I am working on native uint16 tif grayscale images with GDL installed on a Portable Ubuntu Remix TRES runnin' on Vista Home.
I did compile with Magick++ (but no Python).
My script reads each image in gdl and mathematically processes it for some defects, firing out a uint16 corresponding matrix. I would need to convert the matrices back to images.
Unfortunately no write_tiff procedure is (still) available in gdl and the idl write_tiff.pro script fails.
Neither the write_bmp and write_png procedures work at all (they're in a loop), resulting in a curious "(BMP or PNG) Image format not recognized" error and in the termination of processing.
However, by reading in the very same original tif images or by manually creating a uint16 array (e.g. uintarr(n,n)) outside of my script, I seem to be able to use write_png or write_bmp for writing images back.
Any clues on how to convert my matrices? Any idea on why this happens? Did anybody succeed in having a write_tiff.pro working in gdl?
The current CVS version of GDL has a Magick++ bug fixed that caused some malfunctioning on Ubuntu - perhaps using the new version might help? If not, it would be great if you could post here some example code that reveals the problem.
I made serious cleanup of the "image_test.pro" in testsuite/ directory (in the CVS), and few in the files in src/pro. All the examples in this code are working for me, on various machines. Please try this code, report success or problems. Don't hesitate to expand it to new cases (I have very few knowledge on these various formats.)
If you have a simple code and the image file on the web, please send code + link