Quick Tips:
• Crop to: white background + one target curve + full X-axis range.
• CSV export auto-deduplicates X (one center point per pixel column) to reduce thick-line duplicates.
• Drag the yellow 2000 guide line to the 2000 tick inside the crop box (required for every image).
1
Upload Spectrum Image
Upload a clear image with low compression noise and strong axis/curve contrast.
2
Crop & Set Parameters
Step 2 Tips
1) The crop box should cover the full X-axis range (left end to right end).
2) Drag the yellow line in the crop box to align with the 2000 tick; split_pixel_x is captured automatically.
Drag inside box to move, drag corners to resize
Spectrum Parameters
Usually high wavenumber (left side)
Usually low wavenumber (right side)
Do not type manually. Drag the yellow 2000 guide line in the crop panel and the value is filled automatically.
3
Background Preprocessing
Step 3 Tips
• Drag multiple boxes on representative background areas (avoid the target curve).
• Set threshold: lower is strict, higher removes more similar shades.
• Choose output mode (Original or Grayscale) here, then click Apply and Continue.
Background Controls
You can draw multiple background boxes.
35
Recommended: 25-45
Boxes: 0
4
Extract Curve
Step 4 Tips
• Auto: try automatic detection first on the preprocessed image from Step 3.
• Color Pick: sample 3-8 points on the curve; tolerance 15-25 is strict, 25-40 is common, 40-60 tolerates color drift but adds noise.
• Manual Trace: vertical only (most stable). Mark multiple guide points along the curve.
Auto Detection Result
System automatically tried to detect curves using grayscale analysis. Check the result below.
Detecting...
Color Extraction
Click on the curve to sample multiple color points
30
Recommended: 25-40. Too low may miss points; too high may include noise.
No color points selected
Manual Trace
Click points along the curve
Trace Strategy: Vertical Search (fixed)
No points
Result
Drag on the extraction canvas to erase outlier points.