Plant care

How to Identify a Plant Disease From a Photo

7 min readUpdated June 2026By the Agrosphere team

A spotted leaf, a patch of white dust, a wilting tip — most plant problems announce themselves visually long before they spread. The good news is that a clear photo captures almost everything an experienced eye looks for. With the right shots and a little pattern knowledge, you can identify a plant disease from a photo quickly, whether you farm an acre of tomatoes or keep a monstera on the windowsill.

This guide walks through exactly what to photograph, the main disease categories and how to tell them apart, how to separate pest damage from true disease, and how AI photo diagnosis works — including where it can be trusted and where it can't.

What to photograph (the part most people get wrong)

A single blurry close-up rarely gives enough to work with. A diagnosis depends on pattern, distribution, and detail — so capture all three. Take a small set of photos:

Use good, even light. Bright shade or an overcast day beats harsh midday sun, which blows out detail and creates confusing shadows. Avoid using flash on glossy leaves. Wipe off dew or water droplets first, and steady your hands or rest the phone against something for a sharp image.

Common plant disease categories at a glance

Most leaf diseases fall into a handful of visual families. Learning these patterns makes both your own eye and any diagnostic tool far more reliable.

Fungal leaf spots

Discrete spots — round or angular, often with a darker margin or a yellow halo. Some show tiny concentric rings (a classic sign of early blight, Alternaria). Spots may merge as they grow. Fungal spots are the most common category and usually start on older, lower leaves.

Blights

Fast-spreading browning or blackening of large areas, often after warm, wet weather. Late blight of potato and tomato produces greasy, dark lesions with a pale fuzzy edge and can collapse a plant within days.

Powdery and downy mildew

Powdery mildew looks like white or grey talcum powder dusted on the upper leaf surface; it wipes off and thrives in warm, dry conditions with humid nights. Downy mildew is different: yellow blotches on top, with a greyish or purplish fuzzy growth on the underside, and it favours cool, wet conditions.

Rusts

Small raised pustules that release orange, rust-brown, or yellow powdery spores, usually on the leaf underside. Rub a leaf and you may get coloured dust on your finger.

Bacterial diseases

Often angular spots bounded by leaf veins, sometimes water-soaked or greasy-looking, occasionally with a yellow halo or a sticky ooze. Bacterial leaf spots can look greasy and translucent when held to light.

Viral diseases

No spots in the usual sense. Instead look for mosaic or mottling (irregular light-and-dark green patches), yellow streaking, curling, puckering, or stunted, distorted new growth. Viruses are commonly spread by insects such as aphids and whiteflies and cannot be cured — affected plants are managed by removal.

Pests vs. disease: how to tell them apart

A surprising share of suspected diseases are actually insect or mite damage. Before treating for a pathogen, rule out pests:

It also pays to rule out non-living causes: nutrient deficiencies (uniform yellowing or interveinal patterns), sunscald, overwatering, drought stress, and fertiliser or spray burn can all mimic disease. These don't spread leaf to leaf the way infections do — another reason your whole-plant photo matters.

How AI photo diagnosis works

AI plant-disease tools are trained on large libraries of labelled images — thousands of examples of healthy and diseased leaves across many crops and species. The model learns the visual features that separate, say, early blight from a bacterial spot, and matches your photo against those learned patterns to suggest the most likely candidates with a confidence level.

In practice this means you can confirm a diagnosis in seconds by scanning the leaf with a free app like Agrosphere, which names the likely disease, scores the plant's overall health, and gives a practical treatment — the product, dose, and timing — in plain language. Because the core scan runs on your device, it works in the field even without a signal, and it speaks dozens of Indian languages, which removes a real barrier for many growers.

Good results still come down to the photo. Feed any AI tool the well-lit, in-focus, close-up-plus-underside shots described above and the suggestion will be far more accurate than from a single distant blur.

The limits of AI diagnosis (and when to get a human)

AI photo diagnosis is a fast, useful second opinion — not an infallible verdict. Keep its limits in mind:

Use AI to narrow the field and act quickly, then escalate to a person when the stakes or the uncertainty are high.

Acting on the diagnosis safely

Once you've identified the likely problem, match the action to the category:

Always read and follow the product label for the correct crop, dose, timing, pre-harvest interval, and safety gear, and follow your local regulations on which products are approved. When in doubt, start with the least-toxic effective option and remove infected material first — good sanitation prevents more disease than any spray.

Not sure what’s wrong with your plant?

Point your camera at the leaf and Agrosphere names the problem and the exact fix — free, offline, in your language.

Open Agrosphere

Frequently asked questions

Can you really identify a plant disease from just a photo?

Often, yes — many diseases have distinctive visual patterns that a clear set of photos can capture. But some bacterial, viral, and nutrient problems look alike and may need a lab test or an expert to confirm, so treat a photo diagnosis as a strong lead rather than a final verdict.

What photos should I take for the most accurate diagnosis?

Take the whole plant (to show where damage occurs), a sharp close-up of an affected leaf, the leaf underside, and ideally a leaf that is half-healthy and half-affected. Use bright, even light, avoid flash on glossy leaves, and wipe off any water droplets first.

How do I tell a pest problem from a disease?

Look under leaves and at growing tips for insects, webbing, or sticky honeydew. Chewed holes mean chewing insects, fine speckling suggests mites or thrips, and a sticky coating points to sap-suckers — none of which are fungal or bacterial diseases.

How accurate is AI plant disease identification?

AI is a fast, helpful second opinion and is most accurate when the photos are well-lit and in focus and the result has high confidence. It can struggle with look-alike symptoms and doesn't know your local conditions, so confirm important or whole-crop decisions with an agronomist.

Does Agrosphere work without internet?

Yes. The core leaf scan runs on your device, so you can identify a disease and get treatment guidance in the field even with no signal, and it works in dozens of Indian languages.

Can a plant disease be cured once I identify it?

It depends on the type. Many fungal diseases can be managed or stopped with sanitation and the right fungicide, bacterial diseases can only be slowed, and viral diseases cannot be cured — affected plants are removed and the insects that spread them are controlled.