Pktools - Open Source Geospatial Tools: Applications in Earth Observation - page 179

70

Milano
Lodi

65

60

55

50

45

40

35

30

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Fig. 12.4 Mean NDVI profiles based on monthly MODIS images. NDVI values are extracted from
pixels covered by the administrative regions of Milano and Lodi (Italy)

pkextract -i modis_ndvi_2010.tif -s nuts_italy.sqlite -ln milano
→

-ln lodi -bn NDVI -polygon -r mean -f SQLite -o

→

modis_ndvi_2010_mean.sqlite

The mean NDVI profiles are shown in Fig. 12.4 . Both profiles show a seasonal
variation with a maximum NDVI reached in summer. After the harvest (August), the
NDVI drastically drops until Winter time. The NDVI reaches higher values in the
region of Lodi than in Milano, which is expected due to the agricultural character
of the region. The example shown here is just for illustrative purposes only. The
analysis of NDVI profiles can be conducted at a much more detailed level if finer
spatial resolution imagery are available and if vectors contain smaller regions. It is
conceivable to evenmonitor individual points and/or pixels using the above approach.

12.4 pkstatogr

The utility pkstatogr calculates basic statistics on attributes of a vector file. Exam-
ples of the basic statistics include: minimum, maximum, median, mean and standard
deviation. Histograms (in percentage or absolute values) can also be calculated. The
attribute of interest can be selected using the option -n|--fname . Values defined
by the -nodata option, or not withing the limits set by the options -src_min and
-src_max are ignored for the statistics.