Pixel density, resolution, and raster-to-vector question

[Anonymous]

Hello All!

First post.

What criteria is the raster-to-vector feature(s) looking for when auto vectoring a raster entity?

For example, is it looking for high contrast areas; for grayscale images, does it look for bit depth.

How does the resolution and the bit depth affect the auto vectorization process? 

Can a greater resolution and/or bit depth help the application "choose" the best possible path for a vector entity?

We are researching solutions to vectorize a large number of scans of old, very low contrast engineering and tooling drawings.

Many times the vector entity will lock-up or not compute due to what seems like a low contrast issue.

Even under the best scanning conditions, and optimizing the raster image with the histogram, high contrast, sharp line edges are not possible.

Thank you for your help!

[bullera]

Hey welcome to the Forum!

Contrast and continuity of pixels on the original image impact things a lot.  If once the image is converted to bitonal and lines have a lot of disconnected pixels in the linear feature(s), the vectorization process starts really breaking down.  I’m honestly not surprised if sometimes you’re locking up while Raster Design tries to "find" adjacent information when pixels are highly disconnected.  Also, with poor color separation, I’ve seen entire linework disappear from the image when converted to bitonal.

In a lot of cases with the scanned image comes "noise" but often running through the Cleanup options helps.  In extreme cases I’ve had to process the images in Adobe first, then bring it into Raster.

Would you be willing to share one of your scanned images so I could take a look and see what could be done?