15 Dec 2014 02:35 AM |
|
Hello,
at time we working with envi 4.6 and we build a decision tree. We want to classify the geometry of roofs by DEM datas. What for a code must i use for the number of Pixels. We want that the number of Pixels
|
|
|
|
Deleted User Basic Member
Posts:228  
18 Dec 2014 11:13 AM |
|
I am having difficulty understanding your question. It sounds like you would like to set up a decision in your classification that leaves out of a class all pixels with a DEM value less than 100. In that case, the expression you would enter for that node in the tree would be b1
Regards,
Peg
Exelis VIS
|
|
|
|
Deleted User New Member
Posts:  
28 Dec 2014 02:15 PM |
|
Hi Peg,
thanks for your answer. No, i want that the area
Here a small example:
({slope} gt 0 and {slope} lt 10)
This is a flat roof. But when this roof is
best Regards Kai
|
|
|
|
MariM Veteran Member
Posts:2396  
30 Dec 2014 05:53 AM |
|
There is no function built in that you could use in the decision tree to calculate area in this way. It would have to be done manually by calculating the number of pixels within an area. You can pass an IDL function to the Decision Tree but I am unsure if it could be this complicated as you would have to have an ROI or shapefile that defines the area of the roof so you could extract the number of pixels. Then knowing the pixel size, you could estimate area. I think using something like Feature Extraction would be a better option.
|
|
|
|