Blue Logo

Plot Digitizer

Back Back Simulation Tools  ·  Free

1Upload
2Calibrate
3Detect
4Fit
5Export
Upload Plot Image

Upload a plot or curve image (PNG, JPG, etc.). The tool will automatically detect colored curves and data points.

Upload file
Click to browse or drag & drop
PNG, JPG, GIF, or WebP
Calibrate Axes
Axis X (Horizontal)
Axis Y (Vertical)
Plot Area (% of image)

Define where the actual chart area is. Pixels outside are ignored. Use the grid preview to align.

The blue rectangle on the preview shows the crop area.
Options
Lower = detect more similar colors. 30 = balanced.
Detect & Refine
Auto-Detection
Detects colored pixels and groups them into curves. Adjust sensitivity in Step 2 if needed.
Detected Series
Curve Fitting
Series to Fit
Fit Method
Polynomial / Linear
Lineary = a·x + b
Quadraticy = a·x² + b·x + c
Cubicy = a·x³ + b·x² + c·x + d
Quarticy = a·x⁴ + b·x³ + c·x² + d·x + e
Quinticy = a·x⁵ + b·x⁴ + c·x³ + d·x² + e·x + f
Nonlinear
Exponentialy = a + b·e^(−c·x)
Half-lifey = a + b / 2^(x/c)
Exp. Growth/Decayy = Y₀ − (V₀/K)·(1 − e^(−K·x))
Powery = a·xᵇ
Gaussiany = a·e^(−(x−μ)²/2σ²)
Michaelis-Menteny = Vmax·x / (Km + x)
4PL Logisticy = D + (A−D) / (1+(x/C)^B)
Spline Interpolation
Natural CubicPiecewise cubic, S″(x₀)=S″(xₙ)=0
AkimaPiecewise cubic, local slope weights
Smoothing Splinemin Σ(yᵢ−S(xᵢ))² + λ∫S″²dx
Export Data

Download your extracted data in multiple formats.

Export Format
Chart
Upload a plot image and configure axes
to extract data curves.

About Plot Digitizer

Extract numerical XY data from any plot image, including scanned papers, screenshots, or exported graphs. The tool detects colored curves automatically and lets you export clean data to CSV, Excel, or JSON. Everything runs locally in your browser. No image is uploaded to any server and no data ever leaves your device.

Workflow Overview

The tool guides you through five steps:

  • Step 1 - Upload: Load a PNG, JPG, GIF, or WebP image of your plot.
  • Step 2 - Calibrate: Set axis min and max values and define the plot boundary. This is the rectangle that encloses the actual chart area, excluding axis labels and margins. You can also adjust sensitivity here if detection is too aggressive or too sparse.
  • Step 3 - Detect: Run auto-detection. The tool scans for colored pixel clusters and groups them into series. You can remove unwanted series or individual points, and hover over points in the list to locate them on the image.
  • Step 4 - Fit: Optionally fit a mathematical curve to any series. Supported options include polynomial, exponential, power law, spline, and more. R² and RMSE are shown for each fit.
  • Step 5 - Export: Download your data as CSV, Excel (.xlsx), or JSON. The JSON export also includes calibration metadata.

Tips for Best Results

  • Distinct colors: Each curve should have a clearly different color since the detector groups by hue.
  • High contrast: Dark lines on a light background work best. Low contrast or anti-aliased edges may require a lower sensitivity value.
  • Plot boundary: The blue rectangle in the preview shows the crop area. Keep it tight so axes, tick marks, and labels are excluded. Stray colored pixels outside the data area can be detected as false series.
  • Sensitivity: Higher values group more hues together, which results in fewer series. Lower values split similar colors into separate series. Start at 30 and adjust if needed.
  • Resolution: Higher resolution images produce more sample points per curve and improve axis accuracy.
  • Spline vs. polynomial: For smooth empirical curves without a known functional form, Akima or Natural Cubic spline is usually the safest choice. Use polynomials only if you have a clear reason to expect one.