Facility-inspired plant phenotyping, now portable

Phenotyping leaves the lab and goes wherever the crop is.

iPhonotype turns iPhone into a LiDAR-guided plant phenotyping tool that removes manual scaling, tripods, and rigid capture setups. Built-in LiDAR and RGB fusion keep scale grounded while users move freely around complex plant structures, and the system benefits from richer scans through oversampling and optimized compute instead of punishing them. Right away, the app can track plants, perform semantic leaf segmentation, estimate full 3D volume, and support broader phenotyping measurements through auto-guided scans that stay approachable for new users and deep enough for serious plant science.

Download See the Mac Companion

Scan to Print Capture on iPhone, refine on Mac, and export ready-to-print color workflows.

Broad Device Support Works across newer LiDAR iPhones and older image-first devices.

Free Mobile Workflow Keep capture simple on the phone while the Mac Companion handles the heavier tasks.

Animated plant phenotyping pipeline showing above-ground LiDAR scanning and below-ground root analysis.

Above-ground traits Below-ground traits 3D scan to print Deep-learning color inference Industrial fruit measurement PAR meter Older + newer iPhones Prusa + Snapmaker output QR tracking Geolocated field capture Free frontend workflow

Privacy

Your scan data stays with you.

iPhonotype is designed as a fully front-end workflow, so user data stays on the device and is never sent to any external server or cloud service. The iPhone app and Mac Companion run locally, which means we do not collect, see, or store user scans, projects, images, models, or other private information.

Users keep full ownership and control over their data from capture through analysis and export. There is no account requirement, no hidden upload path, and no need to trust a remote backend with sensitive research or field data.

At a time when many people are tired of their data being swept into training pipelines for large AI companies, we believe the person who creates the data should remain its sole owner.

Fully on-device All capture, review, and project data remains on your iPhone or Mac instead of being pushed to a server.

We collect nothing We do not see, ingest, store, or monetize user information, scan content, or experimental records.

You stay in control You decide what to keep, export, share, or delete, without relying on any external cloud system.

No cloud. No hidden upload. No external data trust required.

What’s Next We’re committed to expanding access. In the near future, we’ll be launching versions of iPhonotype for Windows PCs and Android devices, bringing the same privacy-first, cloud-free experience to even more users across platforms.

Real workflows

Real images from the iPhonotype innovation stack.

These images focus on Mac Companion validation, captured scene reconstruction, tray-scale 3D geometry, and the pipeline that leads to multi-color printable assets.

At the Netherlands Plant Eco-phenotyping Centre (NPEC) Ecotron Module, we can scan relatively large plants inside a cylindrical habitat only 60 cm in diameter because the device remains compact and the operator keeps enough freedom to move into narrow gaps, tight recesses, and other occluded parts of the canopy.

Mac Companion validation panels alternating between imported captures, overlays, and benchmark comparison views. — **Mac Companion validation** Validation tools enable comparison of benchmark masks, imported captures, overlays, and plant-level outputs within the companion workflow. For facility-grade benchmarking, iPhonotype was calibrated against Helios at NPEC on 4,143 plant samples from 218 matched captures across 56 tray groups, achieving 0.805 plant count MAE, 2.15 mm² area MAE, and near-perfect area correction (R² = 0.999).

Mac Companion workspace showing scan organization, tray groups, and above-ground phenotyping dashboards. — **Mac Companion validation** Validation tools enable comparison of benchmark masks, imported captures, overlays, and plant-level outputs within the companion workflow. For facility-grade benchmarking, iPhonotype was calibrated against Helios at NPEC on 4,143 plant samples from 218 matched captures across 56 tray groups, achieving 0.805 plant count MAE, 2.15 mm² area MAE, and near-perfect area correction (R² = 0.999).

Single-scene reconstruction A 3D reconstruction rendered from the original scan, showing the plant and surrounding soil from above before moving into a slower orbit for inspection, visualization, and treatment-response comparison.

Before flash drought Matched native USDZ camera pass showing Echinacea purpurea before the drought treatment, with a denser and more cohesive canopy structure.

3D height: 39.7 cm
Footprint: 0.339 m²
Hull volume: 0.088 m³

After 4 days of flash drought Matched native USDZ camera pass showing the plant after four days at 35C, with a more open, fragmented, and visibly less saturated canopy.

3D height: 27.5 cm
Footprint: 0.227 m²
Hull volume: 0.030 m³

Stress signature in the reconstruction The stressed plant reads as shorter, more open, and more fragmented, and the 3D trait export shows a clear contraction in footprint and hull volume after treatment. One important caveat: this comparison is measuring visible leaf particles and canopy clumps rather than perfect biological leaf instances. In the pre-stress scan, overlapping leaves merge into fewer larger connected components, so the weak point is leaf splitting when leaves touch, not the segmentation itself.

Photogrammetry base Convex hull overlay Sync viewer Observe mode Mesh opacity 100%

Before flash drought Higher footprint and larger hull volume, with a denser pre-stress canopy.

39.7 cm • 0.339 m² • 0.088 m³

Loading 3D compare...

After 4 days of flash drought Shorter and visibly contracted, with a markedly reduced footprint and hull volume.

27.5 cm • 0.227 m² • 0.030 m³

Loading 3D compare...

Photogrammetry base layer Native RGB points Convex hull wireframe Height-colored geometry Greenness-derived color Leaf-component centroids

Open full drought report

Tray-scale textured output Beyond single plants, the system can assemble textured tray reconstructions for comparison across multiple samples.

Species-aware workflow suggestion A 73-class EfficientNetV2-S classifier, trained on imagery from iNaturalist, GBIF, and Naturalis Biodiversity Center, reached 0.684 top-1 accuracy on 1,688 held-out test images. That model can suggest a species-aware workflow before analysis, while still letting the user confirm or override the recommendation.

3D reconstruction pipeline

From Helios chamber scale to tray-resolved Arabidopsis analysis.

The reconstruction pipeline works with 3D scans from large growth chambers to tray-level capture, starting inside the Helios chamber with hundreds of emerging potato plants.

It is also portable and mobile enough to fit and work inside double-shelf chambers, where Arabidopsis rosettes grow tightly packed.

Stage 1 · Helios scale Helios chamber-scale potato emergence A selfie-stick helped extend the iPhone through the full Helios chamber, capturing about 250 pots with germinating potato plants in one large reconstruction.

Stage 2 · Growth Chamber Growth Chamber Captured at arm’s length with no accessories, this single-shelf growth-chamber view shows how the workflow still recovers chamber-scale structure in a compact space.

Stage 3 · Tray analysis Tray-resolved Arabidopsis analysis

This tray-level Arabidopsis scan captures 17 plants in less than a minute while preserving enough structure for per-plant convex hulls, rosette footprint, apex height, and longest-leaf measurements. The result is a compact but information-rich 3D scene that keeps the whole tray readable while still supporting plant-by-plant geometry and color analysis.

Loading tray analysis viewer...

Illustrated FLIR-style handheld thermal imager for future thermal plant scanning workflows.

Future thermal fusion with FLIR One A compact FLIR One thermal imager can soon become part of the iPhonotype workflow. Because FLIR exposes an SDK for this device, the app is being prepared to tap into live thermal imagery, fuse it with the existing 3D scans, and analyze temperature patterns directly on reconstructed plant geometry. View the Amazon example

ColorChecker reference target placed inside a tray-style scanning setup for color normalization before analysis.

Color-checker calibration for repeatable color and greenness Because iPhonotype is used as a mobile capture system across benches, chambers, and different field spaces, the workflow can now include a ColorChecker reference photo before tray scans, 3DGS packages, and other captures. That gives the Mac Companion a stable target for color normalization, helping keep greenness, color-derived traits, and visual comparisons consistent over time and between environments.

iPhone prompt asking whether to add a color target photo before starting a 3DGS scan.

iPhone Color-checker calibration The app can prompt the user to take a dedicated color-target photo before a scan begins. Users can shoot a fresh reference, choose an existing one, or continue without it, depending on how tightly they need appearance, color, and greenness measurements standardized for that session.

Root phenotyping tracking Root images acquired through the workflow can also be post-processed through the pipeline to track root structures over time and support downstream root phenotyping analysis. On newer devices that support 4K 120 fps capture, including iPhone 17 Pro Max, scanner-mode video can run with the lens almost touching the Petri dish (or glass-tube) to build extremely large, high-magnification panoramas for fine root tracking.

Multi-color 3D printing

Automatic print-ready plant models from a single scan.

The same iPhonotype reconstruction pipeline can turn a captured plant scene into a ready-to-print multi-color model bundle without rebuilding the geometry by hand. This example was prepared into a five-color Prusa XL workflow with preserved structure, mapped color regions, and slicer-ready exports.

In practice, that means teams can move from mobile capture to a printable physical replica for outreach, teaching, inspection, or downstream fabrication, while keeping the geometry, learned color inference, and print segmentation tied to the original scan.

5-color Prusa XL bundle OBJ + STL + 3MF export Automatic palette mapping

Download 3MF bundle Download STL

Loading print bundle preview Preparing the five-color OBJ model for orbit and inspection.

Interactive print preview Drag to orbit the generated model and inspect how the automatic color-mapped print bundle is laid out before slicing.

iOS app

Field capture, app UI, and tray-level outputs on iPhone.

This section focuses on the mobile side of iPhonotype: real tray capture in the phenotyping environment and iPhone-side outputs that structure plant detections before deeper Mac-based review.

Soybean tray capture at NPEC Vinicius Lube, iPhonotype developer, scanning soybean trays in a real phenotyping environment using the iPhone capture workflow.

Soybean tray analysis showing segmented overlays, leaf counts, canopy area, and summary traits. — **Soybean analysis result** Immediate tray-level soybean analysis with segmented overlays, estimated leaf counts, canopy area, and summary trait tables.

Arabidopsis tray image with fitted plant bounds and convex hull overlays in 2D. — **Arabidopsis tray detections** Arabidopsis tray detections with fitted plant bounds and convex hull overlays in 2D.

Tray phenotyping and project dashboard The iOS app keeps tray phenotyping and project organization visible alongside live object tracking inside the 3D Gaussian Splatting capture mode in one field-ready workflow.

Real-time root capture The same mobile workflow can also record close-range root sessions in real time for downstream frame decomposition, stitching, and root phenotyping analysis. Baseline deep-learning models support root masking, roots-plus-shoots phenotyping, and root-tip tracking across successive frames, so a narrated iPhone capture can turn directly into structured outputs for trait extraction, growth review, and longitudinal comparison inside the Mac Companion.

Illustrated LED backlight panel recommended for back-lighting petri dishes.

Root phenotyping dashboard A full results surface generated from the same capture workflow, combining the selected frame, segmentation overlays, primary-versus-lateral masks, skeleton QA, dish summaries, and per-plant root traits in one reviewable report.

Open full dashboard

Additional iOS sensing tools Beyond capture and segmentation, the iOS side can also log local light conditions and help users navigate dark imaging spaces without breaking assay conditions.

Leaf reflectance logging The app can standardize and log reflected light measurements directly from leaf surfaces during dedicated sensing rounds.

iPhone ambient PAR screen using the front-facing camera to measure incident light on plants. — **Ambient PAR logging** The front-facing camera can also be used to estimate incident light on the plants and log it alongside scanning rounds for context.

iPhone live depth-assisted awareness screen used for navigating dark imaging rooms. — **Live depth-assisted awareness** A low-light depth view can help users move through very dark rooms, including bioluminescence assays and confocal setups, without flooding the space with visible light.

Local AI assistance on Mac The Mac Companion can run local LLMs through Ollama, so users can ask questions about scans, summaries, traits, metadata, and analysis warnings without sending anything to the cloud. In practice, this workflow can run models such as gemma4:e4b, but any suitable open-weights model can be used depending on the machine and the task.

Local AI assistant panel in the Mac Companion showing Ollama-backed scan question answering.

Key capabilities

One capture system, multiple phenotyping realities.

Capture plants in growth chambers, factory lines, petri-dish assays, orchards, and field plots without switching platforms every time the biology changes.

3D scanning to multi-color printing

Recover canopy geometry, plant height, and scene structure, then use deep learning for color inference and printer mapping into ready-to-print outputs for Prusa and Snapmaker.

Root workflows

Ingest petri-dish photos or videos, decompose frames, stitch root coverage, and extract below-ground traits.

QR traceability

Track trays, petri dishes, and fruit lots with QR-linked sessions for repeat imaging and clean dataset provenance.

Geolocation-aware capture

Record where scans were acquired so field phenotyping stays connected to place, treatment, and route.

Runs on many iPhones

Use advanced iPhones for LiDAR-rich capture or older iPhones for lighter image-first workflows without changing the overall platform.

Frontend-first and free to run

The mobile experience is designed as a lightweight frontend, so users can run the iPhone side for free while the Mac Companion handles heavier processing.

Mac companion app

The heavy lifting happens where it should.

iPhone capture stays fast in the field and on the bench. The Mac Companion takes over for validation, organization, model runs, reconstruction, and export, so the mobile experience stays simple while the desktop side handles the deeper work.

Two practical hardware tiers

For the most affordable setup, a MacBook Neo paired with an iPhone 15 Pro covers the full platform for under €1,400.
For heavier reconstruction and the most demanding trait workflows, a MacBook Pro with the latest iPhone 17 Pro Max offers more processing headroom for about €3,300.

These are real screenshots from the Mac Companion app, covering benchmark validation, scan organization, and above-ground phenotyping analysis.

Review scans, benchmark sets, overlays, and validation results.
Organize projects, trays, sessions, and imported datasets.
Run analysis for shoots, roots, fruit, and 3D reconstruction workflows.
Export print-ready 3D assets for downstream production and presentation.

Explore use cases Discuss deployment

Real desktop workflows Benchmark validation, batch phenotyping, tray summaries, and project organization inside the Mac Companion.

Mac Companion benchmark alignment studio comparing benchmark masks and app masks. — **Batch phenotyping** Per-image summaries, CSV exports, and segmented overlays from above-ground analysis.

Mac Companion benchmark alignment studio showing benchmark RGB frame comparison. — **RGB frame comparison** Side-by-side review of benchmark RGB frames against imported app captures.

Mac Companion above-ground phenotyping screen with batch summary and segmented overlays. — **Benchmark alignment studio** Mask-first validation for benchmark comparisons, app overlays, and import quality control.

Mac Companion phenotyping screen for potato and soybean tray analysis. — **Tray-level summaries** Quick summaries for canopy area, leaf counts, and plant-by-plant tray breakdowns.

Mac Companion phenotyping screen for Arabidopsis tray analysis with many plant detections. — **Dense tray analysis** High-count tray workflows for Arabidopsis and other compact phenotyping layouts.

Mac Companion library view organizing flash drought scans and tray groups. — **Workspace and scan library** Organize projects, trays, sessions, and scan groups before moving into validation and downstream analysis.

Use cases

Built for research groups, agri-food operators, and field teams.

Research phenotyping

Track rosettes, seedlings, roots, and treatment responses across repeat sessions with benchmark-linked calibration, model evaluation, reproducible exports, optional multi-color 3D outputs, and video-to-panorama workflows for close-range macro plant imaging.

Agri-food quality workflows

Measure fruit geometry, surface condition, and object dimensions on benches or processing lines with QR-linked lots and Mac-side model orchestration.

Field phenotyping

Capture geolocated plant records in real environments on many classes of iPhone, then sync to the Mac Companion for deeper analysis and 3D post-processing, as well as curated dataset management.

Why this tool

Phenotyping should be portable, accessible, and easy to deploy.

Democratization of phenotyping

Replace single-purpose lab infrastructure with a platform that can move from controlled environments to industrial sites and real field conditions, while also producing shareable and printable 3D outputs with learned color mapping.

Broad device reach

Use the hardware people already carry, from advanced capture-capable iPhones to older models, then add deeper processing only where it matters with the Mac Companion.

Accessible cost model

A frontend-first mobile app plus the Mac Companion lowers the barrier to entry and lets users run the mobile workflow for free.

Dizitalizing house plants

Capture on iPhone. Reconstruct on Mac. Infer color. Print in multi-color.

iPhonotype bridges plant biology, machine vision, and operational usability so teams can move from mobile capture to trait insight, printable 3D outputs, deep-learning color mapping, and multi-color print preparation for Prusa and Snapmaker.

Review capabilities

House plant scan Explore a real iPhonotype house plant scan captured from an `.ftscan` package. Drag to orbit, scroll to zoom, and open the Apple-ready USDZ below if you want a portable scan asset for Apple devices.

Loading interactive scan Preparing the house plant mesh for orbit, zoom, and clean visual review.

Open USDZ Download optimized scan

3DGS scanning pipeline

Portable single-plant field capture with photogrammetry, height, and greenness review.

One portable field scan resolves a compact ornamental plant with 33.7 cm canopy height, 0.071 m² footprint, 0.014 m³ hull volume, a 15.2 cm longest leaf, and 22,464 retained 3D points. The same capture can then be reviewed through photogrammetry, convex-hull context, height mapping, and greenness analysis without leaving the browser.

Portable field reconstruction Three recovered 3D components, a 15.2 cm longest leaf, and 22.5K retained points make this a compact but measurement-ready field scan.

33.7 cm • 0.071 m² • 0.014 m³

Photogrammetry base Convex hull overlay Observe mode Mesh opacity 100%

Loading interactive field scan...

Photogrammetry base layer Native RGB points Convex hull wireframe Height-colored geometry Greenness-derived color

Front-facing 3D capture

Front-facing TrueDepth capture for compact plants on iPhones without LiDAR.

Many iPhones do not carry LiDAR, but they still include the front-facing TrueDepth sensor. That makes close-range 3D capture possible for compact rosettes and other small structures, keeping the plant centered for quick shape, height, and color review on a much broader range of devices.

Contact & partnership

Working on crop science, 3D scanning, phenotyping infrastructure, or industrial measurement?

We want to hear from anyone measuring plants, whether that work happens in a growth chamber, a greenhouse, a processing line, or out in the field. Many of those conversations grow into academically grounded collaborations around portable capture, quantitative analysis, and reproducible 3D plant workflows.

[email protected]

Email [email protected]

Research collaborations

Industrial pilots

3D printing outputs

Model deployment

Dataset partnerships

About us

Built by a phenomics practitioner and refined where real plant science happens.

By day, Vinicius works on large-scale phenotyping projects across controlled environments, operational workflows, and research-grade measurement systems. In his free time, he built iPhonotype out of both passion and rigor, turning day-to-day phenomics challenges into a portable tool that could actually move with the biology.

NPEC has been the proving ground for the app, helping him refine it with real plants in real lab conditions, and out in the field as well. When you use iPhonotype, you are using something that has been tested in the same spaces where cutting-edge plant science is actively happening.

“I hope it helps you as much as it has helped me and my students.”