search

AlanRace / SpectralAnalysis

Software for the analysis and interactive exploration of spectral imaging data
https://alanrace.github.io/SpectralAnalysis/
Apache License 2.0
29 stars 9 forks source link

Normalisation on multi image composites not working #2

Closed Rory-Steven closed 7 years ago

Rory-Steven commented 7 years ago

Hi Alan, Using TIC or L-2 norm (possibly others I've not tried) on a composite image comprised from multiple, smaller images (combined using your converter) gives rise to errors. L-2 norm doesn't seem to work at all. TIC norm appears to work but when I try to create a mean spectrum with TIC norm a blank spectrum is returned also images created are binary with pixel values of 0 or nan. I've checked it on one of the original single images which made up the composite and the issue still occurs for TIC norm mean spectra and image generation but L-2 norm seems to work. Original images were non-rectangular. Rory

AlanRace commented 7 years ago

Hi Rory, This should be fixed in the latest commit (https://github.com/AlanRace/SpectralAnalysis/commit/41a83fe22c05d199c53d4d94622042e193a1c215). Please download the latest version of both MOOGL and SpectralAnalysis and let me know if you're still experiencing this issue.