Global Mapper Pro

Extracting water features from imagery

arklatexgeo
arklatexgeo Global Mapper User
edited May 2014 in Raster Data
Has anyone experimented with extraction of water features from hi-res imagery? I am looking for a way to analyze a set of imagery such that potential water features can be extracted from an image and converted to a vector feature dataset.

Comments

  • global_mapper
    global_mapper Administrator
    edited January 2013
    If your water features tend to be the same color you could potentially extract them by right-clicking on the layer in the Control Center and selecting the option to create equal-value areas for similar colors. However water features often have a variety of colors due to waves, etc. so this is far from foolproof. Often you might be better suited extracting water features, streams in particular, using the watershed delineation capabilities from a terrain surface rather than imagery.

    Thanks,

    Mike
    Global Mapper Guru
    gmsupport@bluemarblegeo.com
    http://www.bluemarblegeo.com/
  • arklatexgeo
    arklatexgeo Global Mapper User
    edited January 2013
    Thanks Mike. That sounds like a good place to get started. We are looking for smaller sources of water in one particular region so your color method might just work.
  • superstefan
    superstefan Global Mapper User Trusted User
    edited May 2014
    Hi Mike,
    is there a guide or webinar on how to use watershed delineation? I need to "paint" the lakes and rivers in my imagery, but it's not possible to select equal-value colored areas.

    Thanks!
  • global_mapper
    global_mapper Administrator
    edited May 2014
    The webinar on using 3D data in Global Mapper at Blue Marble Geographics Previously Held Webinars has a section on doing watershed analysis if I'm not mistaken. Note sure where it is in the video, but if you jump around you can probably find it.

    Thanks,

    Mike
    Global Mapper Guru
    geohelp@bluemarblegeo.com
    Blue Marble Geographics for Coordinate Conversion, Image Reprojection and Vector Translation
  • arklatexgeo
    arklatexgeo Global Mapper User
    edited May 2014
    I had almost forgotten about this thread. I am glad it was resurrected.

    I am currently experimenting with UAV-acquired imagery for aerial surveying, water feature location, and precision agriculture. The cameras I am using are a Canon SX260HS (12.1MP, 3-band NC imagery) and a Canon SX230HS (12.1MP, 3-band NIR/infrablue imagery). The cameras are mounted on an autonomous hexacopter with a brushless 2-axis gimbal. I am using a Trimble GNSS (Juno 3B data logger, Pro XH receiver, Tornado antenna) to establish GCPs and using Agisoft Photoscan Pro to post-process the imagery. I would love to be able to do all of the post-processing and imagery analysis/feature extraction within the Global Mapper environment. Having a one-stop-shop for this step in the process would allow complete automation of the process.

    Is there any existing way to do all of this within the GM environment? Photoscan allows us to align, mosaic, and generate dense point clouds as well as export DEM/DSM all in one workflow. Having the ability to do all of this and follow the flow with feature extraction and even NDVI would be a wonderful addition to GM. If this capability doesn't currently exist within the GM framework, are there future plans to include it?
  • global_mapper
    global_mapper Administrator
    edited May 2014
    You can manually rectify the imagery if it doesn't already have geo-location in it, but there isn't currently any way to take the raw camera parameters and turn that into an ortho-rectified mage. Once you get a geo-located image then you can use the Raster Calculator under the Analysis menu to perform NDVI and NDWI (or whatever else you want) calculations, then you could extra areas from those grid layers using the right-click option in the Control Center to create areas from equal-value areas.

    Thanks,

    Mike
    Global Mapper Guru
    geohelp@bluemarblegeo.com
    Blue Marble Geographics for Coordinate Conversion, Image Reprojection and Vector Translation
  • arklatexgeo
    arklatexgeo Global Mapper User
    edited May 2014
    That's kind of what I thought. Photoscan does a wonderful job of ortho-rectification and mosaic production. The resulting files are often LARGE due to the sheer size of the input files. With 12.1MP cameras at 122 meters AGL flying height, we can produce imagery on the order of 5cm/px resolution. You can imagine how large these files can get. The processing time can be quite long as well. Anyway, I digress. The one thing about GM that I like the best is it's handling of elevation and 3D analysis. IMO, GM is head and shoulders above ArcGIS in terms of elevation data analysis. The ability to work with LIDAR (and other point clouds) within the GM environment is one of the primary reasons we are using it. If GM could rectify and extract elevations from our (geo-tagged) images, we would have a single-package solution. Being able to go from camera to final deliverable within one software environment would be HUGE.
Sign In or Register to comment.