Overview
This service involves deploying AI models in authentic agricultural environments to validate their predictions against real-world data. The process includes field deployment where AI models are integrated with drones or other data collection systems. Ground truth data is collected using manual measurements or industry-standard equipment, serving as a benchmark for accuracy evaluation. Models are compared against this data to assess performance and robustness across various conditions. If required, a feedback loop can be established to refine models based on discrepancies found, ensuring continuous improvement and reliable predictions for agricultural decision-making.
More about the service
How can the service help you?
The service helps customers by validating AI models against real-world conditions, ensuring accurate predictions. Before the service, customers may rely on models trained on controlled datasets; after the service, they receive insights into the model's performance in diverse environments, enabling informed adjustments.
How the service will be delivered
The service involves testing the AI model or AI-based system or equipment in real farming conditions. It can be delivered during growing seasons with no specific location restrictions. The duration depends on the customer’s requirements. Customers receive detailed reports on model performance and recommendations for model improvements. Customers must provide initial model configurations and equipment.
Service customisation
The service can be customised by selecting specific crops or environmental conditions for testing.
Sector
Arable farming
Greenhouse
Horticulture
Tree Crops
Viticulture
Test type
Data
Physical system
Software or AI model
Type of service
Collection of test data
Data analysis
Desk assessment
Performance evaluation
Test design
Test execution
Test setup
Country of delivery
Country of delivery:
Service provider(s):
-
Danish Technological Institute (DTI)
| Website
