Digitizer v1.3 help

Valter Kiisk

Last revised: November 17, 2012

Table of contents

Overview

The program aims to offer assistance in the digitizing of graphs, i.e. recovering the tabulated X-Y dependence that the graph represents. This is the frequently required task when the original numerical data is not available. Common examples are scanned images taken of paper documents and screen shots taken of electronic media. Manual as well as (semi)automatic digitizing is possible. While only rectangular coordinate system is supported, the image can be rotated to achieve precise alignment of axes. The recovered and properly formatted data points can be saved to a text file or copied to clipboard.

Manual digitizing

Follow these steps to manually digitize a graph:

  1. Load the image from disk open or paste from clipboard paste.
  2. Zoom in zoom in or zoom out zoom out to achieve convenient view of the graph.
  3. Rotate the image to the left rotate left or right rotate right to align axes horizontally and vertically.
  4. Click move axes to show axis guides. Drag the guides with the left mouse button to make them coincident with known coordinate positions on the graph.
  5. Click axes setup to open Setup Scale dialog and specify the coordinates of the guides and the types of axes (linear, log, reciprocal).
  6. Carefully position the mouse pointer on the plot line and then click the left mouse button to mark a data point.* Continue until sufficient number of data points have been collected. Use the right mouse button to remove erroneous points.
  7. Click view data to view and format the list of data points.
  8. Within the same dialog, click save to save the formatted data to a text file. Alternatively, click copy to copy all data to clipboard.

* You can use arrow keys for accurate movement of mouse cursor with 1 pixel step (holding the Shift key makes the step of 10 pixels) and press Enter to mark the data point.

Automatic digitizing

It is possible to digitize a graph automatically provided that the curve of interest has sufficient color contrast against the remaining features on the graph (i.e. background, gridlines, annotation and other curves). Generally, a complicated curve needs to be divided into a number of fragments so that within each fragment the curve is single-valued either along the X-axis or along Y-axis. The first and last steps of the procedure are the same as for manual operation.

  1. Load, zoom and rotate the image as necessary.
  2. Click move axes to show the borders of digitizing area. Drag the borders with the left mouse button to define a properly behaving fragment of the curve.
  3. Click move axes to show Automatic Digitizing dialog.
  4. Select color of the curve by left-clicking on the color preview box and, while holding down the mouse button, drag the cursor onto the curve of interest.
  5. Set color tolerance. The tolerance of zero takes into account only pixels with the exact color selected while the tolerance of unity accepts all pixels. The contribution of a pixel will be weighted according to the color difference.
  6. Check horizontal or vertical depending on the prevalent orientation of the curve within the defined area.
  7. Click OK. Right-click to remove any misplaced data points.
  8. Repeat steps 2–7 until the whole curve is digitized.
  9. Setup the scale, format and save the data.