The following situation:
I have a map (JPG) in which cross sections are drawn and the corresponding cross sections as image files (also JPG). The cross sections have (as it is usually the case) different scales for the x and y axis. Furthermore, the cross sections are slightly tilted.
What would be the smartest way, to create a point cloud or DEM for each lithological layer based on the cross sections? Are there any QGIS Plugins that allow georeferencing and rescaling of cross sections and then to "draw" the layer interfaces?
Or is there any other free software, which I could use? Since I dont have access to any licenses apart from Move, GoCAD and ArcGIS, I would appreciate suggestions, that are either free or one of these (free would be best of course). However, as I have never worked with Move or GoCAD I stray away from them a bit.... but if they are the only solution, then so be it.
I was not able, to find tutorials or even a manual for Move. If anyone has a link or could share something like that with me, I would appreciate it, so that I also know how to do it then.
Just so I am clear. You have an image map with cross-section marked, you have the individual cross-section as seperate image files and you want to make a 3d surface for individual lithographic units from these data. I do not know any plug-ins to do it, and I am only now testing open-source software, but I can see how to do it in ArcGIS, so the process described will be in ESRI ArcGIS. A few assumptions I'm making:
1. For a lithographic unit, I assume elevation is defined as a depth below the ground surface, therefore we can use current DEMs of the surface as a baseline. I suggest depth, because once we combine this with a DEM, there will be more laterally continuous data. However, this alse requires us to think a little more about where this assumption cannot be made.
2. The frequency and distribution of the cross-sections is sufficent for interpolation. If not, you will either have a high elevation uncertainty or the data will be best suited for defining relative slope angles.
3. The cross-sections are accurate. Often these are built from just a few cores and are already interpolated. It is possiple this could be compounding errors.
Step 1. Put the image map with cross-sections into real world coordinates. You could potentially do this usuing the "georeferencing" tool and referencing to aerial imagery or other referenced baseline maps.
Step 2. Define a common point sampling interval along your cross-sections. A distance that is easily defined in the seperate cross-sections is best. Example point every 0.5 km. Make this a shapefile with attributes for each of the different units you are interested in mapping.
Step 3. Sample the cross-sections at the defined horizontal interval of the point-shapefile. You will record depth below the surface and put these values into the point shapefile attributes. There maybe an automated way to make these measurements but if not, it may just have to be done manually. If you are doing it digitally, you can bring the image into ArcGIS, scale it with the georeferencing tool and use the ruler tool.
Step 4. Interpolation. For each unit attribute you will use the point shapefile to interpolate depths. There are several interpolation options, try a few and see which "make sense." Generally the simplest methods, like nearest neighbors, make the best and most defensable products. The end of this step will have a few rasters representing each units depth below the surface.
Step 5. Raster math. Now, "minus" the unit depths from the DEM surface. The end of this step will have a few rasters representing each units elevation before quality control.
Step 6. Quality control. Go through your surface and assess where the assumed depth does not make sense. Make a polygon around areas that should not be used. For example, rivers will downcut through units so depth relationships will be different, so you may need to remove them from consideration.
Step 7. Re-interpolate. Extract areas were you believe the depth relations are correct. Convert back to points and re-interpolate.
Step 8. Validation. If this is for publication, you will probably need to validate your 3D models. Although coring is the best way, there may be places, like road cuts, where the units are exposed but were not used for interpolation.
I think that's very difficult to do with some app. If you have only jpg files of cross section, you may need to digitize those pictures into a polyline (converting your raster into a vectorial file like shapefile o cad drawing). Once you have vectoriced you need to georeference them (changing the axis). With your vectorial georeferenced Xsections you can create a surface 3D of each layer with any GIS 3D tool.
I hope that helps
Dear Tomy Truong,
You can use QGIS, GRASS GIS, ILWIS. ENVI, ArcGis (Aacademic version) etc You need to find known coordinates for geo-referencing cross section maps of different scales. After that, digitize all the cross sections then change it into point clouds or dem using analysis tool.
- Badges
- Science topic
- Similar topics
- Biological Science
- Histology
- Sectioning