Solvi: Drone Analytics for Field Trials & Plant-Level Insights

Solvi: Drone Analytics for Field Trials & Plant-Level Insights is an analytics platform tailored to agricultural research teams, seed and chemical companies, and agronomists who require rapid, accurate, and standardized trial evaluations. Instead of spending days walking plots with clipboards, a single drone flight captures the entire trial and converts imagery into plant counts, per-plot statistics, and crop health mapswithin minutes.

By replacing subjective estimates with consistent, high-resolution measurements, Solvi enables confident decision-making, streamlines multi-location trial workflows, and expands research capacity without adding labor. The same capabilities also help growers and service providers monitor crop establishment, plant health, and maturity at production scale.

Plant Counts & Plant-Level Analytics (PlantAI™)

Solvi’s PlantAI™ applies machine learning to detect, count, and analyze individual plants from drone imagery across a wide spectrum of crops—from early-stage seedlings in field trials to mature orchard trees. The process is straightforward: upload your imagery, annotate a small sample of plants, and let the AI complete the analysis. In minutes, you receive precise counts along with plant size and health metrics—whether you’re working on small research plots or expansive commercial fields.

Key Highlights:

• Accepts RGB or multispectral imagery captured by any drone.
• Recognizes plants across row crops, vegetables, vines, and trees.
• Outputs plant-level metrics: count, diameter, area, and health indices (e.g., VARI, NDVI).
• Supports multi-crop detection plus gap and row analysis for deeper field insight.
• Exports as PDF, Excel, SHP, or shareable web links.
• “Done For You” option returns results in ~24 hours if you prefer not to train models.

Zonal Statistics & Plot-Level Analysis

Zonal Statistics transforms drone imagery into plot-by-plot insights ideal for research trials. Define plot boundaries automatically or manually (with SHP/KML import), then compute detailed metrics—vegetation indices, plant counts, elevation, and height—for every plot. Whether comparing treatments, varieties, or experimental conditions, outputs are consistent, repeatable, and ready for export.

Key Highlights:

• Automatic or manual boundary creation with SHP/KML import support.
• Instant per-plot calculations: indices, counts, height, and elevation.
• Color-coded visualization for rapid pattern recognition.
• Simple export to SHP, Excel, or PDF, plus shareable web links.
• Optimized for micro-plot research and variety trials.

Plant Health Maps, Imagery Processing & Collaboration

Beyond counting and plot analytics, Solvi provides full imagery processing and crop health mapping. Upload flight datasets to produce georeferenced maps, vegetation indices, elevation layers, and management zones. Whether scouting issues, generating prescriptions, or collaborating with colleagues, analysis is fast and accessible.

Key Highlights:

• Automatically stitches imagery to create orthomosaics from RGB, multispectral, or thermal sensors.
• Calculates vegetation indices (NDVI, NDRE, VARI) and builds management zones.
• Generates elevation surfaces and plant height models.
• Exports to GeoTIFF, JPEG, SHP, Excel, and PDF.
• Enables annotations and sharing via web links—no special viewer software required.

How it Works

Solvi’s workflow moves you from raw drone imagery to actionable outputs in a few steps—no GIS background required. Each stage emphasizes speed and accuracy so you can progress from collection to decisions in the same day. Whether assessing trial plots or large production fields, the process remains consistent.

Step 1: Collect Drone Imagery

Fly any drone and camera setup—RGB, multispectral, or thermal—over your fields or trial plots. Capture overlapping images using standard mapping patterns to ensure complete coverage and quality.

Step 2: Upload & Process

Upload images directly to Solvi to stitch and georeference them into a high-resolution orthomosaic. The platform supports automated processing and integrates with tools like Agisoft Metashape for advanced workflows.

Step 3: Apply Analytics

Choose PlantAI™ for plant counts, Zonal Statistics for plot-level insights, or vegetation index mapping for crop health analysis. Apply multiple tools to the same dataset to gain different perspectives without reflying.

Step 4: Export & Share

Download outputs as PDF, Excel, SHP, or GeoTIFF for research and farm management systems. Share interactive maps via a simple web link to streamline collaboration with teammates and partners.

What customers have to say about Solvi:

Solvi is trusted by leading researchers, seed companies, and agricultural organizations worldwide to bring speed, accuracy, and consistency to field trial assessment. From multi-location breeding programs to university research projects, the platform helps teams replace subjective measurements with standardized, data-rich insights they can rely on.

MAS Seeds, France – Scaling trials without scaling costs
“Before Solvi, evaluating emergence across our trial network was a slow, labor-intensive process. Now, we capture all the data we need in a single drone flight, and results are ready in hours. This has allowed us to increase our trial capacity by around 40% without adding more staff, while ensuring consistent data across 200+ trial sites.”

— Boris Calvet, Drones and Large-scale Phenotyping at MAS Seeds 

Nordic Beet Research, Denmark – Confidence in decisions and data
“With Solvi, we’ve replaced guesswork with quantifiable, repeatable measurements. It’s not just about saving time—it’s about building trust in our decisions and in the integrity of our data. That confidence is invaluable when managing research across multiple locations.”

— William English, Project Manager & Industrial Doctoral Candidate at NBR

Rijk Zwaan, Netherlands: – A new way to work with field trials
“We are impressed that Solvi’s PlantAI™ detection model is robust enough to work on different crop types from seedlings to crops to fruits, with relatively little labeling required. Solvi is very easy for a new user to quickly pick up. And the user-support experience from Solvi is the best by far for all SaaS solutions we have used in the past.”
— Mike Poodt, Team Leader Information Management R&D at Rijk Zwaan

Solvi in Production Agriculture

While Solvi is optimized for trial research, Plant Counts and Health Maps are equally valuable in production settings—supporting scouting for emergence, maturity, plant health, canopy cover, and early issue detection.

Key Applications:

• Counting and sizing across large fields.
• Weed detection and canopy cover measurement.
• Prescriptive zone creation for spraying or fertilizing.

The same fast upload-and-analyze workflow applies, making Solvi effective for growers, agronomists, and drone service providers.

What types of imagery and sensors does Solvi support?
Solvi accepts standard RGB photos as well as multispectral and thermal imagery, so teams can process everything from simple visual surveys to advanced plant-health datasets. The platform is sensor-agnostic and works with popular agricultural payloads (e.g., RedEdge/Altum-class multispectral cameras) alongside integrated multispectral drones. Thermal data is converted to temperature maps for interpretation, and multispectral uploads can include radiometric calibration so indices are comparable across flights and locations.

How does PlantAI™ count plants and deliver plant-level analytics?
PlantAI™ uses machine learning to detect individual plants directly from your stitched field map. After upload, you mark example plants (10+ is recommended, capturing normal variability), preview results, and iterate if needed. The system then returns full-field counts and per-plant traits such as diameter, area, and vegetation-index-based health. It works across row crops, vegetables, vines, and trees, and you can even target non-crop detections like weeds for specialized analyses. If you prefer a turnkey workflow, a done-for-you service can produce results without training models yourself.

What image resolution and flight altitude should I use for accurate counts?
Resolution drives accuracy. For early growth stages and small plants, aim for sub-centimeter ground sample distance (often ~0.5–1.0 cm/px) which typically requires low altitudes (about 15–30 m / 50–100 ft depending on camera). For higher-level tasks like boundaries, elevation, or tree counts, you can fly higher (often 80–120 m / 250–400 ft) because each plant canopy is large in the image. Balance detail with practicality: lower altitude yields more photos and longer flights, so use the highest altitude that still delivers the plant-level detail you need.

Can Solvi auto-create microplot boundaries for plot-level statistics?
Yes. You can import SHP/KML layouts, draw custom zones by hand, or use automated trial-plot tools that detect plots or generate templated grids when gaps are hard to see. After labeling plots, Solvi computes per-plot metrics—indices, counts, canopy cover, height, and more—and you can export results for downstream analysis or reporting.

Which vegetation indices are available, and how do I choose?
Solvi supports common RGB and multispectral indices (e.g., VARI, GLI, NDVI, NDRE, SAVI/MSAVI, EVI, CHI, NDWI) and lets advanced users add custom formulas. Choose indices based on crop, growth stage, and question: RGB indices highlight greenness patterns; near-infrared/red-edge indices reveal chlorophyll, water status, and subtle stress. Remember that indices indicate variation, not root cause—ground-truthing remains essential. For cross-date/site comparisons, include radiometric calibration during processing.

How do I focus health analysis on crops while excluding soil or weeds?
Use Plant Focus to create a vegetation mask that filters out non-target pixels. You can build this mask from a thresholded index (e.g., NDVI/VARI) in early stages to exclude bare soil, or from PlantAI detections to separate crop rows from weeds or grass later in the season. With the mask applied, Plant Health maps and Zonal Statistics compute metrics only on vegetation, and canopy cover is calculated automatically for fields and plots.

How fast are results, and what export/share options are available?
After upload, Solvi stitches imagery and generates maps rapidly (smaller datasets can complete in minutes, larger sets in under about an hour), then you can run analytics immediately. Deliverables export to PDF for reports and to data formats like Excel/CSV, SHP, and GeoTIFF for GIS and agronomic tools. Share interactive web links so collaborators can view maps and metrics in a browser—no specialized software required.

Does Solvi work with integrated multispectral drones like DJI’s Mavic 3M?
Yes. Integrated multispectral platforms are fully supported. Typical workflows include flying low enough for sub-centimeter resolution when plant-level detail is required, using sufficient front/side overlap, and including radiometric references (panels and/or a downwelling light sensor) so NDVI/NDRE and related indices are consistent across dates and fields.

How are researchers and organizations using Solvi in real-world trials?
Universities, seed companies, and research institutes use Solvi for standardized microplot analytics, large multi-site trial coordination, and decision-support modeling. Examples include developing variable-rate nitrogen models from multispectral indices, digitizing thousands of sugar-beet plots for weekly assessments, scaling nationwide multispectral programs with cloud processing, and accelerating vegetable breeding with plant counts and per-plant sizing across many locations.

What best practices help avoid common plant-counting errors?
Start with crisp imagery at the right growth stage and adequate resolution; verify that small plants are separable from soil and weeds. Provide enough diverse training examples (sizes, shades, and shapes) and include “distractors” in the sample window so the model learns what to ignore. If you see false positives or missed plants, expand the sample set and adjust your sample area. For late-stage specialty crops, tighten your training around the harvestable portion (e.g., cabbage heads) to size what matters. Calibrate multispectral flights so health metrics are comparable across time and locations.

Mapping Tomato Field for Plant Counts with DJI Mavic 3 Multispectral