Remote sensing-based early disease detection for precision fungicide application in soybean

Overview

This research develops a drone-based remote sensing system that detects early-stage foliar disease infection in soybean crops before visible symptoms emerge. By capturing subtle changes in leaf spectral reflectance using multispectral and thermal cameras, the technology identifies disease indicators that traditional visual scouting methods miss. When paired with varietal information, the system supports more precise fungicide application decisions, helping growers protect yield while reducing unnecessary chemical inputs.

Current disease management relies on visual scouting, which detects disease only after symptoms appear—often too late for fungicides to be fully effective. This indiscriminate approach can lead to yield loss, wasted product, and increased environmental impact. By detecting spectral changes associated with early infection, this approach offers a proactive alternative that improves both economic and agronomic outcomes for soybean producers.

Technical specifications

Core approach:

• Uses drone-mounted multispectral and thermal cameras to capture leaf reflectance data from soybean fields
• Analyzes spectral changes such as those measured by indices like GNDVI (Green Normalized Difference Vegetation Index) to identify early disease indicators
• Incorporates varietal information to refine disease detection and management recommendations
• Builds on prior validation showing that hyperspectral analysis can correlate reduced reflectance with increasing soybean rust severity
• Leverages demonstrated success of thermal imagery in diagnosing infections such as Oidium neolycopersicum in tomato, Plasmopora viticola in grapes, and Pseudoperonospora cubensis in cucumber

Validation methodology:

• Greenhouse trials will measure leaf reflectance of plants from three soybean varieties inoculated with the frogeye leaf spot pathogen, starting one day before inoculation and repeating every two days
• Field trials will assess reflectance across three varieties and five fungicide treatments at two locations, with spectral camera flights conducted every two weeks
• Disease severity and yield data will be recorded to correlate spectral signals with agronomic outcomes

Technology readiness level

This research is at an early-to-mid stage of development. Prior work has established proof of concept for using spectral reflectance to detect soybean rust and for using thermal imagery to diagnose other foliar diseases. The proposed greenhouse and field validation studies will test the approach specifically for frogeye leaf spot and multiple foliar diseases in soybean across multiple varieties and management treatments. Successful validation would position the technology for further development toward grower-ready decision support tools for precision fungicide management.