Articles
Construction of flower bud diagnosis system using AI image analysis in strawberry cultivation
Article number
1360_29
Pages
229 – 236
Language
English
Abstract
For the implementation of smart agriculture in strawberry cultivation system, we examined the construction of a system for diagnosing the state of flower buds by diagnostic imaging using AI technology.
In this study, the authors examined improvements using AI machine learning and image processing for annotations that have a large workload.
In the improvement of annotation, we could propose a semi-automatic annotation method that combines annotation and the previous learning result.
We were also able to propose a fully automatic annotation method that uses image processing techniques such as image color information and background subtraction.
Furthermore, we were able to propose a method for extracting the flowers that are currently in bloom by using image composition and measurement of overlapping objects.
As a result, by diagnosing the growth state of strawberry cultivation using the estimation result of the number of flowers detected by the method proposed this time, the state of the flowers in each fruit cluster and the flowering time of the flowers in the next fruit cluster, it was found that the proposed system was effective in diagnosing the growth state.
In this study, the authors examined improvements using AI machine learning and image processing for annotations that have a large workload.
In the improvement of annotation, we could propose a semi-automatic annotation method that combines annotation and the previous learning result.
We were also able to propose a fully automatic annotation method that uses image processing techniques such as image color information and background subtraction.
Furthermore, we were able to propose a method for extracting the flowers that are currently in bloom by using image composition and measurement of overlapping objects.
As a result, by diagnosing the growth state of strawberry cultivation using the estimation result of the number of flowers detected by the method proposed this time, the state of the flowers in each fruit cluster and the flowering time of the flowers in the next fruit cluster, it was found that the proposed system was effective in diagnosing the growth state.
Authors
S. Matsuo, Y. Takahashi, C. Tokoro, M. Isozaki
Keywords
AI, number of flowers, Yolo, image processing, growth diagnosis
Groups involved
- Division Precision Horticulture and Engineering
- Division Plant-Environment Interactions in Field Systems
- Division Vine and Berry Fruits
- Division Tropical and Subtropical Fruit and Nuts
- Division Plant Genetic Resources, Breeding and Biotechnology
- Division Temperate Tree Nuts
- Division Temperate Tree Fruits
- Division Vegetables, Roots and Tubers
- Division Sustaining Horticulture in a Changing World
- Working Group Mechanization, Digitization, Sensing and Robotics
Online Articles (45)