Methods and systems for use in identifying types of crops based on images
The system uses a trained model with 3D convolution and temporal transformer to accurately identify crop types at a pixel level in satellite imagery, improving accuracy and coverage, and facilitating informed crop rotation and regulatory compliance.
WO2026128748A1PCT designated stage Publication Date: 2026-06-18MONSANTO TECHNOLOGY LLC
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- MONSANTO TECHNOLOGY LLC
- Filing Date
- 2025-12-11
- Publication Date
- 2026-06-18
AI Technical Summary
⚠Technical Problem
Existing methods for identifying crop types in agricultural fields are inaccurate, limited in granularity, and often rely on incomplete or misrecorded human data, which hinders effective crop rotation and regulatory compliance.
⚗Method used
A system and method that processes satellite imagery using a trained model with 3D convolution and temporal transformer to identify crop types at a pixel level, improving accuracy and coverage by analyzing images through semantic segmentation and generating crop type maps.
🎯Benefits of technology
Enhances crop type identification accuracy and extends coverage to adjacent fields, eliminating human intervention and enabling informed crop rotation and regulatory compliance.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure US2025059266_18062026_PF_FP_ABST
Abstract
Systems and methods are provided for use in identifying types of crops in agricultural fields based on images of the agricultural fields. One example computer-implemented method includes accessing a plurality of images, which include a time series of images, each of the plurality of images representative of at least one agricultural field or portion thereof, which includes a crop; calculating at least one index for each pixel of each image in the plurality of images, the index based on one or more color bands of the image, and appending the at least one index to the plurality of images, as a data structure; compiling tubelets from the plurality of images, based on the at least one index; and identifying, using a trained model, based on the compiled tubelets, a crop type of the crop in the at least one agricultural field.
Need to check novelty before this filing date? Find Prior Art