hi .. Im Baraah Amer from birzeit university today I will explain for you accuracy assessment
Collect reference data: "ground truth" • Determination of class types at specific locations Compare reference to classified map •
accuracy should be at least 85%
Accuracy Assessment dtermind on The nature of the region , The aim of the study, choosing reference source .. size of reference plots ..position and number of samples... sampling method
Open Erdas ERDAS IMAGINE 2014 program
open a viewer and display the supervised image from open raster
open a viewer then open raster and display an "un-classified" image from different satellite such as spot if u are using landsate
fit to frame
from raster icon we choose supervised then accuracy assessment
from file icon we choose open
we get our supervised image
okay
select edit then CREATE/ADD RANDOMM POINTS
number of points set it to 100 or set it as u want
Select the distribution you want
either "random", "stratified random", or "equalized random"
RANDOM which observations are randomly placed.
Stratified Random Sampling : a minimum number of observations are randomly placed in each category
I will select the distribution parameters to RANDOM
we have different numbers it depends on your classifications ID
We don't want 0 so I will delete them
right click select Criteria
Class == 0
select Close right click delete selection
select VIEW then SELECT VIEWER
the image you select is the reference
then select VIEW | SHOW ALL.
All of the random points will be displayed as WHITE in the viewer that you have selected earlier as a reference
in REFERENCE column, enter your best guess of a class VALUE for the pixel covered by each reference point "the numeric value"
As you do this the color of the point in the viewer will change to yellow
to see the numeric value you assigned to each class display attribute table for ur supervised image
ten select REPORT | ACCURACY REPORT .
it isn't that difficult to understand most of the report.
The error matrix simply compares the reference points to the classified points. The Kappa coefficient
expresses the proportionate reduction in error generated by a classification process compared with the error of a completely random classification
if u want more information look at the link in the description box.



Không có nhận xét nào:
Đăng nhận xét