Spectral reflectance sensing for potato crop nutrient and health monitoring

Overview

This research develops a non-destructive method for assessing crop health and nutrient intake in potato plants by analyzing the wavelength spectral reflectance of leaves. By correlating spectral data collected via spectrophotometer with laboratory chemical analysis of petioles, the approach identifies specific wavelengths associated with key nutrients such as nitrogen, phosphorus, and calcium. The goal is to enable rapid, field-deployable nutrient diagnostics that support precision agriculture and reduce reliance on time-consuming laboratory testing.

Technical specifications

Data collection approach:

• Petioles and corresponding tip leaves are collected weekly from potato fields, with 40 samples per data point
• Laboratory chemical analysis of petioles yields concentrations for 17 macro and micro nutrients
• Spectral reflectance data is captured across the 400–2500 nm range on both dry and wet leaves to account for water content effects
• Preliminary analysis using R software identified strong correlations between nitrogen and 668 nm, phosphorus and 1140 nm, and calcium and 724 nm
• The 700–1400 nm range shows particular promise for nutrient-specific associations

Planned enhancements:

• Expansion to 200 data points per nutrient in the next growing season
• Controlled-environment growth of potato plants with induced deficiencies (starting with reduced phosphorus) to improve model robustness and avoid overfitting
• Migration from point spectrophotometry to hyperspectral imaging to increase dataset size and reduce analysis time
• Regression and discriminant analysis to build statistical models estimating nutrient quantities and classifying sufficiency versus deficiency

Technology readiness level

The technology is at an early-to-mid research stage. An initial dataset of 42 data points per nutrient was collected between May and September 2020, with preliminary correlation results confirming wavelength-nutrient relationships for nitrogen, phosphorus, and calcium. The research group is actively expanding datasets and refining controlled-environment validation methods. Dedicated sensor development represents the downstream commercialization goal, contingent upon sufficient model validation through expanded field and controlled trials.