RPi as GUI programmable smart camera with stromx - Raspberry Pi Forums
hi all,
during last 2 years developed first (probably beta-quality) version of open-source project "stromx".
http://www.stromx.org
consists of graphical editor "stromx-studio" allows assemble image processing pipelines (= streams in stromx terminology). these streams consists of image operators linked threads. streams can executed in stromx-studio or in other application links against stromx library.
thus, raspberry pi camera module used smart camera in following way: first build vision pipeline on pc in stromx-studio , save file. transfer file raspberry pi , execute stream in slim, gui-less command line tool calls couple of functions of stromx library. in next step parameters of vision stream remotely configured via web-interface.
have far:
- stromx runtime library , stromx-studio known compile , build on raspberry pi. however, not necessary run stromx-studio on raspberry pi if stuck procedure above.
- python bindings stromx libraries exist. simple application executes stromx streams can implemented python script.
# doing:
- adding many more vision operations integrating large parts of opencv library.
# cool in future:
- precompiled packages raspbian. ubuntu packages (x86) stromx build on regular basis should not complicated.
- implement web-interface served on raspberry pi , allows remote clients control vision stream. again, python bindings available tornado [1] + stromx solution option.
- implement end-point operator in stromx stream allows transfer result data smart camera (i.e. raspberry pi) remote client collects data , processes further (most using stromx stream again). idea use boost.asio [2].
think these ideas? happy comments, suggestions etc!
cheers,
matz.
[1] http://www.tornadoweb.org/
[2] http://www.boost.org/doc/libs/1_54_0/do ... _asio.html
during last 2 years developed first (probably beta-quality) version of open-source project "stromx".
http://www.stromx.org
consists of graphical editor "stromx-studio" allows assemble image processing pipelines (= streams in stromx terminology). these streams consists of image operators linked threads. streams can executed in stromx-studio or in other application links against stromx library.
thus, raspberry pi camera module used smart camera in following way: first build vision pipeline on pc in stromx-studio , save file. transfer file raspberry pi , execute stream in slim, gui-less command line tool calls couple of functions of stromx library. in next step parameters of vision stream remotely configured via web-interface.
have far:
- stromx runtime library , stromx-studio known compile , build on raspberry pi. however, not necessary run stromx-studio on raspberry pi if stuck procedure above.
- python bindings stromx libraries exist. simple application executes stromx streams can implemented python script.
# doing:
- adding many more vision operations integrating large parts of opencv library.
# cool in future:
- precompiled packages raspbian. ubuntu packages (x86) stromx build on regular basis should not complicated.
- implement web-interface served on raspberry pi , allows remote clients control vision stream. again, python bindings available tornado [1] + stromx solution option.
- implement end-point operator in stromx stream allows transfer result data smart camera (i.e. raspberry pi) remote client collects data , processes further (most using stromx stream again). idea use boost.asio [2].
think these ideas? happy comments, suggestions etc!
cheers,
matz.
[1] http://www.tornadoweb.org/
[2] http://www.boost.org/doc/libs/1_54_0/do ... _asio.html
raspberrypi
Comments
Post a Comment