I am having a problem with the line of sight tool and viewsheds on ArcGIS 10.1. This is for a viewshed analysis of pillboxes and defences in a particular area in Malta.

I am using a DTM with a 1m cell size. Structures have been filtered out. A discrepancy has arisen with the Line of Sight and the Viewsheds. The observer point of both LoS and Viewshed are the same (1.78 metres) and the target height is also the same (1.7 metres). There are areas which the viewshed shows as visible while the line of sight is stating otherwise. Has anyone ever encountered such a problem before and how can I solve this? enter image description here

  • Please always include images within your question. External links tend to fail over time, reducing the value of both the question and answer. There are also folks who won't touch (or are blocked from) Google Drive or DropBox links, whom you exclude from your potential pool of answerers.
    – Vince
    Oct 29, 2015 at 18:17

1 Answer 1


Line of sight and viewshed are two different things:

  • Line of sight shows the visible locations for one observer along a line (the observer is the first vertex of the line provided as input). If you revert the observer and the target (end vertex of the line), the results will be different.
  • Viewshed shows whether an area is visible for a set of observers: if an area is visible for at least one observer, it will be marked as 'visible'.

This is why there are more areas marked as 'visible' along the line by the Viewshed result than by the Line of Sight result.

The Visibility tool (new to ArcGIS 10.2.1), with its analysis_type parameter, will provide you with more options to determine whether an area is visible for one or more observers, and which ones.

Or, if you want to align the output of the Viewshed to the Line of Sight's, export the first vertex of the line to a point and use it as input for Viewshed, instead of using the line.

  • 1
    Is the Visibility tool, which @GISGe was referring to, available in ArcGIS 10.1? Further, why "Viewshed shows whether an area is visible for a set of observers"? Isn't a viewshed calculated taking into account 1 observer? Thank you for any clarification.
    – NewAtGis
    Oct 30, 2015 at 11:46
  • @NewAtGis: The tool is new to 10.2.1, thanks for pointing this out. According to the help, "A viewshed identifies the cells in an input raster that can be seen from one or more observation locations". So yes, there can be several locations.
    – GISGe
    Oct 30, 2015 at 11:59
  • Thanks for the follow-up. As for the observed location(s), I understand that is you have a polyline, the observer locations will be the vertices of the polyline. But I think that what the poster is asking is relative to a viewshed from a single observation location (i.e., point).
    – NewAtGis
    Oct 30, 2015 at 12:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.