Taking photographs for raster to vector conversion

by Auntie on February 16, 2010

Users sometimes send us photos of objects they’d like to vectorize the outline of using our Scan2CAD raster to vector conversion software.

Unfortunately, using raster to vector software to vectorize the outline of an object on a photo is not as simple as it sounds. One reason for this is that before you can vectorize the object outline you need to divide the photo into two discrete color areas: The object and the background. This is usually hard to do automatically because the object and the background tend to share colors that make it difficult or impossible to separate them.

If you just take any old snap of an object, like the one of the Nice Niece below, you won’t be able to separate the object from the background. In this instance it is quicker to simply trace around the outline you want manually.

Object and background share colors
There is too much commonality between the colors in the object and the background. It is impossible to separate them.

The trick is to photograph the object against a background which is a different color to the color of the object.

I got someone to hold a sheet of black card behind the Nice Niece while I took a photo. Now, the background and the Niece are largely different colors and it is possible to separate them using Scan2CAD’s thresholding or color palette tools.

Object and background different colors   Object and background separated
The object and the background are different colors.
 
  They can be separated into two discrete color areas …
Vectorization of object outline    
… and vectorized.    

Leave a Comment

Previous post: What are TWAIN and WIA?

Next post: Do affordable A2 / C-size scanners exist for CAD and CNC work?