| 14 Mar 2012 05:54 PM |
|
|
When I use Compute Statistics under Basic Tools > Statistics and check the histogram option, the histogram of a relatively small file comes out fine, whereas the histogram of another, larger file does not seem right. It looks very regular, and every 7th bin or so has 0 in it, which looks odd and can certainly not be explained by the data. My question is: does ENVI has a known issue with computing histograms for larger datasets? This one has 643,467 points.
|
|
|
|
Deleted User
New Member
Posts:  
| 16 Mar 2012 07:22 AM |
|
|
I am wondering what the range of data values in the image is? If there are less possible data values than histogram bins, then some of the bins will end up empty. So, for example, if your data are byte or integer data, with values ranging from 0 to 64, but there are 256 histogram bins, then you will end up with a bunch of empty histogram bins, because there aren't enough distinct pixel values to assign at least one to each bin.
I hope that helps.
Peg
|
|
|
|
Deleted User
New Member
Posts:7
| 19 Mar 2012 07:02 PM |
|
|
Thanks for your reply, Peg. The data are floating point and range from 0 to 0.987. The empty bins are very regular - exactly every 8th bin is empty. It looks pretty artificial and is definitely not the distribution of the dataset.
|
|
|
|
Deleted User
New Member
Posts:
| 20 Mar 2012 09:58 AM |
|
|
Hmm. I'm not sure how to explain that. I recommend sending this question in to Exelis VIS Tech Support (support@exelisvis.com or (303) 413-3920 or your local distributor). If you can transfer your data over to Exelis, then we can explore it in more detail, and hopefully come up with an explanation.
Regards,
Peg
|
|
|
|