X-B8, 2012 
sures able to cap- 
nd provide an ac- 
ound. The patches 
^H are detected in 
at map. For these 
alculated on a per 
ue in the pixels of 
1yte patches these 
than those for the 
larger heterogene- 
r all bands of the 
criminatory capa- 
ire of energy, local 
d in the calculated 
riations in texture 
el of the selected 
bod is considered. 
dow of predefined 
' the pixel intensi- 
ened to the central 
eaningful on a per 
alues of the pixels 
n of the intra-patch 
e patch vegetation 
Juickbird bands. 
ntropy-based mea- 
ed in information 
rms, an indication 
ise, heterogeneous 
arded, as the ones 
to have higher en- 
Similarly to vari- 
n a per pixel basis. 
calculated in a sur- 
de (1) 
intensities, or gray 
quency of appear- 
the number of pix- 
ber of pixels of the 
on the one hand, 
to increase spatial 
«ture. On the other 
statistical analysis, 
xugh further quan- 
> decreases. As an 
pixels, the image 
s, so that we have 
n forming the his- 
‘wo schemes were 
ing requantized to 
y requantized indi- 
1all neighborhood 
5 variations in the 
  
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B8, 2012 
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia 
  
  
  
Figure 1: LPH/MPH and TPH habitats in Le Cesine protected site. 
extent of a larger surrounding area, we introduce the Local En- 
tropy Ratio (LER) measure. Two concentric windows of different 
sizes are considered around each pixel. A local entropy value is 
extracted for each window, H; and H, for the inner and outer 
windows, respectively, and their ratio 
Hi 
o 
  
LER = (2) 
is assigned to the central pixel. The smaller the ratio, the more 
homogeneous the close neighborhood of the central pixel, com- 
pared with its broader surroundings. Two versions of the measure 
are produced: in the first case, the pixels of the small window are 
included in the calculation of the entropy of the large one, while 
in the second a more unbiased approach is offered by excluding 
the pixels of the inner window from the entropy calculation of the 
outer one. A point of particular importance in the latter case is 
that the outer window should be large enough to allow for sensi- 
ble statistical analysis. Therefore, since a statistically meaningful 
number of pixels has to be at least one order of magnitude larger 
than the number of gray levels, in case of image quantization in 
eight gray levels and a small window size of 9 x 9 pixels, a large 
window of a minimum dimension of 13 x 13 pixels needs to be 
created around the central pixel, thus having 169—81 — 88 pixels 
after the exclusion of the central window. As previously, quanti- 
zation can be performed for either the whole region or separately 
for each window. 
3.4 Local Binary Patterns 
Local binary patterns (Petrou and García-Sevilla, 2006) are also 
tested in capturing local changes in texture. As all previous mea- 
sures, local binary patterns are computed on a per pixel basis. For 
each pixel, its surrounding pixels in a circle of predefined radius 
are considered. Each such pixel is flagged with a value of 1 if it 
is larger than the central pixel, or 0 otherwise. Scanning the sur- 
rounding pixels in a clockwise order, a binary number is formed 
from their assigned values. This number is converted to the dec- 
imal system and assigned to the central pixel. The value of the 
measure for a specific patch is calculated through averaging the 
resultant pixel values for all pixels of the patch. 
The measure can be converted to rotation invariant if all possi- 
ble binary numbers, formed by changing the starting point of the 
clockwise counting for each pixel, are considered, and the largest 
or smallest of them is finally assigned to the central pixel. Homo- 
geneous regions are expected to be characterized, in general, by 
smaller binary numbers than heterogeneous regions, since more