I am using the landsat tree cover continuous fields dataset from Hansen et al. (2013) to calculate the mean forest cover within HUC 12 watershed boundaries. However I can't seem to figure out how to calculate the mean. The attribute table for the data does not contain the pixel values (0-100) for percent canopy cover. I can only find these by identifying a certain pixel or by looking in the metadata. I tried copying the metadata into excel but there are way too many zeros. There has to be a way to do this.
See the attached images. The one labeled attribute table shows that the pixel values do not show up in the attribute table. The next image labeled PXVL shows that when you identify a pixel the value shows up. The image attached labeled metadata shows the metadata for the image and that there are a lot of zeroes that surround the outline of the image in this dataframe. Is there a way to extract these values to calculate the mean?
Any help will be much appreciated!
I would use ImageJ for this purpose, but ArcGIS Zonal Statistics works as well...
Hey Sean,
This is simply to do in ArcGis. Just import your data and look to metadata. If you cant see the values in the attribute table, you can try to work with raster tools, such as Zonal Statistics.
I think your problem is not how to calculate the mean, but that you didn't find one band or field which contains the pixels of percent canopy cover. This is a common problem faced for people who use land cover data. So, the first step is that you should extract these pixels that contain percent canopy cover. I haven't used ERDAS, I don't know how to solve this problem in ERDAS. But I know that ArcGIS can do it.
Your image contains more than one field or band in the attribute table. In ArcGIS, you can see the field that contain percent canopy cover in the attribute. Then you can use "Reclassify" tool to select one field to output (see the figure below). Of course, the pixel values of image must be the type of integer. In the figure, I take an example one image that contains two field (i.e., value and count) and extract the count field pixels from this image.
After this field is output, you can use Zonal Statistics to calculate the mean.
This question is relevant to the structure of your image. If you don't solve this problem, you can send your image (if small enough) to my email. Good luck!
Arc GIS zonal statistics is not working for me. I dont see anywhere where it gives you a value of a mean it just changes my image to black and white.
Sean, I'm not sure if this question is still active, but i thought to add a comment anyway. You don't have to extract the values to calculate mean. This is already done for you through the Image Metadata. Simply click on the 'General' tab in the 'Image Metadata' and you would find the 'Mean' under the 'Statistics Info:' block. I have used the third image (pixel data) that you have attached with your question and have added the general tab over it to show the Mean. Check the attached image and let me know if it works.
There will be lots of option I.e. minimum cell value, maximum cell value, mean cell value etc. Hope this will help you
