Ground Truthing and Crop Validation

Value of Ground Truthing

Ground-truth observations connect satellite-derived patterns with agricultural conditions observed on the ground. They help local experts and Oracle assess whether crop interpretation reflects local crop types, cropping practices, and seasonal conditions.

Ground truthing is especially valuable during initial onboarding, when adding a crop or region, after material changes in farming practices, and when users identify an unexpected pattern that requires local validation.

Plan a Representative Collection

Plan observations with local agronomists and field teams so that the collection reflects the range of relevant crops, regions, and growing conditions. Use the agreed sampling approach and record limitations, such as inaccessible areas or atypical conditions, so they can be considered during review.

Record Useful Field Observations

For each observation, record the information needed to understand the field context:

• location and observation date;
• observed crop and, where relevant, variety or cropping pattern;
• crop growth stage and visible condition;
• the observer or responsible organization; and
• supporting notes or photographs when permitted by the agreed process.

Use clear, consistent crop names and review coordinates, dates, and labels before submission. Do not collect or share personal information unless it is necessary, authorized, and handled through the agreed process.

Review and Use the Results

Government agronomists and data stewards should review observations for completeness, plausibility, and consistency with local agricultural knowledge before submission. Oracle and government teams can then use the approved information to validate crop interpretation and identify where additional collection or clarification is needed.

Ground truthing supports validation; it does not make every individual map feature a field-verified observation.