Field-Level Yield Forecasting Using Segmented Aerial Imagery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current crop yield forecasting models are limited in accurately predicting yields at the field level due to the large volume of data required, which is impractical to process and store, and existing remote sensing imagery only estimates yields at county, regional, or state levels, neglecting field-specific variations.
Innovation Solution
The use of aerial image data from satellites, transformed and processed using machine learning models, specifically deep learning algorithms, to generate field-level yield predictions by reducing data complexity and requiring fewer computational resources, allowing for accurate field-level yield mapping and strategic decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If field-level imagery data is used for crop yield forecasting, then measurement precision is improved, but device complexity increases due to large data volume requiring specialized computing devices
Solution Approach 1:
The patent segments the agricultural land into multiple distinct parcels or polygons, processing imagery data for each segment separately. This allows field-level precision yield forecasting by analyzing each parcel's specific characteristics while managing data volume through spatial segmentation rather than treating the entire region as a single unit.
Solution Approach 2:
The patent transforms the problem from processing raw pixel data to processing extracted features and metadata dimensions. By converting imagery into structured data representations (parcels, crops, growth stages, yield estimates), the system reduces computational complexity while maintaining field-level precision through multi-dimensional data organization.
2Productivity
If in-season satellite imagery is processed for yield forecasting, then productivity is improved through timely decisions, but loss of time occurs due to extensive data processing requirements
Solution Approach 1:
The system performs preliminary processing of satellite imagery to extract parcel boundaries, crop types, and growth stage information before final yield forecasting. This preliminary action prepares data structures in advance, enabling faster yield estimation when decisions are needed during the growing season without reprocessing raw imagery from scratch.
Solution Approach 2:
The patent replaces traditional mechanical field assessment methods with automated satellite imagery processing and machine learning models. This substitution enables rapid analysis of large areas without physical field visits, significantly improving productivity while reducing the time loss associated with manual agricultural assessment.
3Device complexity
If county or regional level estimations are used, then device complexity is reduced, but measurement precision deteriorates by neglecting field-specific variations
Solution Approach 1:
The patent applies local quality by identifying and analyzing distinct parcels within regions, each with its own crop type, growth stage, and yield characteristics. This allows the system to maintain simplified regional processing while incorporating local field-specific variations through parcel-level metadata, achieving both reduced complexity and improved precision.
Data Source
AI summary
In an embodiment, digital images of agricultural fields are received at an agricultural intelligence processing system. Each digital image includes a set of pixels having pixel values, and each pixel value of a pixel includes a plurality of spectral band intensity values. Each spectral band intensity value describes a spectral band intensity of one band among several bands of electromagnetic radiation. For each of the agricultural fields, spectral band intensity values of each band are preprocessed at a field level using the digital images for that agricultural field resulting in preprocessed intensity values. The preprocessed intensity values are provided as input to a machine learning model. The model generates a predicted yield value for each field. The predicted yield value is used to update field yield maps of agricultural fields for forecasting and can be displayed via a graphical user interface (GUI) of a client computing device.