ROI to ASCII, band values reported that don't match image overlay

Prev Go to previous topic

Go to next topic

Last Post 20 Oct 2011 09:04 AM by anon

2 Replies

Sort:


You are not authorized to post a reply.

Author

Messages

anon

New Member

Posts:

20 Oct 2011 09:04 AM
In order to determine the pixel value for multiple bands of an image at different point locations, I opened a shapefile of the points- saved as EVF file- exported to ROI file and then from that ROI file selected Output ROI to ASCII. However in the ASCII file, the values that are reported for the bands don't always match what I see in the image overlaid with the points determined by cursor location. Specially there are some points that should obviously have a band value of 0 but instead a value is reported. The image is a stack of vegetation indices for different dates. Any suggestions? Is ENVI averaging or summing pixel values? I have double checked that the image file and point file are in the same coordinate system. Are bands not labeled in the order they are stacked? Thanks.

Deleted User

New Member

Posts:

20 Oct 2011 10:46 AM
Looks like part of the problem is the "ID" field the ASCII file uses. Any idea where that comes from? The ID numbers don't match the order of points in the attribute table of my point file, or the "PlotID" field I had created to go with the shapefile. (Currently I am trying to match up the values in the ASCII file to values a table I created in ArcMap using extract values to points, where it did keep a meaningful ID column, but considering ArcMap has been crashing a lot with multiple band rasters I would like to avoid having to do this for each image I need the pixel values from).

Deleted User

New Member

Posts:

02 Nov 2011 07:57 PM
Looks to me like the ID number field is just the pixel number. Do you have the same number of pixels in the ROI as you have points in a shapefile? I wouldn't be surprised if you had to reorder the points to make the two lists match up. For the bands that are supposed to be zero but aren't, are they really close to zero by any chance?

You are not authorized to post a reply.