Hi all,
I'm trying to find some software that will aid in converting a TIFF image to vector format for use in CAD.
This solution from Autodesk doesn't work well when I input big TIFF images.
Does anyone have experience with this software from BLOM?
Another we are looking at is WiseImage, but it's a bit pricey at $6k.
I've also found a method using QGIS, but I'm always skeptical of open source solutions working reliably.
If anyone has any experience vectorizing their orthophotos/aerial maps, that would be really helpful!
Replies
http://www.gdal.org/gdal_polygonize.html
Its open source software, but rock solid. I have used it for many years. Its highly regarded in GIS/RS spheres. Here's a list of software using gdal: http://trac.osgeo.org/gdal/wiki/SoftwareUsingGdal (so you dont have to take my word for it)
typically you dont vectorize aerial photography - because you'd probably get a big mess. So if you need to classify the photos. Give QGIS a shot. In their plugin repositories they have a tool that should do the job. I never used it but it looks promising. Its called, Semi-Automatic Classification Plugin
Hi David,
Let's take a step back. What are you trying to accomplish?
The vectorization of QGIS you mention wasn't made to do a direct conversion. It was made when you have clear classification criteria, such as vectorize all raster points with elevation > 500 into a polygon. Don't even try to vectorize without such conditions, because you get 800MB files.
If you're interested into classification of land cover, you should look into using artificial intelligence solutions. One commercial tool that does this is "Overwatch" (now Textron):
https://www.youtube.com/watch?v=MxT39BK1_iU
http://www.textronsystems.com/products/geospatial/feature-analyst
It uses neural networks to train on classification of raster data to extract features. So it can extract polygons of vegetation, water and buildings separately.
QGIS has been around for years and it has an immense amount of plugins. I wouldn't discard it as any bit of opensource software, because a lot of the software is being used by hundreds of GIS users around the world. QGIS is actually composed of a number of subprojects, like GDAL and GRASS (which also have their own GUI's or CLI's).
In GRASS (and QGIS) you can also use AI algorithms for land cover classification, but it probably takes a couple more steps. Look for "maxlik" on google to see more tutorials on this technique:
https://www.youtube.com/watch?v=_6JDzNcE2A0
What many people do is simply digitize a raster manually by hand using polygon drawing tools on a new vector layer and try to use existing datasets to quicken the pace. This works very well if the area is small and you just want specific elements of the data.
You can look into merging your data with alternative data sources like OpenStreetMap for example. Then you get roads and some houses for free already and often the water bodies.
If you provide more info on what you're trying to accomplish, we can take it from there.
Hello, newbie here. QGIS can be used to assemble images taken with a uav? For mapping?
No, you need an app to do photogrammetry for that. See the wiki of project-ion above and the processing of data section. It shows some tools you could use for that.
I also wanted to perform this myself for the purposes of creating thematic map that reduce the amount of detail of raster maps (a bit like cartography). I have a written tutorial here, which requires some refinements and screenshots that I add later. This is using QGIS. Is this what you intend to achieve?
https://github.com/gtoonstra/project-ion/wiki/Automated-feature-cla...
Hey Gerard,
Thanks for your response!
While classification looks useful, our end goal is to produce something like this... So actually creating the survey drawing is the goal. While some people do it by hand, it is incredibly time consuming. One of my colleagues said that he'd found a couple of softwares that would actually draw lines on building edges, etc. Probably building on some of the classification methods/softwares you listed above.