How are the recommendations from the photo analysis calculated?

In short, the photo analysis process has three main steps:

1. Image processing – detecting leaf cover from your photos

2. Estimating N-uptake – using field trial relationships between leaf cover and nitrogen uptake

3. Calculating the N-recommendation – applying crop-specific agronomic inputs

Limits: Reliable only in early growth.
• Winter wheat ≤70 % leaf cover (≤40 kg N/ha)
• Winter oilseed rape ≤80% leaf cover (≤80 kg N/ha)
The app blocks denser images. For later cereal stages, use the Yara N-Tester.

1) From photo to leaf cover

1. Green-red difference: The red channel is subtracted from the green channel. Green vegetation reflects more green light and absorbs more red light, helping us identify leafy areas more easily.
2. Automatic thresholding: An algorithm is applied to separate leaf pixels (foreground) from soil/background. For oilseed rape after winter, the algorithm also recognises red/brown pixels typical of senescent leaves.
3. Noise reduction and smoothing: Small artefacts and smooth edges are removed to improve accuracy.
4. Result: The app calculates leaf cover (%) as the share of the image covered by leaves.

2) Estimating N-uptake from leaf cover

Through field trials, Yara established empirical relationships between leaf cover and nitrogen uptake. N-uptake rises with increasing leaf cover, but accuracy declines at high canopy density, making the analysis reliable only in early growth stages.

• Winter wheat (BBCH 20-30): reliable up to 70 % leaf cover (≤40 kg N/ha)
• Winter oilseed rape: reliable up to 80 % leaf cover (≤80 kg N/ha)

The app blocks images above these thresholds to avoid misleading results. If you need nitrogen guidance in cereals at later growth stages, you can use the Yara N-Tester.

3) From N-uptake/leaf cover to N-recommendation

Winter oilseed rape – total seasonal N

What you provide:

• Measurement date: before/after winter

• Mineralisation potential: low/medium/high

• Yield expectation: 1-6 t/ha
• If after winter: autumn N-uptake (kg N/ha), or (if unknown) fraction of brown leaves (%)

How we calculate:

• Estimate N-uptake and fresh matter weight from leaf cover (three pathways: autumn only, autumn and spring, spring only)

• Convert N-uptake into a total seasonal N-recommendation, adjusted for mineralisation potential and yield expectation

• Recommended total can be split into applications (e.g. BBCH 20-27 and BBCH 30-39)

Winter wheat – first dressing only

What you provide:

• Seeding date: early/normal/late

• Common first N dressing: 50-90 kg N/ha

• Variety: selected from list

• Growth stage: BBCH 20-30

How we calculate:

• Compute a first-dressing recommendation from leaf cover (no fresh matter/N-uptake output)

• Apply correction factors for variety and growth stage (based on field trial data)

• Combine corrected leaf cover and seeding date to retrieve a relative recommendation (offset against a local standard for the first dressing)