What causes contour lines in this Viewshed?

3707
12
Jump to solution
06-23-2021 11:01 AM
OrleansOscar
New Contributor II

I have produced a viewshed output here. What causes the (what appears to be) contour lines?

The raster surface was derived from 1m lidar data.

 

Can this be remedied? What would a good workaround be?


Edit: England

Tags (2)
1 Solution

Accepted Solutions
waxg
by
New Contributor

With the DEM data in your area of interest, the following are some of the experiments I did in order to see the source of the contouring effects.


1. With the original DEM at 2m resolution, I ran the Visibility tool and I can reproduce the contouring effects in the output (rendered using Orange color in below). The contouring effect in the Visibility output is due to the presence of terraces in the data, which is visible in the hillshade/elevation background.

waxg_0-1626459038192.png

 

2. After the DEM is resampled to 5m resolution using the Nearest Neighbor method, I ran the Visibility tool again and the new output based on the 5m resolution still has some contouring effects, as shown below:

waxg_1-1626459038196.png

 


3. After the DEM is resampled to 10m resolution using the Bilinear method, I ran the Visibility tool again and the new output based on the 10m resolution is free of contouring effects, as shown below:

waxg_2-1626459038194.png

 

4. With the original 2m DEM, I changed the observer height to 10 meters, and re-ran the Visibility tool. The new Visibility output (purple) is overlayed underneath the original Visibility output (orange) as follows. The new output (purple) contains less contouring effects:

waxg_3-1626459038219.png

 

5. With the original 2m DEM, I changed the observer height to 20 meters, and re-ran the Visibility tool. The new Visibility output (purple) is overlayed underneath the original Visibility output (orange) as follows. The new output (purple) contains even less contouring effects:

waxg_4-1626459038220.png

 

In summary,
(1) For the given DEM (2m with terrace like landform), the tool is producing correct results. Only the edges of the terrace is visible, not the top, which is producing contouring effect.
(2) You can change the observer height (if applicable) to see the top of the terrace.
(3) Or, you can change the resolution of the DEM with the bilinear resampling method to remove the terraces from the DEM and make it smoother, which will remove the contouring effect. However, by doing so, you will be changing your DEM, which may or may not be appropriate for your analysis.

Please let me know if you have additional questions.

Thanks
Xuguang Wang

View solution in original post

12 Replies
JayantaPoddar
MVP Esteemed Contributor

Does the contour-like area represent a landscape of terrace- farming? And maybe the observer(s) can only view the vertical sides of the landscape, not the flat-tops.



Think Location
0 Kudos
OrleansOscar
New Contributor II

I should have mentioned that this is in England, and the area in question is grassy fields for livestock.

Good thinking with the terrace farming though - admittedly I did have to check for my own sanity!

0 Kudos
JayantaPoddar
MVP Esteemed Contributor

Could you share a few XY Coordinates (Decimal Degrees) or the extent of the area?

Even the DEM looks like a terraced-terrain. You can see the sharp edges (delineations).



Think Location
0 Kudos
DuncanHornby
MVP Notable Contributor

Terracing is not a common land use in England certainly in the agricultural sense. I believe @OrleansOscar question remains valid. @OrleansOscar  you need to describe what tools you used, parameter setting to create your output, also provide the source of the data.

0 Kudos
OrleansOscar
New Contributor II

Hi  @DuncanHornby & @JayantaPoddar 

I have uploaded an image of both the surface and viewshed for more context. The AOI is that of the area surrounding 54.00992254131516, -2.234673540228602

The tool face is also uploaded, it's the Visibility Analysis tool, inputs are classic 360 / 90 for a number of points in this area. But the contouring also appears from a single observer point too.

Surface data was created using the Mosaic to New Raster tool using 1m DSM lidar tiles (dated 2004) from DEFRA.

Thanks again

0 Kudos
DuncanHornby
MVP Notable Contributor

You say "The tool face is also uploaded..." I don't see anything? No indication of what the tool used or parameter settings?

0 Kudos
ChrisM
by
New Contributor III

Probably an issue with the high res of the raster. What is the pixel size, 1 m?

OrleansOscar
New Contributor II

Hi Chris,

Yep! 1m pixel size...

0 Kudos
ChrisM
by
New Contributor III

You might resample the raster to increase the pixel size and smooth it out more. I have similar problems even using 5 m LiDar. They often pick up stuff that might be interesting but are obstacles to what I want to do. These two links might help. Conversely you might see if you can download some data for your area with a lower res and rerun the viewshed.

https://desktop.arcgis.com/en/arcmap/10.3/tools/spatial-analyst-toolbox/altering-the-resolution.htm

https://desktop.arcgis.com/en/arcmap/10.3/tools/spatial-analyst-toolbox/how-filter-works.htm