DOI QR코드

DOI QR Code

딥 러닝 기반 드론 영상을 활용한 벼 포장의 재식거리 평가

Evaluation of Planting Distance in Rice Paddies Using Deep Learning-Based Drone Imagery

  • 박혁진 (국립식량과학원 ) ;
  • 권동원 (국립식량과학원 ) ;
  • 임우진 (국립식량과학원 ) ;
  • 이지현 (국립식량과학원 ) ;
  • 김은지 (국립식량과학원 ) ;
  • 정남진 (전북대학교 작물생명과학과 ) ;
  • 조정일 (국립식량과학원 ) ;
  • 황운하 (국립식량과학원 ) ;
  • 장재기 (국립식량과학원 ) ;
  • 상완규 (국립식량과학원 )
  • Hyeok-jin Bak (National Institute of Crop Science, Rural Development Administration, RDA) ;
  • Dongwon Kwon (National Institute of Crop Science, Rural Development Administration, RDA) ;
  • Woo-jin Im (National Institute of Crop Science, Rural Development Administration, RDA) ;
  • Ji-hyeon Lee (National Institute of Crop Science, Rural Development Administration, RDA) ;
  • Eun-ji Kim (National Institute of Crop Science, Rural Development Administration, RDA) ;
  • Nam-jin Chung (Jeonbuk National University, Department of Crop Science and Biotechnology) ;
  • Jung-Il Cho (National Institute of Crop Science, Rural Development Administration, RDA) ;
  • Woon-Ha Hwang (National Institute of Crop Science, Rural Development Administration, RDA) ;
  • Jae-Ki Chnag (National Institute of Crop Science, Rural Development Administration, RDA) ;
  • Wan-Gyu Sang (National Institute of Crop Science, Rural Development Administration, RDA)
  • 투고 : 2024.07.31
  • 심사 : 2024.08.21
  • 발행 : 2024.09.01

초록

본 연구는 농업 분야에서의 드론으로 촬영되어진 영상을 딥 러닝 기반의 영상분석 기술을 사용하여 벼 재식거리를 추정하는 알고리즘을 작성하였다. 다양한 포장에서 수집된 영상데이터를 전처리 과정을 거처 YOLOv5x 딥 러닝 모델의 학습데이터로 활용하였다. 영상 데이터를 학습시킨 결과, 높은 정밀도와 재현율을 보여 각 영상에서 벼의 위치 정보를 효과적으로 추정할 수 있음을 확인하였고, 학습된 모델은 다양한 논 포장의 환경에서 벼의 중심부 좌표를 기반으로 벼의 위치를 정확하게 추정할 수 있었다. 이를 통해 각 포장의 벼의 주수를 추정하였고 실제 값과 유사한 값을 얻을 수 있었다. 또한 영상의 벼의 위치정보를 기반으로 하여 재식거리를 정확히 파악할 수 있는 새로운 알고리즘을 제시하였다. 포장의 일부를 촬영한 다양한 지역의 드론 영상에서 본 연구에서 작성한 알고리즘이 실제 재식거리와 R2=0.877의 높은 상관성을 확인하였다. 이는 본 연구에서 작성한 알고리즘이 실제 농업 현장에서 효과적으로 적용될 수 있다는 가능성을 시사한다.

In response to the increasing impact of climate change on agriculture, various cultivation technologies have been recently developed to improve agricultural productivity and reduce carbon emissions for carbon neutrality. This study presents an algorithm for estimating rice planting density in agriculture using drone-captured images and deep learning-based image analysis technology. The algorithm utilizes images collected from various paddies; these images are processed through pre-processing steps and serve as training data for the YOLOv5x deep learning model. The trained model demonstrated high precision and recall, effectively estimating the position information of rice plants in each image. By accurately estimating the position of rice plants based on the central coordinates in diverse unpaved environments, the model allowed for estimation of rice plant density in each paddy, producing values closely aligned with actual measurements. Moreover, the algorithm proposed in this study provides a novel approach for precise determination of rice planting density based on the position information of rice plants in the images. Analysis of drone footage from different regions capturing portions of paddies revealed that the developed algorithm exhibited a significant correlation (R2 =0.877) with actual planting density. This finding suggests the potential effective application of the algorithm in real-world agricultural settings. In conclusion, we believe that this research contributes to the ongoing digital transformation in agriculture by offering a valuable technology that supports the goals of enhancing efficiency, mitigating methane emissions, and achieving carbon neutrality, in response to the challenges posed by climate change.

키워드

과제정보

본 연구는 2024년도 농촌진흥청 국립식량과학원 전문연구원 과정 지원사업(RS-2022-RD010389)에 의해 이루어진 것임.

참고문헌

  1. Alzubaidi, L., J. Zhang, A. J. Humaidi, A. Al-Dujaili, Y. Duan, O. Al-Shamma, J. Santamaria, M. A. Fadhel, M. Al-Amidie, and L. Farhan. 2021. Review of deep learning: concepts, CNN architectures, challenges, applications, future directions. Journal of Big Data 8 : 53. https://doi.org/10.1186/s40537-021-00444-8
  2. Bak, H. J., D. W. Kwon, W. G. Sang, H. Y. Ban, S. Y. Chang, J. K. Beak, Y. H. Lee, W. J. Im, M. C. Seo and J. I. Cho. 2023. Deep learning approaches for accurate weed Area assessment in maize fields. Korean Journal of Agricultural and Forest Meteorology 25(1) : 17-27.
  3. Ban, H. Y., J. K. Beak, W. G. Sang, J. H. Kim, and M. C. Seo. 2021. Estimation of the lodging area in rice using deep learning. Korean J. Crop Sci. 66(2) : 105-111.
  4. Bouguettaya, A., H. Zarzour, A. Kechida, and A. M. Taberkit. 2022. Deep learning techniques to classify agricultural crops through UAVimagery: a review. Neural Computing and Applications 34 : 9511-9536. https://doi.org/10.1007/s00521-022-07104-9
  5. Dobosz, B., D. Gozdowski, J. Koronczok, J. Zukovskis, and E. Wojcik-Gront. 2023. Evaluation of maize crop damage using UAV-Based RGB and multispectral imagery. Agriculture 13(8) : 1627. https://doi.org/10.3390/agriculture13081627
  6. Hama, A., K. Tanaka, B. Chen, and A. Kondoh. 2021. Examination of appropriate observation time and correction of vegetation index for drone-based crop monitoring. Journal of Agricultural Meteorology 77(3) : 200-209. https://doi.org/10.2480/agrmet.D-20-00047
  7. Hayashi, S., A. Kamoshita, and J. Yamagishi. 2006. Effect of planting density on grain yield and water productivity of rice grown in flooded and non-flooded fields in Japan. Journal of Plant Prod. Sci. 9(3) : 298-311.
  8. Hwang, W. H., H. S. Lee, S. Y. Yang, and C. G. Lee. 2023. Change in yield characteristics by transplanting density in major cultivated rice. Korean J. Crop Sci. 68(1) : 1-7
  9. Kanetaka, M., A. Takahashi, and S. Azuma. 2004. Transplanting culture by dense sowing and sparse planting of Koshihikari. The Hor. Crop Sci. 40 : 11-14.
  10. Kazemi, F. and E. Parmehr. 2022. Evaluation of rgb vegetation indices derived from uav images for rice crop growth monitoring. ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences X-4/W1-2022 : 385-390. https://doi.org/10.5194/isprs-annals-X-4-W1-2022-385-2023
  11. Kim, G. Y., S. B. Lee, J. S. Lee, E. J. Choi, and J. H. Ryu. 2012. Mitigation of greenhouse gases by water management of SRI (System of Rice Intensification) in rice paddy fields. Korean J. Soil Sci. 45(6) : 1173-1178.
  12. Lee, Y. H., H. J. Bak, J. W. Kim, B. K. Kim, H. S. Jang, D. W. Kim, J. T. Youn, and T. Y. Hwang. 2022. Evaluation of leaf area, canopy coverage light interception, and exposure setting in wheat using RGB color vegetation indices. Journal of Agriculture & Life Science 56(6) : 1-13.
  13. Li, B., X. Xu, J. W. Han, L. Zhang, C. Bian, L. Jin, and J. Liu. 2019. The estimation of crop emergence in potatoes by UAV RGB imagery. Plant Methods 15 : 15. https://doi.org/10.1186/s13007-019-0399-7
  14. Norman, U. 2008. The System of Rice Intensification (SRI) as a system of agricultural innovation. Jurnal Tanah dan Lingkungan 10(1) : 27-40
  15. Redmon, J., S. Divvala, R. Girshick and A. Farhadi. 2016. You only look once: Unified, Real-Time Object Detection. arXiv:1506.02640v5.
  16. Shen, J., Q. Wang, M. Zhao, J. Hu, J. Wang, M. Shu, Y. Liu, W. Guo, H. Qiao, Q. Niu and J. Yue. 2024. Mapping maize planting densities using unmanned aerial vehicles, multispectral remote sensing, and deep learning technology. Drones 8, 140. https://doi.org/10.3390/drones8040140
  17. Shorten, C. and T. M. Khoshgoftaar. 2019. A survey on image data augmentation for deep learning. Journal of Big Data. 6:60 https://doi.org/10.1186/s40537-019-0197-0
  18. Ultralytics. 2020. YOLOv5: A family of object detection architectures and models. Ultralytics GitHub Repository. Available at: https://github.com/ultralytics/yolov5
  19. Yang, S. Y., W. H, Hwang, J. H. Jeong, H. S. Lee, and C. G. Lee. 2021. Changes in growth and yield of different rice varieties under different planting densities in low-density transplanting cultivation. Korean J. Crop Sci. 66(4) : 279-288.