Smart Media Journal (스마트미디어저널)

THE KOREAN INSTITUTE OF SMART MEDIA (한국스마트미디어학회)
권호 표지
DOI QR코드

DOI QR Code

Searching Spectrum Band of Crop Area Based on Deep Learning Using Hyper-spectral Image

초분광 영상을 이용한 딥러닝 기반의 작물 영역 스펙트럼 밴드 탐색

  • 이광형 (한국전자기술연구원 IT응용연구센터) ;
  • 명현정 (한국전자기술연구원 IT응용연구센터) ;
  • 디팍 기미레 (한국전자기술연구원 IT응용연구센터) ;
  • 김동훈 (한국전자기술연구원 IT응용연구센터) ;
  • 조세운 (한국전자기술연구원 IT응용연구센터) ;
  • 정성환 (한국전자기술연구원 IT응용연구센터) ;
  • 김병준 (한국전자기술연구원 IT응용연구센터)
  • Received : 2024.07.31
  • Accepted : 2024.08.16
  • Published : 2024.08.30
Download PDF
⟨ Previous Next ⟩

Abstract

Recently, various studies have emerged that utilize hyperspectral imaging for crop growth analysis and early disease diagnosis. However, the challenge of using numerous spectral bands or finding the optimal bands for crop area remains a difficult problem. In this paper, we propose a method of searching the optimized spectral band of crop area based on deep learning using the hyper-spectral image. The proposed method extracts RGB images within hyperspectral images to segment background and foreground area through a Vision Transformer-based Seformer. The segmented results project onto each band of gray-scale converted hyperspectral images. It determines the optimized spectral band of the crop area through the pixel comparison of the projected foreground and background area. The proposed method achieved foreground and background segmentation performance with an average accuracy of 98.47% and a mIoU of 96.48%. In addition, it was confirmed that the proposed method converges to the NIR regions closely related to the crop area compared to the mRMR method.

최근 초분광 영상을 활용하여 작물의 생육 분석 및 질병을 조기에 진단하는 다양한 연구들이 등장하였지만, 수많은 스팩트럼 밴드를 사용하거나 최적의 밴드를 탐색하는 것은 어려운 문제로 남아 있다. 본 논문에서는 초분광 영상을 이용한 딥러닝 기반의 최적화된 작물 영역 스펙트럼 밴드를 탐색하는 방법을 제안한다. 제안한 방법은 초분광 영상 내 RGB 영상을 추출하여 Vision Transformer 기반 Segformer을 통해 배경과 전경 영역을 분할한다. 분할된 결과는 그레이스케일 전환한 초분광 영상 각 밴드에 투영 후 전경과 배경 영역의 평균 픽셀 비교를 통해 작물 영역의 최적화된 스펙트럼 밴드를 탐색한다. 제안된 방법을 통해 전경과 배경 분할 성능은 평균 정확도 98.47%와 mIoU 96.48%의 성능을 나타내었다. 또한, mRMR 방법에 비해 제안 방법이 작물 영역 밀접하게 연관된 NIR 영역에 수렴하는 것을 확인하였다.

Keywords

Acknowledgement

본 논문은 농림축산식품부의 재원으로 농림식품기술기획평가원의 디지털육종전환기술개발사업의 지원을 받아연구되었음(No. 322065-3).

References

  1. Byoungjun Kim, You-Kyoung Han, Jong-Han Park, and Joonwhoan Lee, "Improved Vision-Based Detection of Strawberry Diseases Using a Deep Neural Network," Frontiers in Plant Sciences, vol. 11, Jan. 2021.
  2. Ferentinos, Konstantions P. "Deep learning models for plant disease detection and diagnosis," Computers and Electronics in Agricultures, Vol. 145, pp. 311-318, 2018. https://doi.org/10.1016/j.compag.2018.01.009
  3. 고유진, 이현준, 정희자, 위리, 김남호, "딥러닝 기반 작물 질병 탐지 및 분류 시스템," 스마트미디어저널논문지, 제12권, 제7호, 9-17쪽, 2023년 https://doi.org/10.30693/SMJ.2023.12.7.9
  4. 박준, 심춘보, 김준영, 박성욱, 정세훈, "ResNet 기반 작물 생육단계 추정 모델 개발," 스마트미디어저널논문지, 제11권, 제2호, 53-62쪽, 2022년
  5. 김서정, 김형석, "Multi-Tasking U-Net 기반 파프리카 병해충 진단," 스마트미디어저널논문지, 제9권, 제1호, 16-22쪽, 2020년
  6. Moghadam, P., Ward, D., Goan, E., Jayawardena, S., Sikka, P., and Hernandez, E. "Plant Disease Detection using Hyperspectral Imaging," 2017 International Conference on Digital Image Computing: Techniques and Application, pp. 1-8, 2017.
  7. 최강인, 노혜민, 정성환, 유철중, "초분광 영상과 심층 신경망을 이용한 작물 생육상태 분류 기법 : 파프리카 잎 사례 연구," 한국정보기술학회논문지, 제17권, 제12호, 1-12쪽, 2019년 https://doi.org/10.14801/jkiit.2019.17.12.1
  8. Keunho Park, Young ki Hong, Gook hwan Km, and Joonwhoan Lee, "Classifiation of apple leaf conditions in hyper-spectral images for diagnosis of marssonia blotch using mRMR and deep neural network," Computers and Electronics in Agriculture, vol. 148, pp. 179-187, 2018. https://doi.org/10.1016/j.compag.2018.02.025
  9. Vaiphasa, C., Skidmore, A. K, De Boer, W. F., and Vaiphasa, T, "A hyperspectral band selector for plant species discrimination," ISPRS Journal of Photogrammetry and Remote Sensing, vol. 62, no. 3, pp. 225-235, 2007. https://doi.org/10.1016/j.isprsjprs.2007.05.006
  10. Cordts, M., Omran, M., Ramos, S., Rehfeld, T., Enzweiler, M., Benenson, R., and Schiele, B., "The cityscapes dataset for semantic urban scene understanding," In Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 3213-3223, 2016.
  11. A Dosovitskiy, L. Beyer, A Kolesnikov, D. Weissenbom, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Geliy, J. Uszkoreit, and N. Houlsby, "An image is worth 16x16 words: Transformers for image recognition at scale," arXiv preprint arXiv:2010.11929, Jun. 2020.
  12. E. Xie, W. Wang, Z. Yu, A. Anandkumar, J. M. Alvarez, and P. Luo, "SegFormer: Simple and efficient design for semantic segmentation with transformers," Advances in Neural Information Processing Systems, vol. 34, Oct. 2021.
  13. Giuffrida, M. V., Minervini, M., and Tsaftaris, S., "Learning to count leaves in rosette plants," in Proceedings of the Computer Vision Problems in Plant Phenotyping, pp. 1-13, 2015.
  14. Shruti Jadon, "A survey of loss functions for semantic segmentation," 2020 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB), pp. 1-7, Oct. 2020.
  15. Zhao, Zhenyu, Radhika Anand, and Mallory Wang. "Maximum relevance and minimum redundancy feature selection methods for a marketing machine learning platform," 2019 IEEE international conference on data science and advanced analytics, pp. 442-452, 2019.
  16. Jonathan Long, Evan Shelhamer, and Trevor Darrell, "Fully convolutional networks for semantic segmentation. In Pcroceedings of the IEEE conference on computer vision and pattern recognition, 2015.
  17. Li, Gongyang, Zhi Liu, and Haibin Ling, "ICNet: Information conversion network for RGB-D based salient object detection," IEEE Transactions on Image Processing, vol. 29, pp. 4873-4884, 2020. https://doi.org/10.1109/TIP.2020.2976689
  18. Zhao, H., Shi, J., Qi, X., Wang, X., and, Jia, J., "Pyramid scene parsing network," In Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 2881-2890, 2017.
  19. Changqian Yu, Jingbo Wang, Changxin Gao, Gang Yu, Chunhua Shen, and Nong Sang. "Context prior for scene segmentation," In Pcroceedings of the IEEE conference on computer vision and pattern recognition, 2020.
  20. Ilya Loshchilov, Frank Hutter, "Decoupled weight decay regularization," arXiv preprint arXiv:1711.05101, 2017.
  21. Zhong, T., Zhao, J., Guo, X, Su, Q., & Fox, G., "RINAS: Training with Dataset Shuffling Can Be General and Fast," arXiv preprint arXiv:2312.02368, 2023.