Wearable Crop Identification Database Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Wearable terminal display systems require relearning when new crop relationships are encountered, leading to prolonged updating times.
Innovation Solution
A data structure for machine learning is used to build a database with training data including image data and meta-IDs linked to reference information, allowing for efficient selection and updating of reference information without rebuilding the database.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If the wearable terminal display system uses machine learning to identify crop types from images, then the system can automatically determine harvest time, but the system requires extensive retraining when new crop relationships are encountered, leading to prolonged updating times
Solution Approach 1:
The patent segments the machine learning model into two independent components: a crop identification model that recognizes crop types from images, and a harvest time prediction model that determines harvest timing based on crop type and growth conditions. This segmentation allows the identification model to remain fixed while only the prediction model needs updating when new crop relationships are encountered, significantly reducing retraining time
Solution Approach 2:
The patent performs preliminary action by pre-training the crop identification model on a comprehensive dataset of various crop types and growth conditions before deployment. This pre-training establishes a robust foundation that can accurately identify crops without requiring retraining when new prediction relationships are added, separating the identification learning phase from the prediction updating phase
2Reliability
If the system rebuilds the entire machine learning database when new relationships are acquired, then the system maintains accuracy, but the updating process becomes time-consuming
Solution Approach 1:
The patent extracts the harvest time prediction logic from the crop identification process, creating a separate prediction module that receives identification results as input. This extraction allows the identification model to remain unchanged and reliable, while only the prediction module needs to be updated with new crop relationships, maintaining accuracy while improving updating efficiency
Solution Approach 2:
The patent introduces dynamics by making the prediction model adaptable and modifiable without affecting the stable identification model. The prediction model can be dynamically updated with new crop relationships, growth conditions, and harvest time data, allowing the system to evolve while maintaining the reliability of the core identification function
Data Source
AI summary
An information providing system includes a content association database storing external information, in which content is divided in a chunk structure, and chunk reference information, in which reference information having a different attribute than the content is divided in a chunk structure, in association with each other. An external information similarity calculation database is built on machine learning using the external information. A processor acquires specific external information, and compares the external information with the acquired specific external information. The processor looks up the external information similarity calculation database, and calculates a similarity between the external information and the specific external information, when the external information does not match with the acquired specific external information. The processor selects first external information from the stored external information, based on the calculated similarity, and extracts chunk reference information associated with the first external information as first chunk reference information.