[Vineyard & Winery Elf EN](https://help.vineyardelf.com/index.md) / [VineyardElf](https://help.vineyardelf.com/vineyardelf.md)

# [Machine Vision - AI leaf photo analysis](https://help.vineyardelf.com/machine-vision-ai-leaf-photo-analysis.md)

## What is Machine Vision?

Machine Vision is a feature that analyzes photos of your plants using artificial intelligence. You upload a photo of a leaf - the system tells you whether the plant is healthy or something might be wrong with it.

It's like having an expert plant pathologist in your pocket!

## What can it detect?

### Diseases
- **Powdery mildew** - white, floury coating
- **Downy mildew** - yellow spots, greyish fuzz on the underside of the leaf
- **Grey mould (Botrytis)** - grey, fluffy coating on fruit
- **Black rot** - brown spots with a dark rim
- **Esca** - discolouration between the leaf veins

### Nutrient deficiencies
- **Nitrogen** - general yellowing (starting from the bottom of the plant)
- **Iron** - yellow leaves with green veins
- **Magnesium** - yellowing between the veins
- **Potassium** - brown leaf margins

## How to run an analysis?

### Step 1: Take a photo

- Photograph the leaf close up (15-30 cm)
- The leaf should be well lit
- Avoid shadows and reflections
- A plain background works best

**Tip:** Photograph leaves showing symptoms, but also a few healthy ones for comparison.

### Step 2: Send it for analysis

In the VineyardElf app:
1. Go to **Leaf analysis** in the sidebar
2. Click "Choose file" or drag in your photo
3. Pick a category (diseases, deficiencies, pests, stress)
4. Click "Analyze"

### Step 3: Review the results

After a few seconds you will see:
- **Annotated photo** - boxes around the detected problems
- **List of detected problems** with a confidence level (confidence %)
- **Class** - the name of the detected disease or condition

## Where can I find my analysis history?

At the bottom of the Machine Vision page you'll see a gallery of your previous analyses:
- Photo thumbnails
- Dates
- Detected problems

## Tips for better results

**A good photo:**
- Sharp and well lit
- The leaf fills most of the frame
- Both healthy and affected parts are visible

**Avoid:**
- Photos in shadow or backlit
- Blurry shots
- Photos taken from a distance (the whole vine)

## What if the system gets it wrong?

AI is not infallible! If a result looks odd to you:
- Take more photos from different angles
- Compare it with the disease descriptions in the knowledge base
- When in doubt, consult an expert

---

## Vineyard camera - automatic monitoring

Since February 2026 there is also **automatic camera monitoring** with Reolink. A camera mounted in the vineyard takes a photo every 30 seconds and sends it for AI analysis without any action on your part.

A live preview from the camera is available on the [dashboard](/a/elf-dashboard), and the full configuration is in the [Vineyard camera](/a/elf-vineyard-cam) tab.

**Planned:**
- Automatic alerts when the AI detects a disease
- Drone integration (analysis of photos from flyovers)
- Bird detection + deterrence

---

## Related articles

- [Vineyard camera - Reolink + AI monitoring](/kb/vineyard-elf/kamera-winnicy-monitoring-reolink-ai)
- [Disease monitor](/kb/vineyard-elf/monitor-chorob-4-modele-predykcyjne-DzWeasYW)
- [Panel Reader - reading gauges](/kb/vineyard-elf/odczyt-paneli-urzadzen-integracja-z-zewnetrznymi-systemami-nhzY22Ny)
- [Dashboard](/kb/vineyard-elf/dashboard-podglad-danych-i-sterowanie-winnicy-XNkpfXCF)