Inteligencia artificial y visión por ordenador para evaluar los componentes del rendimiento de la vid en viñedos comerciales

  1. Íñiguez, Rubén 12
  2. Poblete-Echeverría, Carlos 12
  3. Hernández, Inés 12
  4. Gutiérrez, Salvador 3
  5. Barrio, Ignacio 12
  6. Tardáguila, Javier 12
  1. 1 Universidad de La Rioja
    info

    Universidad de La Rioja

    Logroño, España

    ROR https://ror.org/0553yr311

  2. 2 Instituto de Ciencias de la Vid y del Vino
    info

    Instituto de Ciencias de la Vid y del Vino

    Logroño, España

    ROR https://ror.org/01rm2sw78

  3. 3 Universidad de Granada
    info

    Universidad de Granada

    Granada, España

    ROR https://ror.org/04njjy449

Show affiliations +
Proceedings:
BIO Web of Conferences: 44th World Congress of Vine and Wine
  1. Pau Roca (ed. lit.)

Publisher: Web of Conferences

ISSN: 2117-4458

Year of publication: 2023

Volume: 68

Congress: 44th World Congress of Vine and Wine (44º. 2023. Cádiz/Jerez)

Type: Conference paper

DOI: 10.1051/BIOCONF/20236801023 GOOGLE SCHOLAR lock_openOpen access editor

Abstract

Yield estimation is very important for the wine industry since provides useful information for vineyard and winery management. Climate change effects such as higher temperatures and lower water availability can affect vineyard yield components. In general, traditional yield forecasts are based on destructive manual counting of bunches and berry weight. These conventional methods do not provide accurate estimations and are time-consuming, expensive, and labour-intensive. In this study, novel methods using digital technologies such as computer vision and artificial intelligence were used to estimate yield in commercial vineyards. Computer vision was used for the automatic detection of different canopy features and the calibration of regression equations for the prediction of yield per vine. Artificial intelligence was used for the automatic counting of bunches. The results showed that the deep learning algorithm was able to detect bunches with hight accuracy. Our results demonstrate the applicability of these novel methods to assess yield components in commercial vineyards.

Bibliographic References

  • Martin S., Dunstone R., Dunn G.. GWRDC 100 (2003)
  • Mohimont, (2022), Agronomy, 12, pp. 2463, 10.3390/agronomy12102463
  • Liu, (2017), Computers and Electronics in Agriculture, 137, pp. 88, 10.1016/j.compag.2017.03.013
  • Palacios, (2020), Computers and Electronics in Agriculture, 178, pp. 105796, 10.1016/j.compag.2020.105796
  • Palacios, (2022), Biosystems Engineering, 218, pp. 175, 10.1016/j.biosystemseng.2022.04.015
  • Xin, (2020), Australian Journal of Grape and Wine Research, 26, pp. 207, 10.1111/ajgw.12444
  • Hacking, (2020), OENO One, 54, pp. 793, 10.20870/oeno-one.2020.54.4.3361
  • Xiong, (2018), Sensors, 18, pp. 969, 10.3390/s18040969
  • Íñiguez, (2021), J. Tardáguila. Agronomy, 11, pp. 1003, 10.3390/agronomy11051003
  • Fuentes A., Yoon S., Kim S.C., Park E.D.S.. Sensors 17 (2017)
  • Sozzi, (2022), Agronomy, 12, pp. 319, 10.3390/agronomy12020319
  • Tzutalin D.. LabelImg. GitHub Repository 6 (2015)
  • Bochkovskiy A., Wang C. Y., Liao H. Y. M.. arXiv preprint arXiv:2004.10934 (2020)
  • Redmon J.. Darknet: Open source neural networks in C (2013)
  • Bochkovskiy A.. Yolo v4 repository [source code]. https://github.com/AlexeyAB/darknet (2020)