Short answer

Integrate vis-NIR spectroscopy and PLS regression modelling into agricultural equipment design to enable real-time, site-specific soil property assessment and optimise resource application.

Field: Modelling
Source: CERES (Cranfield University) (2015)
Method: Quantitative research using spectral analysis and statistical modelling.
Sample: Approximately 1500 measurement points per hectare were collected across multiple fields over two growing seasons.
Evidence: Strong effect

Visible and Near-Infrared (vis-NIR) spectroscopy, when coupled with Partial Least Squares (PLS) regression, can create reliable models to predict essential soil properties for optimising nitrogen fertiliser application in vegetable crops. This modelling research insight is drawn from a 2015 study published in CERES (Cranfield University). Using Quantitative research using spectral analysis and statistical modelling. with Approximately 1500 measurement points per hectare were collected across multiple fields over two growing seasons., researchers explored how this design variable affects real-world outcomes. The key design takeaway: Integrate vis-NIR spectroscopy and PLS regression modelling into agricultural equipment design to enable real-time, site-specific soil property assessment and optimise resource application.

Study

ModellingHigh ImpactStrong effect

Vis-NIR Spectroscopy Models Accurately Predict Soil Properties for Precision Fertilisation

CERES (Cranfield University) · 2015

Key Findings

01Vis-NIR spectroscopy can be effectively used to predict soil moisture content (MC), soil organic carbon (OC), pH, and total nitrogen (TN).
02Partial Least Squares (PLS) regression is a suitable method for developing accurate calibration models for these soil properties.
03Model accuracy for soil moisture content was best with regional calibration sets (SC2).

Application

Design takeaway

Integrate vis-NIR spectroscopy and PLS regression modelling into agricultural equipment design to enable real-time, site-specific soil property assessment and optimise resource application.

How to apply

Design agricultural machinery with integrated vis-NIR sensors and on-board processing capabilities to analyse soil properties as the machinery operates, adjusting fertiliser application rates in real-time.

Project actions

01When investigating material properties, consider using non-destructive sensing techniques.
02Explore different regression techniques to build predictive models for your chosen properties.

Method & Evidence

AimTo develop reliable calibration models using on-line vis-NIR spectroscopy for predicting soil moisture content, organic carbon, pH, and total nitrogen to refine nitrogen fertilisation management in vegetable crops.

MethodQuantitative research using spectral analysis and statistical modelling.

ProcedureSoil spectra were collected using a mobile vis-NIR spectrophotometer in vegetable crop fields. Calibration models for soil properties (moisture content, organic carbon, pH, total nitrogen) were developed using Partial Least Squares (PLS) regression analysis, testing different calibration set sizes and geographical scales.

SampleApproximately 1500 measurement points per hectare were collected across multiple fields over two growing seasons.

ContextAgricultural technology, precision farming, soil science.

Variables

IVSpectral data (vis-NIR reflectance), calibration set characteristics (size, geographical scale).

DVPredicted soil properties (moisture content, organic carbon, pH, total nitrogen).

CVSoil type, crop type, measurement depth, time of measurement, spectrophotometer settings.

Strengths & Limitations

Strengths

+Demonstrates a practical application of advanced sensing and modelling techniques.
+Provides a robust methodology for developing calibration models for soil properties.

Limitations

The accuracy of spectral analysis can be affected by factors like soil texture, moisture levels, and the presence of surface residues.

Reliability & validity

The study uses cross-validation to assess model performance, which helps ensure the reliability and generalisability of the calibration models. The use of a spectrophotometer and established statistical methods contributes to the validity of the findings.

Think critically

How might the environmental conditions (e.g., sunlight intensity, soil surface condition) during spectral data collection impact the reliability of the developed models in a real-world application?

Design Principles

"Utilise spectral analysis and predictive modelling for non-destructive, real-time assessment of material properties to inform adaptive system behaviour."

This approach offers a significant advancement over traditional soil sampling methods, which are often time-consuming, costly, and less accurate. By enabling rapid, non-destructive, and site-specific soil analysis, designers can develop more efficient and cost-effective agricultural systems that reduce waste and improve crop yields.

What This Means for Your Design

Using light to 'see' what's in the soil and a computer model to guess the amounts of water, carbon, and nitrogen, which helps farmers use the right amount of fertiliser.

How to use in your project

1.Reference this study when discussing the use of spectral analysis or predictive modelling for material characterisation in your design project.

Add to My Project

Quick Cite

Paragraph starter

The research by Jiménez-Donaire (2015) highlights the potential of vis-NIR spectroscopy and PLS regression modelling for accurately predicting soil properties such as moisture content and nutrient levels. This approach offers a non-destructive and timely method for site-specific analysis, which can significantly refine resource management strategies in agricultural design.

Source

CERES (Cranfield University)

On-line measurement of selected soil properties towards the refinement of Nitrogen fertilisation management in vegetable crops

journal · 2015

View source

Questions About This Research

What does the research say about vis-nir spectroscopy models accurately predict soil properties for precision fertilisation?: Integrate vis-NIR spectroscopy and PLS regression modelling into agricultural equipment design to enable real-time, site-specific soil property assessment and optimise resource application. Evidence: CERES (Cranfield University) (2015).
Why does "Vis-NIR Spectroscopy Models Accurately Predict Soil Properties for Precision Fertilisation" matter for design?: This approach offers a significant advancement over traditional soil sampling methods, which are often time-consuming, costly, and less accurate. By enabling rapid, non-destructive, and site-specific soil analysis, designers can develop more efficient and cost-effective agricultural systems that reduce waste and improve crop yields.
How can designers apply this research?: Integrate vis-NIR spectroscopy and PLS regression modelling into agricultural equipment design to enable real-time, site-specific soil property assessment and optimise resource application.
What were the main findings?: Vis-NIR spectroscopy can be effectively used to predict soil moisture content (MC), soil organic carbon (OC), pH, and total nitrogen (TN).. Partial Least Squares (PLS) regression is a suitable method for developing accurate calibration models for these soil properties.. Model accuracy for soil moisture content was best with regional calibration sets (SC2).
What research method was used?: Quantitative research using spectral analysis and statistical modelling. with Approximately 1500 measurement points per hectare were collected across multiple fields over two growing seasons..
How strong is the evidence?: Evidence strength is rated Strong effect, based on a 2015 journal from CERES (Cranfield University).
What should I do differently in my next project?: Design agricultural machinery with integrated vis-NIR sensors and on-board processing capabilities to analyse soil properties as the machinery operates, adjusting fertiliser application rates in real-time.
What are the limitations?: The optimal calibration set size and geographical scale for model accuracy can vary depending on the specific soil property being predicted.