Skip to main content
  • EN

Hi everyone,

 

 

I was wondering whether anyone had strategies for automating the georeferencing process. I have a raster displaying parking lots and am trying to create a dataset that shows the painted parking space lines. In the past, I've manually drawn the parking space lines, but was wondering whether anyone had any suggestions on how to automate this process.

Basically, I've been thinking about how to draw a rectangle over the entire area of the parking lot and then somehow divide the parking lot into a specific number of spaces. However, since each parking lot has a different number of spaces, I'm stuck on how I'd automate this process.

Any ideas?

Thanks

Hi @jennamkdavis,

This sounds like an interesting challenge for FME. There are some good resources around on the knowledge centre for help when it comes to converting images to vector. Check out this tutorial which shows an example of isolating parts of an image: https://knowledge.safe.com/articles/1226/using-conditions-with-the-rasterexpressionevaluato.html?smartspace=raster_2 This type of thing could help you. There is also a great FME Hub transformer called the RasterConvoluter which might help a little with things like edge detection. Another article is this one which talks a little more about Raster expressions: https://knowledge.safe.com/articles/43753/raster-calculations-and-raster-palettes.html

 

When it comes to actually performing the conversion you could check out this blog which gives a few ideas on methods. https://blog.safe.com/2017/08/3-ways-to-convert-a-raster-image-to-vector/

 

I encourage you to post an example of a raster you're working with as a starting point.

Hi, @jennamkdavis

this is interesting. Drawing a series of parking boxes is not a problem. The problematic part of it is the logistic around the parking. So if we ommit the circulation, be it one way or bi-directional, if we ommit having one or more exits, if we do not take handicaped parking places into account, I would probably go ahead with rotating and orienting the parking lot so that the future divisions would be exactly horizontal/vertical. Then a tiler who splits the area to tiles of 2,4 x 5,3m (or however you wish) and then I would filter out the "incomplete" tiles. Maybe at this point would do some shifting left/right/up/down and see if some position gives more tiles than others.

Having the ful tiles only, use a tester to assign them either as a parking lot or a driveway. Such as (tile)rows 1, 4, 5, 8, 9, 12, 13 parking lots, the others driveways. Similar with columns.

I want to say that to me it seems unrealistic to automate it fully as you will always need to do manual adjustments due to factors that you cannot anticipate.

Reply


Community Stats

30,469
Posts
115,707
Replies
38,288
Members
Terms & ConditionsCookie settings
A product of
Safe Software respectfully acknowledges that we live, learn and work on the traditional and unceded territories of the Kwantlen, Katzie, and Semiahmoo First Nations.