Satellite Ponding Water Detection with Flow-Spectral Coupling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing satellite imagery techniques struggle to precisely detect ponding water within agricultural fields due to mixed spectral responses and atmospheric corrections, making it challenging to differentiate water signals from other land cover signals, especially in high-resolution images.

Innovation Solution

A computer-implemented model using spectral analysis, flow simulation, and coupling logic to analyze satellite imagery, precipitation, elevation, and soil data to estimate ponding water probability, incorporating a cropland mask and Markov Random Field to enhance accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If simple thresholding using infrared bands is used for water detection, then the detection process is simple and fast, but the precision of detecting ponding water is poor due to mixed spectral responses

Engineering Contradiction:
Improvedetection speedVSAvoidponding water detection precision
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent combines multiple spectral indices (NDWI, MNDWI, NDVI) with auxiliary variables (elevation, soil properties, precipitation) into an integrated machine learning model. This merging of multiple detection approaches allows the system to maintain simplicity in implementation while achieving high precision in ponding water detection by leveraging the complementary strengths of different indices and environmental factors.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent creates a composite detection model that integrates multiple spectral indices and environmental variables. Similar to how composite materials combine different substances to achieve superior properties, this composite model combines multiple detection approaches to achieve high precision ponding water detection while maintaining operational simplicity through automated processing.

Inventive Principle:
Principle #40Composite materials

2Measurement precision

If multispectral indices and auxiliary variables are combined for water detection, then the detection precision is improved, but the device complexity increases

Engineering Contradiction:
Improveponding water detection precisionVSAvoiddetection system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The machine learning model performs self-service by automatically learning the optimal relationships between multiple spectral indices, auxiliary variables, and ponding water presence from training data. This eliminates the need for manual configuration of complex detection parameters and reduces system complexity while maintaining high precision detection capabilities.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent transforms the complex multi-parameter detection problem into a unified probability output by the machine learning model. Instead of manually managing multiple indices and variables, the model internally processes these parameters and outputs a single ponding water probability value, simplifying the detection system while preserving precision.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If high-resolution satellite imagery is used for ponding water detection, then the spatial detail is improved, but the difficulty of detecting and measuring water signals increases due to mixed pixels

Engineering Contradiction:
Improvespatial detection precisionVSAvoidwater signal differentiation difficulty
Core Design Contradiction:
Measurement precisionVSDifficulty of detecting and measuring

Solution Approach 1:

The patent segments the detection problem by analyzing multiple spectral bands and auxiliary variables separately, then integrating their results through a machine learning model. This segmentation approach allows the system to handle mixed pixels at high resolution by evaluating multiple independent indicators rather than relying on a single water signal, thereby reducing detection difficulty while maintaining spatial precision.

Inventive Principle:
Principle #1Segmentation

4Measurement precision

If physical field traversal is used to detect ponding water, then the detection accuracy is high, but the productivity and coverage area are limited

Engineering Contradiction:
Improveponding water detection accuracyVSAvoidfield coverage efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent replaces the mechanical system of physical field traversal with a remote sensing-based machine learning model. Satellite imagery and environmental data are processed computationally to detect ponding water across large areas simultaneously, maintaining high detection accuracy while dramatically improving productivity and coverage efficiency compared to manual field inspection.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

The model effectively pinpoints areas likely to contain ponding water, enabling farmers to efficiently locate and drain them, thereby preserving crop yield by addressing the issue of overwatering.

Implementation Method 1

These spectral approaches typically rely on the fact that water strongly absorbs incoming radiation in the near to mid-infrared wavelengths

Methodology Applied
Scientific EffectAbsorption (EM radiation): Absorption (EM radiation)

Data Source

PatentEP3353715B1Ponding water detection on satellite imagery
Publication Date: 2026.01.21 MONSANTO TECHNOLOGY LLC
  • EP3353715B1 patent drawingFigure 1
  • EP3353715B1 patent drawingFigure 2(a)~2(b)
  • EP3353715B1 patent drawingFigure 3

AI summary

In an approach, an image of an agricultural field is analyzed using a classifier that has been trained to identify a probability for each pixel within an image that the pixel corresponds to water. A flow simulation is performed to determine regions of the field that are likely to pool water after rainfall based on precipitation data, elevation data, and soil property data of the field. A graph of vertices representing the pixels and edges representing connections between neighboring pixels is generated. The probability of each pixel within the graph being ponding water is set based on the probability pixel being water, the likelihood that water will pool in the area represented by the pixel, the probability of neighboring pixels being ponding water, and a cropland mask that identifies which pixels correspond to cropland. A class for each pixel is then determined that maximizes the joint probability over the graph.