이전|다음

화성 표면 돌 조각의 합성곱 자기 주의 네트워크 인식 방법

  • role： First author第一作者
  • Affiliation：

    College of Automation, Beijing Information Science and Technology University, Beijing 100192, China

    State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Researc Institute, Chinese Academy of Sciences, Beijing 100101, China

  • Email：2023020486@bistu.edu.cn
  • Introduction：曹学欢， 研究方向为行星制图与遥感。E-mail： 2023020486@bistu.edu.cn
CAO Xuehuan12,  
  • role： Corresponding author通信作者
  • Affiliation：

    State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Researc Institute, Chinese Academy of Sciences, Beijing 100101, China

  • Email：pengman@aircas.ac.cn
  • Introduction：彭嫚，研究方向为行星制图与遥感。E-mail： pengman@aircas.ac.cn
PENG Man2*,  
  • Affiliation：

    State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Researc Institute, Chinese Academy of Sciences, Beijing 100101, China

WAN Wenhui2,  
  • Affiliation：

    State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Researc Institute, Chinese Academy of Sciences, Beijing 100101, China

    University of Chinese Academy of Sciences, Beijing 100049, China

WANG Biao23,  
  • Affiliation：

    State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Researc Institute, Chinese Academy of Sciences, Beijing 100101, China

WANG Yexin2,  
  • Affiliation：

    State Key Laboratory of Remote Sensing and Digital Earth, Aerospace Information Researc Institute, Chinese Academy of Sciences, Beijing 100101, China

DI Kaichang2,  
  • Affiliation：

    College of Automation, Beijing Information Science and Technology University, Beijing 100192, China

LI Lu1

추상적인

화성 표면에 널리 분포된 돌 조각들은 화성 탐사차의 안전한 주행에 잠재적인 위협을 제기하며, 동시에 돌 조각의 분포 특성은 화성 착륙 지역의 지질 진화를 연구하는 데 중요한 단서를 제공합니다. 그러나 화성 탐사차 이미지 내 돌 조각 인식은 여러 도전에 직면해 있습니다: 돌 조각과 배경의 경계가 흐릿하여 윤곽 추출이 어려우며, 표면 질감 특징의 유사성으로 인해 오판이 발생하고, 기존의 화성 돌 조각 실제 데이터셋이 부족하여 모델 학습에 제약이 있습니다. 본 논문에서는 화성 탐사차 이미지 내 돌 조각의 정밀 인식을 위해 합성곱 자기 주의 네트워크 기반의 돌 조각 자동 인식 모델을 제안하여 이미지의 픽셀 단위 분할을 실현하였습니다. 본 모델은 인코더-디코더 구조를 채택하였으며, 인코더는 합성곱 신경망을 기반으로 이미지 특징을 추출하고 개선된 자기 주의 모듈을 삽입하여 모델의 문맥 정보 인지 능력을 강화하며, 디코더는 인코더가 추출한 특징을 이미지 공간으로 매핑하여 정밀한 분할을 수행합니다. 모델 성능 검증을 위해 본 논문에서는 “주룽” 화성차 영상을 주석 처리하고 천문 데이터셋을 구축하였으며, 시뮬레이션 돌 조각 데이터셋 Synmars, Simmars6k, “큐리오시티” 영상 데이터셋 MarsData-v2 등 여러 데이터셋과 결합하여 모델의 성능을 테스트하고 검증하였습니다. 또한 본 모델을 DeepLabv3+, Unet++, Segformer, Marsnet 등 다양한 방법과 정밀도를 비교하였습니다. 평균 픽셀 정밀도, 재현율, 교집합비율 등의 지표를 사용한 평가 결과, 본 모델이 돌 조각을 정확하게 인식할 수 있음을 보여주었으며, 시뮬레이션 데이터셋에서는 정밀도와 재현율이 90% 이상에 도달하였고 실제 데이터셋에서도 최고 성능을 기록하였습니다.

키워드

화성; 돌 조각 추출; 합성곱 신경망; 트랜스포머; 특징 추출

References

  1. 1.
    Cao H, Wang Y Y, Chen J, Jiang D S, Zhang X P, Tian Q and Wang M N. 2021. Swin-Unet: Unet-like pure transformer for medical image segmentation. arXiv: 2105.05537
  2. 2.
    Chen F X. 2008. Image segmentation technology based on region growing method. Science and Technology Information, (15): 58-59
  3. 3.
    Chen G P, Li L, Zhang J X and Dai Y. 2023. Rethinking the unpretentious U-Net for medical ultrasound image segmentation. Pattern Recognition, 142: 109728
  4. 4.
    Chen J N, Lu Y Y, Yu Q H, Luo X D, Adeli E, Wang Y, Lu L, Yuille A L and Zhou Y Y. 2021. TransUNet: transformers make strong encoders for medical image segmentation. arXiv: 2102.04306
  5. 5.
    Chen L C, Zhu Y K, Papandreou G, Schroff F and Adam H. 2018. Encoder-decoder with atrous separable convolution for semantic image segmentation. arXiv: 1802.02611
  6. 6.
    Ding S F, Qi B J and Tan H Y. 2011. An overview on theory and algorithm of support vector machines. Journal of University of Electronic Science and Technology of China, 40(1): 2-10
  7. 7.
    Duan R L, Li Q X and Li Y H. 2005. Summary of image edge detection. Optical Technique, 31(3): 415-419
  8. 8.
    Fang K N, Wu J B, Zhu J P and Xie B C. 2011. A review of technologies on random forests. Statistics and Information Forum, 26(3): 32-38
  9. 9.
    Gou S, Yue Z Y, Di K C and Niu S L. 2021. Rampart craters in the Isidis Planitia, Mars: remote sensing analysis and environment implications. National Remote Sensing Bulletin, 25(7): 1374-1384
  10. 10.
    Han S Q and Wang L. 2002. A survey of thresholding methods for image segmentation. Systems Engineering and Electronics, 24(6): 91-94, 102
  11. 11.
    He K M, Zhang X Y, Ren S Q and Sun J. 2016. Deep residual learning for image recognition//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas: IEEE: 770-778
  12. 12.
    Hendrycks D and Gimpel K. 2023. Gaussian error linear units (GELUs). arXiv:1606.08415
  13. 13.
    Jain A K. 2008. Data clustering: 50 years beyond K-means//Daelemans W, Goethals B and Morik K, eds. Machine Learning and Knowledge Discovery in Databases. Antwerp: Springer: 3-4
  14. 14.
    Jia Y T, Su Z J, Wan G, Liu L and Liu J. 2023. AE-TransUNet+: an enhanced hybrid transformer network for detection of lunar south small craters in LRO NAC images. IEEE Geoscience and Remote Sensing Letters, 20: 6007405
  15. 15.
    Jia Y T, Wan G, Li W H, Li C, Liu J, Cong D W and Liu L. 2024. EDR-TransUnet: integrating enhanced dual relation-attention with transformer U-Net for multiscale rock segmentation on mars. IEEE Transactions on Geoscience and Remote Sensing, 62: 4601416
  16. 16.
    Li H C, Qiu L W, Li Z, Meng B, Huang J B and Zhang Z M. 2021. Automatic rocks segmentation based on deep learning for planetary rover images. Journal of Aerospace Information Systems, 18(11): 755-761
  17. 17.
    Lin T Y, Goyal P, Girshick R, He K M and Dollár P. 2020. Focal loss for dense object detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 42(2): 318-327
  18. 18.
    Liu H Q, Yao M B, Xiao X M and Xiong Y G. 2023. RockFormer: a U-shaped transformer network for Martian rock segmentation. IEEE Transactions on Geoscience and Remote Sensing, 61: 4600116
  19. 19.
    Liu Z, Lin Y T, Cao Y, Hu H, Wei Y X, Zhang Z, Lin S and Guo B N. 2021. Swin transformer: hierarchical vision transformer using shifted windows. arXiv: 2103.14030
  20. 20.
    Liu Z Q, Wan W H, Peng M, Zhao Q, Xu B, Liu B, Liu Y L, Di K C, Li L C, Yu T Y, Wang B F, Zhou J L and Chen H M. 2014. Remote sensing mapping and localization techniques for teleoperation of Chang’e-3 rover. Journal of Remote Sensing (in Chinese), 18(5): 971-980
  21. 21.
    Long L H, Zhu Y T, Yan J W, Liu J J and Wang Z Y. 2023. New building extraction method based on semantic segmentation. National Remote Sensing Bulletin, 27(11): 2593-2602
  22. 22.
    Lv W K, Wei L H, Zheng D, Liu Y and Wang Y M. 2023a. MarsNet: automated rock segmentation with transformers for Tianwen-1 mission. IEEE Geoscience and Remote Sensing Letters, 20: 3506605
  23. 23.
    Lv Z Y, Huang H T, Sun W W, Lei T, Benediktsson J A and Li J H. 2023b. Novel enhanced UNet for change detection using multimodal remote sensing image. IEEE Geoscience and Remote Sensing Letters, 20: 2505405
  24. 24.
    Ma C H, Li Y, Lv J Y, Xiao Z X, Zhang W M and Mo L S. 2024. Automated rock detection from mars rover image via Y-shaped dual-task network with depth-aware spatial attention mechanism. IEEE Transactions on Geoscience and Remote Sensing, 62: 4600418
  25. 25.
    Mao A Q, Mohri M and Zhong Y T. 2023. Cross-entropy loss functions: theoretical analysis and applications. arXiv:2304.07288
  26. 26.
    Min L T, Fan Z M, Xie X X and Lü Q Y. 2023. Feature reassembly and self-attention for oriented object detection in remote sensing images. National Remote Sensing Bulletin, 27(12): 2716-2725
  27. 27.
    Morid M A, Borjali A and Del Fiol G. 2021. A scoping review of transfer learning research on medical image analysis using ImageNet. Computers in Biology and Medicine, 128: 104115
  28. 28.
    Nair V and Hinton G E. 2010. Rectified linear units improve restricted Boltzmann machines//Proceedings of the 27th International Conference on Machine Learning. Haifa: Omnipress: 807-814
  29. 29.
    Ogohara K and Gichu R. 2022. Automated segmentation of textured dust storms on mars remote sensing images using an encoder-decoder type convolutional neural network. Computers and Geosciences, 160: 105043
  30. 30.
    Qian S, Chen Z H, Lin M Q and Zhang C B. 2015. Saliency detection based on conditional random field and image segmentation. Acta Automatica Sinica, 41(4): 711-724
  31. 31.
    Ronneberger O, Fischer P and Brox T. 2015. U-Net: convolutional networks for biomedical image segmentation. arXiv: 1505.04597
  32. 32.
    Rozenberszki D, Litany O and Dai A. 2024. UnScene3D: unsupervised 3D instance segmentation for indoor scenes//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle: IEEE: 19957-19967
  33. 33.
    Shelhamer E, Long J and Darrell T. 2017. Fully convolutional networks for semantic segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(4): 640-651
  34. 34.
    Shen Y, Pan D, Cao H T, Yuan B F, Jia Y, He L B and Zou M. 2025. Terrain classification and rock abundance analysis at Utopia Planitia using Zhurong image data based on deep learning algorithms. Journal of Terramechanics, 117: 101022
  35. 35.
    Simonyan K and Zisserman A. 2015. Very deep convolutional networks for large-scale image recognition. arXiv:1409.1556
  36. 36.
    Song B G, Shao P, Shao W, Zhang X D and Dong T. 2025. A two-branch building extraction network integrating main body and edge separation and multi-scale information extraction. National Remote Sensing Bulletin, 29(8): 1-13
  37. 37.
    Sudre C H, Li W Q, Vercauteren T, Ourselin S and Cardoso M J. 2017. Generalised Dice overlap as a deep learning loss function for highly unbalanced segmentations//Proceedings of the 3rd International Workshop, DLMIA 2017, and 7th International Workshop on Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support. Québec City: Springer: 240-248
  38. 38.
    Vaswani A, Shazeer A, Parmar N, Uszkoreit J, Jones L, Gomez A N, Kaiser Ł and Polosukhin I. 2017. Attention is all you need//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach: Curran Associates Inc.: 6000-6010
  39. 39.
    Xie E Z, Wang W H, Yu Z D, Anandkumar A, Alvarez J M and Luo P. 2021. SegFormer: simple and efficient design for semantic segmentation with transformers//Proceedings of the 35th International Conference on Neural Information Processing Systems. Virtual Event: Curran Associates Inc.: 12077-12090
  40. 40.
    Xie J, Yan K, Kang Z Z, Xu X J, Xue B, Yang J F and Tao J Y. 2021. Verification study for the mineral and rock identification using multispectral camera of the “Zhurong” Mars Rover on the earth. National Remote Sensing Bulletin, 25(7): 1385-1399
  41. 41.
    Xiong Y G, Xiao X M, Yao M B, Liu H Q, Yang H and Fu Y G. 2023. MarsFormer: Martian rock semantic segmentation with transformer. IEEE Transactions on Geoscience and Remote Sensing, 61: 4600612
  42. 42.
    Yang C, Zhao H S, Bruzzone L, Benediktsson J A, Liang Y C, Liu B, Zeng X G, Guan R C, Li C L and Ouyang Z Y. 2020. Lunar impact crater identification and age estimation with Chang’E data by deep and transfer learning. Nature Communications, 11(1): 6358
  43. 43.
    Yi Q, Zhang Y H, Zong Y T and Dai Y B. 2023. A survey of hyperspectral image classification algorithms based on convolutional neural networks. Electronics Optics and Control, 30(3): 70-77
  44. 44.
    Yu C Q, Wang J B, Gao C X, Yu G, Shen C H and Sang N. 2020. Context prior for scene segmentation//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle: IEEE: 12413-12422
  45. 45.
    Yuan Y H, Chen X L and Wang J D. 2020. Object-contextual representations for semantic segmentation//Vedaldi A, Bischof H, Brox T and Frahm J M, eds. 16th European Conference Computer Vision-ECCV 2020. Glasgow: Springer: 173-190
  46. 46.
    Zhao H S, Shi J P, Qi X J, Wang X G and Jia J Y. 2017. Pyramid scene parsing network. arXiv:1612.01105
  47. 47.
    Zheng S X, Lu J C, Zhao H S, Zhu X T, Luo Z K, Wang Y B, Fu Y W, Feng J F, Xiang T, Torr P H S and Zhang L. 2021. Rethinking semantic segmentation from a sequence-to-sequence perspective with transformers//Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Nashville: IEEE: 6877-6886
  48. 48.
    Zheng W, Kang G W, Chen W F and Li X W. 2008. Unsupervised segmentation of remote sensing images based on fuzzy Markov random field model. Journal of Remote Sensing (in Chinese), 12(2): 246-252
  49. 49.
    Zheng Z, Zhong Y F, Wang J J and Ma A L. 2020. Foreground-aware relation network for geospatial object segmentation in high spatial resolution remote sensing imagery//Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle: IEEE: 4095-4104

더 읽기

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