Filrer/Identify cloudy or dew images
When doing the Registration, the number of images that matched the reference image can be an indicateur to exclude poor quality image. Several possibility : Stacked Images. Usually in this case the value dropped by at least 50%. Cloudy images : It depends of the could cover, Dew on objective : Same things. The number of star that match is reduced. When imaging milky way regions, with DSLR and large field the number of stars drop significativelly (20% or more). This criteria is significant. I.e with more than 1500 stars in the picture. When using a CMOS with small field or less stars (short exposure) it is not relevant.
The idea would be either or both :
- have a new "Matching" indicator that can be used in the Stacking panel (like FWHM, Roundess or Quality) and add a filter with a level that would exclude all the frame with a matching value below this level.
- have an option into Registration panel to exclude all frame that have a matching value below than some level. The user could be able to select them later after checking if it is relevant or not.
Using this command to parse the log you get the top list of frame to exclude : grep Found 2020-01-03T18.06.50.324624Z.log | awk '{ if(($3/1781)<0.75){print $3, $7} }' | sort -rn where
- 2020-01-03T18.06.50.324624Z.log is the log file
- 1781 is in this case the max number if detected star
- 0.75 is the percentage of matching
Result on 350 images : Number of star image number, 1312 371, 1312 351, 1310 377, 1308 379, 1308 374, 1306 382, 1298 352, 1297 375, 1296 380, 1294 378, 1293 359, 1282 354, 1273 353, 1263 355, 1254 266, 1248 357, 1233 356, 1202 295, 1191 114, 1165 335, 1157 196, 1139 14, 1107 205, 1065 362, 981 381, 935 358, 884 383, 873 312, 360 289, Once check manually, the identified images have to be to be rejected.
Just test by yoursellf :-)