---
license: cc-by-nc-nd-4.0
task_categories:
  - image-classification
  - object-detection
tags:
  - drone
  - UAV
  - multispectral
  - soybean
  - precision-agriculture
  - remote-sensing
  - stink-bug
  - crop-monitoring
pretty_name: Mavic3M Anhumas Soybean Stink Bug Resistance 2026
size_categories:
  - 100K<n<1M
---

# Mavic3M Anhumas Soybean Stink Bug Resistance Dataset 2026

## Dataset Description

This dataset contains UAV imagery of soybean fields collected for stink bug (*Nezara viridula* / *Euschistus heros*) resistance research. Images were acquired using a **DJI Mavic 3 Multispectral** drone at 10 m flight altitude over experimental plots in Anhumas, São Paulo, Brazil, spanning **10 time points from January 14 to April 13, 2026**.

> **Access is gated.** Please submit an access request with your institutional affiliation and intended use. Approval is required before downloading.

---

## Experimental Design

Two treatment conditions were compared across all time points:

| Condition | Folder suffix | Description |
|-----------|--------------|-------------|
| Treated | `vcpcontrol` | Plots with insecticide application |
| Untreated | `VCPnocontrol` | Control plots without treatment |

Each flight date includes one flight per treatment condition (2 folders per date, 20 folders total).

---

## Data Collection

| Parameter | Details |
|-----------|---------|
| Platform | DJI Mavic 3 Multispectral (Mavic 3M) |
| Flight altitude | 10 m AGL |
| Location | Anhumas, São Paulo, Brazil |
| Date range | January 14 – April 13, 2026 |
| Time points | 10 dates |
| Total size | ~450 GB |

### Time Points

| # | Date |
|---|------|
| 1 | January 13, 2026 |
| 2 | January 19, 2026 |
| 3 | January 29, 2026 |
| 4 | February 4, 2026 |
| 5 | February 27, 2026 |
| 6 | March 17, 2026 |
| 7 | March 23, 2026 |
| 8 | March 30, 2026 |
| 9 | April 7, 2026 |
| 10 | April 13, 2026 |

---

## Data Structure

```
Mavic3M_Anhumas/
├── DJI_YYYYMMDDHHMMSS_NNN_vcpcontrol10m/
│   ├── DJI_YYYYMMDDHHMMSS_NNNN_MS_G.TIF      # Green band (560 nm)
│   ├── DJI_YYYYMMDDHHMMSS_NNNN_MS_R.TIF      # Red band (650 nm)
│   ├── DJI_YYYYMMDDHHMMSS_NNNN_MS_RE.TIF     # Red Edge band (730 nm)
│   ├── DJI_YYYYMMDDHHMMSS_NNNN_MS_NIR.TIF    # Near-Infrared band (860 nm)
│   ├── RGB/
│   │   └── DJI_YYYYMMDDHHMMSS_NNNN_D.JPG     # RGB downward camera
│   ├── *_PPKNAV.nav    # PPK GNSS navigation file
│   ├── *_PPKOBS.obs    # PPK GNSS observation file
│   ├── *_PPKRAW.bin    # PPK raw GNSS data
│   └── *_Timestamp.MRK # Image timestamp markers
└── DJI_YYYYMMDDHHMMSS_NNN_VCPnocontrol10m/
    └── (same structure as above)
```

### Image Modalities

| Type | Format | Bands | Description |
|------|--------|-------|-------------|
| Multispectral | GeoTIFF (.TIF) | G, R, RE, NIR | 4 individual band images per capture |
| RGB | JPEG (.JPG) | R, G, B | Downward-facing visible camera |

### Auxiliary Files

PPK (Post-Processing Kinematic) GNSS files are included for precise georeferencing of each flight:
- `.nav` — Navigation/ephemeris data
- `.obs` — GNSS observation data (RINEX format)
- `.bin` — Raw binary GNSS data
- `.MRK` — Image trigger timestamps for synchronization

---

## File Naming Convention

**Multispectral:** `DJI_YYYYMMDDHHMMSS_NNNN_MS_[BAND].TIF`  
**RGB:** `DJI_YYYYMMDDHHMMSS_NNNN_D.JPG`

- `YYYYMMDDHHMMSS` — Acquisition datetime
- `NNNN` — Sequential image number within the flight
- `[BAND]` — Spectral band: `G`, `R`, `RE`, or `NIR`

---

## Intended Use

This dataset is intended for research on:
- UAV-based crop monitoring and phenotyping
- Multispectral vegetation index analysis (NDVI, NDRE, GNDVI, etc.)
- Detection of stink bug damage in soybean canopy
- Time-series analysis of treatment effects on soybean growth

---

## Citation

If you use this dataset in your research, please cite accordingly. Citation information will be updated upon publication.

---

## Contact

For access requests or questions, please contact the dataset owner via Hugging Face.