Title:
METHOD AND SYSTEM FOR GENERATING ROOT MORPHOLOGICAL DATA-BASED ARTIFICIAL INTELLIGENCE MODEL
Document Type and Number:
WIPO Patent Application WO/2023/022433
Kind Code:
A1
Abstract:
The present invention relates to a method for generating a root morphological data-based artificial intelligence model. The method for generating a root morphology data-based artificial intelligence model may comprise the steps of: acquiring a root shape image that changes by decision making of a previously trained learning model, and learning the correlation between the decision making of the learning model and root morphology in consideration of the root shape image; generating growth and cultivation information by linking cultivation image tag information related to a cultivation process of a plant and provided from a user terminal with an image of the plant, and learning a current state of the plant and a cultivation target state of the plant in consideration of the growth and cultivation information; generating an educational algorithm for learning a root image change relationship in consideration of the learned correlation and the growth and cultivation information; and constructing a cultivation learning model being with reference to the generated education algorithm while using a plant image of an aboveground part and a root shape image, obtained from an image input device, as an input, and a control signal for controlling a plant cultivation system as an output.
Inventors:
PARK HAE YOUNG (KR)
Application Number:
PCT/KR2022/011982
Publication Date:
February 23, 2023
Filing Date:
August 11, 2022
Export Citation:
Assignee:
MIDBAR CO LTD (KR)
International Classes:
G06N20/00; G01B11/24; G06Q50/02
Foreign References:
| KR20210035948A | 2021-04-02 | |||
| JP2019054779A | 2019-04-11 | |||
| KR20200029657A | 2020-03-19 | |||
| KR101887503B1 | 2018-08-10 | |||
| KR20210096439A | 2021-08-05 |
Attorney, Agent or Firm:
AJU INTERNATIONAL LAW & PATENT GROUP (KR)
Download PDF: