Problem images

1. Diagonal edges: Although there is a diagonal edge finding algorithm in the program, if the data around the edge is sparse (i.e. the edge is poorly-defined) or there are edges running from side to side (which are assumed to be absent) or the edge is close to and nearly parallel with an edge, the routine may not find the edge or it may misrepresent it. In these cases, the resulting background surface near the edge may not be as accurate as desired. This can lead to spurious source detections within PSS.
2. Extended sources: The algorithm makes use of the appropriate point spread function (PSF) to parameterise sources. Since extended sources can not be adequately described by the PSF, their presence may not be detected, or parts of them (their wings in particular) may be missed. This can result in a contaminated background region and a correspondingly overluminous background surface around the location of the extended source.
3. Bright sources: Very bright sources possess broad wings which are difficult to completely allow for (they can cover a very substantial portion of the entire image). As a result, the output background surface may contain evidence of `domed' emission where the source was found. This should be at a low level compared to the flux in the source before removal, but could still jeopardise the identification of other weak sources by PSS.
4. Double sources. Close double sources can not be adequately parameterised by the PSF and so determination of their spatial extent on the image is not reliable. Again this can result in an unrealistic background surface estimate in the vicinity of the sources.