Orfeo Toolbox - Open Source Geospatial Tools: Applications in Earth Observation - page 206

(a) (b)

Fig. 13.9 Edge detection filters applied to the SWIR band of the LISS-III scene. a Gradient filter.
b Sobel filter

For the purposes of the demonstration of the utility we will continue to use the
subset of the LISS-III scene over Shannon airport. In a first instance, we will run the
command line tool using the gradient filter, which is the default selection on band 4,
which is the Short Wave Infrared band.

otbcli_EdgeExtraction -in subset_snn.tif -channel 4 -out
→

subset_snn_edges.tif uint16

The edge extraction is presented in Fig. 13.9 and it clearly shows that the edges
between land cover types are clearly identified as bright colored pixels. We can
re-apply the edge detection algorithm, but by specifying the sobel filter, which results
in a similar output image but with subtle differences. The touzi filter is also available,
but it is generally used for radar images with speckle noise.
If you prefer, it is possible to redisplay the grayscale images using the color
mapping tool, otbcli_ColorMapping , which generates an 8-bit color image
from a look up table. The application includes a number of predefined look-up tables
that are defined using the option -method.continuous.lut . The command
line options are outlined below:

-progress
< boolean > Report progress

-in
<

string > Input Image (mandatory)

-out
< [pixel] Output Image [pixel=uint8/uint16/int16/uint32/int32/float/

double] (default value is float) (mandatory)
-ram
<

string >

int 32 > Available RAM (Mb) (optional, off by default, default value is 128)