of Space Science, 32(2): 298-304
  24. 24.
    Zhong Y F, Li W Q, Wang X Y, Jin S Y and Zhang L P. 2020. Satellite-ground integrated destriping network: a new perspective for EO-1 Hyperion and Chinese hyperspectral satellite datasets. Remote Sensing of Environment, 237: 111416

Leer el texto completo

The above content is generated by Large Model Translation. The translated content is for reference only. We do not assume any commercial or legal responsibilty for any consequences arising from the use of our website