This is a program which utilizes the included functions in order to calculate 8 image indices (Bias, Correlation coefficient, DIV, Entropy, ERGAS, Q, RASE and RMSE). The purpose of the program is to produce the results fast, easily and in a convenient way for the user (see Outputs). Initially, its purpose was to perform index analysis in hyperspectral and multispectral satellite imagery. It has been used and tested in fused hyperspectral products for quality assessment of the spectral fidelity. However, it is estimated that it can be used for image comparison of similar or processed images, of completely different origin. Every included function can be used separately.
*For citation please go to bottom of description.*
==================
Program Structure:
==================
0) User runs the program by typing 'imanalysis' in the command window.
Inputs:
1) User must provide the program with the number (nin) of the test images
(test) he desires to compare, with the original image (orig).
*** All images must have the exact same resolution ***
2) After image inputs, user is being asked for the h/l ERGAS ratio.
3) Then, the user has to input the filenames, first that of the original
image and afterwards, those of the (nin) test images declared in step 1.
Index analysis:
4) Program performs computation of all eight indices for every image and
for every band, by using seven independent functions. The average value
is calculated for every index. Total values are also computed for
Entropy, ERGAS and RASE indices.
Outputs:
5) Program outputs an Excel file, containing each index analysis results
in a homonymous spreadsheet. For ease, or later statistical operations,
a column has been added to the left, numbering the bands of the tested
imagery and a row above, containing the filename. User of course, can
examine and plot the index results from Matlab command window. By typing
before the index 'c' and after the index 's', the cell array containing
the certain index is shown. For example, to display ERGAS index, we must
type 'cergass'. See lines 326-333 in imanalysis, for every index (2nd arg in xlswrite).
-Compatibility-
-Oldest Matlab version tested: 7.0.1 (R14SP1). Bear in mind that you will
not be able to analyze hyperspectral images with this version.
-Oldest Matlab version known to have full functionality: 7.6 (R2008a).
*This program does NOT use sliding windows in index computations.*
------------------------------
Citation / Link to paper
------------------------------
This is the citation and link to the study/paper, where you can find an application of the present program in fused hyperspectral and multispectral satellite imagery.
If you are using this program for academic purposes, it would be very kind of you if you cite it properly:
A. D. Vaiopoulos, "Developing Matlab scripts for image analysis and quality assessment", Proc. SPIE 8181, 81810B (2011); doi:10.1117/12.897806
Link of the paper above:
http://dx.doi.org/10.1117/12.897806
Link to program download (current page):
http://www.mathworks.com/matlabcentral/fileexchange/32637
|
