[otb-users] Grouping methods (otbcli_LSMSSegmentation)

[j***@gmail.com]

Hi guys,

I'm using LSMSS as part of my Msc. thesis where I apply OBIA on VHR (RGB)
imagery in an effort to characterize potato plants by their spatial
features (shape, volume, texture, etc.). So far I'm really impressed by the
LSMSS tool (and the toolbox in general) but I have a question concerning
the spatial radius parameter in the second step (coined the grouping step
in Michel et al., 2015). In that same paper it is mentioned that *all *neighboring
pixels are grouped together if their local (spectral and spatial modes are
under the set thresholds (*hr and hs).* Now regarding the *hs *threshold,
do spatial offsets (from the spatial image/displacement image) get summed
until the *hs *value has been reached in clusters already determined by *hr*?
Or are potential clusters determined by (re)gathering all pixels that are
within that *hs *(and *hr) *threshold for each individual pixel? Looking at
the lack of *n_iter* variables and the fact that '*neighboring'* pixels
are explicitly mentioned, I would think the former makes more sense. Am I
interpreting it correctly?

Kind regards,

Jasper
--
--
Check the OTB FAQ at
http://www.orfeo-toolbox.org/FAQ.html

You received this message because you are subscribed to the Google
Groups "otb-users" group.
To post to this group, send email to otb-***@googlegroups.com
To unsubscribe from this group, send email to
otb-users+***@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/otb-users?hl=en
---
You received this message because you are subscribed to the Google Groups "otb-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to otb-users+***@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[j***@gmail.com]

I came to that assumption after reading this post
(https://groups.google.com/d/msg/otb-users/PM5XWBrAr3Q/b8qFCkuiBwAJ) where
the author himself mentioned that "h_s will operate on the displacement map
during the segmentation step".

Post by j***@gmail.com
Hi guys,
I'm using LSMSS as part of my Msc. thesis where I apply OBIA on VHR (RGB)
imagery in an effort to characterize potato plants by their spatial
features (shape, volume, texture, etc.). So far I'm really impressed by the
LSMSS tool (and the toolbox in general) but I have a question concerning
the spatial radius parameter in the second step (coined the grouping step
in Michel et al., 2015). In that same paper it is mentioned that *all *neighboring
pixels are grouped together if their local (spectral and spatial modes are
under the set thresholds (*hr and hs).* Now regarding the *hs *threshold,
do spatial offsets (from the spatial image/displacement image) get summed
until the *hs *value has been reached in clusters already determined by
*hr*? Or are potential clusters determined by (re)gathering all pixels
that are within that *hs *(and *hr) *threshold for each individual pixel?
Looking at the lack of *n_iter* variables and the fact that '
*neighboring'* pixels are explicitly mentioned, I would think the former
makes more sense. Am I interpreting it correctly?
Kind regards,
Jasper

--
--
Check the OTB FAQ at
http://www.orfeo-toolbox.org/FAQ.html

You received this message because you are subscribed to the Google
Groups "otb-users" group.
To post to this group, send email to otb-***@googlegroups.com
To unsubscribe from this group, send email to
otb-users+***@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/otb-users?hl=en
---
You received this message because you are subscribed to the Google Groups "otb-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to otb-users+***@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[j***@gmail.com]

I came to that assumption after reading this post (
https://groups.google.com/d/msg/otb-users/PM5XWBrAr3Q/b8qFCkuiBwAJ) where
the author himself mentioned that "h_s will operate on the displacement map
during the segmentation step". If it would 'just' (re)select the pixels in
a set spatial radius for *hr* evaluation, you could just use the pixel
coordinates from the filtered image.

Post by j***@gmail.com
Hi guys,
I'm using LSMSS as part of my Msc. thesis where I apply OBIA on VHR (RGB)
imagery in an effort to characterize potato plants by their spatial
features (shape, volume, texture, etc.). So far I'm really impressed by the
LSMSS tool (and the toolbox in general) but I have a question concerning
the spatial radius parameter in the second step (coined the grouping step
in Michel et al., 2015). In that same paper it is mentioned that *all *neighboring
pixels are grouped together if their local (spectral and spatial modes are
under the set thresholds (*hr and hs).* Now regarding the *hs *threshold,
do spatial offsets (from the spatial image/displacement image) get summed
until the *hs *value has been reached in clusters already determined by
*hr*? Or are potential clusters determined by (re)gathering all pixels
that are within that *hs *(and *hr) *threshold for each individual pixel?
Looking at the lack of *n_iter* variables and the fact that '
*neighboring'* pixels are explicitly mentioned, I would think the former
makes more sense. Am I interpreting it correctly?
Kind regards,
Jasper

--
--
Check the OTB FAQ at
http://www.orfeo-toolbox.org/FAQ.html

You received this message because you are subscribed to the Google
Groups "otb-users" group.
To post to this group, send email to otb-***@googlegroups.com
To unsubscribe from this group, send email to
otb-users+***@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/otb-users?hl=en
---
You received this message because you are subscribed to the Google Groups "otb-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to otb-users+***@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[j***@gmail.com]

Disregard this topic, I took another fresh look at the Orfeo Cookbook and
it dawned on me that *h_s *will operate on the displacement map as that
allows you to check the spatial distance of the local modes uses to 'set'
the values in the filtered image. It allows you to limit the merging of
neighboring pixels if their local modes are further away spatially than
what you consider desirable and/or plausible for each particular segment.

Post by j***@gmail.com
I came to that assumption after reading this post (
https://groups.google.com/d/msg/otb-users/PM5XWBrAr3Q/b8qFCkuiBwAJ) where
the author himself mentioned that "h_s will operate on the displacement map
during the segmentation step". If it would 'just' (re)select the pixels in
a set spatial radius for *hr* evaluation, you could just use the pixel
coordinates from the filtered image.

Post by j***@gmail.com
Hi guys,
I'm using LSMSS as part of my Msc. thesis where I apply OBIA on VHR (RGB)
imagery in an effort to characterize potato plants by their spatial
features (shape, volume, texture, etc.). So far I'm really impressed by the
LSMSS tool (and the toolbox in general) but I have a question concerning
the spatial radius parameter in the second step (coined the grouping step
in Michel et al., 2015). In that same paper it is mentioned that *all *neighboring
pixels are grouped together if their local (spectral and spatial modes are
under the set thresholds (*hr and hs).* Now regarding the *hs *threshold,
do spatial offsets (from the spatial image/displacement image) get summed
until the *hs *value has been reached in clusters already determined by
*hr*? Or are potential clusters determined by (re)gathering all pixels
that are within that *hs *(and *hr) *threshold for each individual
pixel? Looking at the lack of *n_iter* variables and the fact that '
*neighboring'* pixels are explicitly mentioned, I would think the
former makes more sense. Am I interpreting it correctly?
Kind regards,
Jasper

--
--
Check the OTB FAQ at
http://www.orfeo-toolbox.org/FAQ.html

You received this message because you are subscribed to the Google
Groups "otb-users" group.
To post to this group, send email to otb-***@googlegroups.com
To unsubscribe from this group, send email to
otb-users+***@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/otb-users?hl=en
---
You received this message because you are subscribed to the Google Groups "otb-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to otb-users+***@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.