Measure by sampling

Run a reproducible image point count

Generate random, systematic or stratified points, classify them manually and export percentages, Wilson intervals, coordinates and an annotated image.

Start point counting

Completely free No account or payment Image stays on this device

USE THIS WHENUse this when composition or surface coverage should be estimated through transparent human-classified image points.SUPPORTED INPUTOne image · up to 1,000 points

What you will getA clear result before technical detail

01A reproducible set of normalized sampling coordinates

02Category percentages with 95% Wilson intervals

03An annotated PNG and JSON decision record

Private analysis labFiles stay inside this browser

SAMPLING / 1,000 POINTS

Choose the image you want to sampleOne JPEG, PNG or WebP · petrography, phases, quadrats, corrosion or general surface coverage

Selection is processed locally

Session diagnosticsNo processing errors

Stored only in this browser tab. Image bytes, filenames and metadata are never included.

No tool error has been recorded in this tab.

No upload or account Originals remain unchangedReview methodology

Useful and careful

What is built into the result.

Important findings come first. Technical fields remain available without taking over the page.

One transparent engine supports several human-classified image fields

Seeded sampling reproduces the same point positions

The observer remains responsible for every category assignment

Understand the method

A useful result with its reasoning attached.

This page explains what the lab measures, how to interpret it and where human review remains essential.

One sampling engine, several real image workflows

Point counting estimates composition by classifying locations selected from an image. The same statistical structure appears in petrographic mineral counts, metallographic phase fractions, ecology quadrats, surface corrosion review and other coverage studies. The categories change; the need for inspectable sampling remains.

AnalyzeImage provides general and field-oriented presets without pretending that one vocabulary fits every discipline. Categories are editable, and the report preserves the exact normalized coordinate and assigned label for every point.

Petrography and mineral composition
Metallographic phase review
Vegetation and quadrat coverage
Corrosion and coating coverage

Random, systematic and stratified designs

Simple random points avoid a fixed spatial pattern but may cluster. A systematic grid provides even coverage and easy repetition but can align with periodic structures. Stratified random sampling places one randomized point inside each grid cell, balancing broad coverage with randomized positions.

A numeric seed makes the point coordinates reproducible. Running the same method, count and seed generates the same normalized locations independent of screen size. Resetting classifications does not silently change the sample.

10–1,000 points
Normalized x/y coordinates
Deterministic seed
Visible method tradeoffs

What the confidence interval means

For each category, the lab reports the observed percentage and a 95% Wilson interval based on classified points. Wilson intervals behave more carefully than the simple plus-or-minus formula when a category is rare or the sample is small.

The interval is not total measurement uncertainty. Image selection, scale, lighting, focus, category definitions and disagreement between observers can dominate the result. The export therefore records method and decisions rather than presenting the percentage as an unquestionable material property.

Live classified count
Wilson interval per category
Unclassified points remain visible
Annotated evidence image

Three simple steps

Know what happens before you start.

Use this when composition or surface coverage should be estimated through transparent human-classified image points.

DesignChoose random, systematic or stratified sampling, categories, point count and seed.
ClassifyAssign every visible point while ambiguous decisions remain explicit.
EstimateReview percentages and Wilson intervals, then export coordinates and an annotated PNG.

RESULT ORDER

01 · A reproducible set of normalized sampling coordinates02 · Category percentages with 95% Wilson intervals03 · An annotated PNG and JSON decision record04 · Limits and next step

Clear before you rely on it

Questions this tool should answer upfront.

Short answers keep important privacy, evidence and professional-use limits visible.

01Does the point count tool classify the image automatically?

No. It generates and records the sample, while the user assigns each category. This keeps ambiguous decisions visible and avoids an unsupported automated diagnosis.

02Can another reviewer reproduce the same points?

Yes. Use the same image, sampling method, point count and seed. The JSON report also contains every normalized coordinate.

03Is the reported interval a complete accuracy guarantee?

No. It describes sampling uncertainty for the classified points. Observer judgment, image acquisition and calibration require separate controls.

Method referencesStraboSpot StraboMicro point counting manual NIST proportion confidence interval guidance

Continue when useful

Your next step, without starting over.

Move to another page only when its outcome matches what you need.

Check image dimensionsConfirm the source has enough pixels for the intended review.Inspect the fileRecord source dimensions and metadata separately.

Important limitations

Sampling intervals do not include observer or acquisition bias · The tool does not identify minerals, tissue or structural defects automatically · Professional standards may require a different sampling count, calibration or review protocol

Full limitations