Estimation of Days after Transplanting using an Artificial Intelligence CNN (Convolutional Neural Network) Model in a Closed-type Plant Factory

Youngtaek Baek, Seounggwan Sul, Young-Yeol Cho

doi:10.7235/HORT.20230008

All Issue

2023 Vol.41, Issue 1 Preview Page Next Page

Research Article

Estimation of Days after Transplanting using an Artificial Intelligence CNN (Convolutional Neural Network) Model in a Closed-type Plant Factory 완전제어형 식물공장에서 인공지능 CNN(Convolutional Neural Network)모델을 이용한 상추의 정식 후 일수 예측

28 February 2023. pp. 81-90

PDF XML

Abstract

작물의 재배에서 실시간으로 생육상태를 측정하는 것은 중요하다. 완전제어형 식물공장에서는 정확한 환경관리와 재배식물의 균일한 생육으로 인해 정식 후 일수(DAT)가 생육상태의 지표가 될 수 있다. 본 연구는 폐쇄형 식물공장 시스템에서 인공지능 모델을 이용하여 상추의 DAT를 이미지 데이터를 통해 추정하기 위해 수행되었다. 사용된 종묘는 Nunhems사의 청상추 ‘Thimble’품종으로, DAT별 이미지 데이터를 수집하였고 재배시스템은 NFT 시스템[240×60×200cm(L×W×H)]이었으며 LED 조명 RGB의 비율은 8:1:1로 세팅하였고 12시간 일장으로 광량은 265(±50)µmol·m^-2·s^-1로 조절하였다. 재배 환경은 온도 19–21°C, 상대습도 55–65%, CO₂ 500–800µmol·mol^-1였다. 동일한 환경에서 수행한 3번의 실험을 통해 총 12개체의 ‘Thimble’에서 DAT별 이미지를 촬영하였다. 모든 실험에서 정식 4일, 8일, 12일, 15일, 18일과 22일 후에 중복으로 이미지 데이터를 확보하였다. 수집된 DAT별 이미지 데이터 중 10개체에서 촬영된 총 240장의 이미지 데이터는 학습에 사용하였고, 학습에 사용되지 않은 2개체의 총 48장의 이미지 데이터는 인공지능 모델의 정확도를 검정하기 위해 사용하였다. 파이썬의 텐서플로를 활용하여 생성한 합성곱 신경망(CNN)을 통해 생성된 인공지능 모델은 검정 정확도가 91.7%였으며, 티처블 머신을 통해 생성한 인공지능 모델은 86.7%의 검정 샘플에서 정확한 정식 후 일수를 최대확률로 예측하였다. 본 실험의 경우, 적은 수의 이미지 데이터로 생성한 인공지능 모델이지만, 생육 환경을 정밀하게 조절할 수 있고 재배 식물의 생장 상태가 균일하여 규격화된 이미지 데이터가 생산되는 식물공장의 특성에 의해 통계적으로 유의미한 정확도를 보였다. 데이터가 많아질수록 모델의 정확도가 상승하는 인공신경망의 특성으로 미루어 추후 실험들을 통하여 더 많은 이미지 데이터를 수집한다면 인공지능 모델이 보다 더 정밀해질 것으로 보인다. 이를 통해 식물공장 재배 현장에 인공지능 예측, 판별을 도입하여 재배 식물의 생장을 실시간으로 확인하고 피드백할 수 있는 시스템을 갖출 수 있을 것으로 생각된다.

키워드

인공신경망

딥러닝

이미지 데이터

머신러닝

티처블 머신

It is important to measure the growth status in real time during the cultivation of crops. In a fully controlled plant factory, the number of days after transplanting (DAT) can be an indicator of growth status due to accurate environmental management and the uniform growth of cultivated plants. This study was conducted to estimate the number of DAT of lettuce through image data using an artificial intelligence model in a closed plant factory system. The seedlings used were the 'Thimble' variety of green lettuce from Nunhems, and image data were collected according to the number of DAT. The RGB ratio was set to 8:1:1, and the amount of light was adjusted to 265(±50) μmol·m^-2·s^-1 with a photoperiod of 12 hours. The cultivation environment maintained a temperature of 19–21°C, relative humidity of 55–65%, and CO₂ in the range of 500–800 μmol·mol^-1. Through three experiments conducted in identical environments, images were taken according to the date of growth from 12 specimens of ‘Thimble’ in total. In all three duplicated experiments, images were collected 4th, 8th, 12th, 15th, 18th, and 22nd DAT. Among collected image set, 240 images from ten objects were used for learning dataset and 48 images from two objects were used to test dataset the accuracy of the generated artificial intelligence model. The artificial intelligence model created through the convolutional neural network (CNN) generated with Python's TensorFlow had a test accuracy rate of 91.7%, and the artificial intelligence model created through a teachable machine showed accuracy of 86.7% for the test samples in this experiment. The number of DAT was predicted with the maximum probability. In this research, although the artificial intelligence model was created with a small amount of image data, it showed significantly high accuracy due to the characteristics of the plant factory where standardized image data were produced by precise environmental control and the uniform growth of the cultivated plants. Given the nature of the artificial neural network, which increase the accuracy of the model as more data are inputted, it is expected that the artificial intelligence model will become more precise if more image data can be collected through additional experiments in the future. Through these efforts, it will be possible to have a system capable of checking and offering feedback of the growth of cultivated plants in real time by introducing artificial intelligence prediction and discrimination to the plant factory cultivation site.

Keywords

artificial neural network

deep learning

image data

machine learning

teachable machine

References

Anwar SM, Majid M, Qayyum A, Awais M, Alnowami M, Khan MK (2018) Medical image analysis using convolutional neural networks: a review. J Med Syst 42:1-13. doi:10.1007/s10916-018-1088-1 10.1007/s10916-018-1088-130298337

Carney M, Webster B, Alvarado I, Phillips K, Howell N, Griffith J, Jongejan J, Pitaru A, Chen A (2020) Teachable machine: Approachable Web-based tool for exploring machine learning classification. In Extended abstracts of the 2020 CHI conference on human factors in computing systems, pp 1-8. doi:10.1145/3334480.3382839 10.1145/3334480.3382839

Cha MK, Kim JS, Cho YY (2014) Growth model of common ice plant (Mesembryanthemum crystallinum L.) using expolinear functions in a closed-type plant production system. Hortic Sci Technol 32:493-498. doi:10.7235/hort.2014.14013 10.7235/hort.2014.14013

Choi J (2018) A study on the standardization strategy for building of learning data set for machine learning applications. Journal of Digital Convergence 16:205-212. doi:10.14400/JDC.2018.16.10.205 10.14400/JDC.2018.16.10.205

Hao X, Jia J, Khattak AM, Zhang L, Guo X, Gao W, Wang M (2020) Growing period classification of Gynura bicolor DC using GL-CNN. Comput Electron Agric 174:105497. doi:10.1016/j.compag.2020.105497 10.1016/j.compag.2020.105497

Hinton GE, Srivastava N, Krizhevsky A, Sutskever I, Salakhutdinov RR (2012) Improving neural networks by preventing co-adaptation of feature detectors. arXiv preprint arXiv:1207.0580. doi:10.48550/arXiv.1207.0580 10.48550/arXiv.1207.0580

Hoagland DR, Arnon DI (1950) The water-culture method for growing plants without soil. Circular. California agricultural experiment station, p 347

Iitsuka S, Fujii N, Kokuryo D, Kaihara T, Nakano S (2019) CNN-based growth prediction of field crops for optimizing food supply chain. In: Ameri F, Stecke K, von Cieminski G, Kiritsis D, eds, Advances in Production Management Systems. Production Management for the Factory of the Future. APMS 2019. IFIP Advances in Information and Communication Technology, Vol 566. Springer Cham doi:10.1007/978-3-030-30000-5_20 10.1007/978-3-030-30000-5_20

Kate V, Shukla P (2021) A 3 Tier CNN model with deep discriminative feature extraction for discovering malignant growth in multi-scale histopathology images. Informatics in Medicine Unlocked 24:100616. doi:10.1016/j.imu.2021.100616 10.1016/j.imu.2021.100616

Kingma DP, Ba J (2014) Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980. doi:10.48550/arXiv.1412.6980 10.48550/arXiv.1412.6980

LeCun Y, Bengio Y (1995) Convolutional networks for images, speech, and time-series. In Arbib MA (Ed.), The handbook of brain theory and neural networks MIT Press

Lee JW, Kang WH, Moon T, Hwang I, Kim D, Son JE (2020) Estimating the leaf area index of bell peppers according to growth stage using ray-tracing simulation and a long short-term memory algorithm. Hortic Environ Biotechnol 61:255-265. doi:10.1007/s13580-019-00214-9 10.1007/s13580-019-00214-9

Lee KJ, Kacorri H (2019) Hands holding clues for object recognition in teachable machines. In Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems, pp 1-12. doi:10.1145/3290605.3300566 10.1145/3290605.3300566

Moon TW, Park JY, Son JE (2020) Estimation of Sweet Pepper Crop Fresh Weight with Convolutional Neural Network. J Bio-Env Con 29:381-387. doi:10.12791/KSBEC.2020.29.4.381 10.12791/KSBEC.2020.29.4.381

Nair V, Hinton GE (2010) Rectified linear units improve restricted boltzmann machines. In Proceedings of the 27th International Conference on International Conference on Machine Learning, pp 807-814

Nam DS, Moon T, Lee JW, Son JE (2019) Estimating transpiration rates of hydroponically-grown paprika via an artificial neural network using aerial and root-zone environments and growth factors in greenhouses. Hortic Environ Biotechnol 60:913-923. doi:10.1007/s13580-019-00183-z 10.1007/s13580-019-00183-z

Rawat W, Wang Z (2017) Deep convolutional neural networks for image classification: A comprehensive review. Neural Comput 29:2352-2449. doi:10.1162/neco_a_00990 10.1162/neco_a_0099028599112

Tran TT, Choi JW, Le TTH, Kim JW (2019) A comparative study of deep CNN in forecasting and classifying the macronutrient deficiencies on development of tomato plant. Appl Sci 9:1601. doi:10.3390/app9081601 10.3390/app9081601

Information

Publisher :KOREAN SOCIETY FOR HORTICULTURAL SCIENCE
Publisher(Ko) :원예과학기술지
Journal Title :Horticultural Science and Technology
Journal Title(Ko) :원예과학기술지
Volume : 41
No :1
Pages :81-90
Received Date :2022. 06. 22
Revised Date :2022. 09. 08
Accepted Date : 2022. 09. 30
DOI :https://doi.org/10.7235/HORT.20230008

Horticultural Science and Technology 원예과학기술지 ISSN:1226-8763(Print) 2465-8588(Online)

All Issue