map2map does not downsample correctly.

[rbeyer]

I have a 4-band ISIS cube in a PointPerspective projection (at 10 km/pixel).

If I map2map it to an Orthographic projection, map2map works fine, but if I try degrading the resolution (via setting the PixelResolution in the map file, and doing pixres= map or matchmap=true, or setting it manually via pixres= mpp and resolution= ), a little bit of downsampling works as expected, but somewhere between 19 and 19.5 km/pixel (not sure of just shy of 2x degradation is relevant) the resulting cube is just full of NULLs in all of its bands.

Changing the INTERP method makes no difference. This bug is in ISIS 3.5.2 and 3.6.0.

[jlaura]

@scsides @USGSsakins Can you all verify this and get a response on here even just saying that we are able to reproduce? Then we can slot for a support fix at an appropriate time.

[scsides]

I tried different approaches to reproduce what you are seeing, but was not successful. Can you make the data, map files, and command line available to us?

[rbeyer]

Sure.

PointPerspective projected cube: http://byss.arc.nasa.gov/rbeyer/mvic_0299104952_0x545_sci_ch.cub
Orthographic Map file: http://byss.arc.nasa.gov/rbeyer/charon_0299105209.map

Example command is: map2map fr=mvic_0299104952_0x545_sci_ch.cub to=mvic_0299104952_0x545_sci_ch_map2map.cub map=charon_0299105209.map pixres=map

[scsides]

I can reproduce this using those files.

[rbeyer]

Whoa, so its something specific about these files. That's a little frightening.

[ascbot]

I am a bot that cleans up old issues that do not have activity.

Happy Birthday to this issue! 🎂

Unfortunately, this issue has not received much attention in the last 12 months. Therefore, I am going to close it. Please feel free to reopen this issue or open a new issue sometime in the future. If this issue is a bug report, please check that the issue still exists in our newest version before reopening.

[rbeyer]

Whoops, while there has been no activity, this remains an Issue.

I just used those files, and the commands included above with ISIS 4.2.0, and have the same behavior: the output of map2map is a multi-band cube with nothing but NULL values.

[Kelvinrr]

@scsides @rbeyer The issue seems to be that when the image is too small it fails to use the slower interpolation method, and the faster interpolation fails when there are far more quads than pixels and also have null data around the edges. Putting in a simple check comparing the sizes seems to have fixed it.

Pumped it up to 100 KM per pixel and it seems to be okay. PR Incoming.