search

TheImagingSource / tiscamera

The Linux SDK for The Imaging Source cameras.
https://www.theimagingsource.com
Apache License 2.0
300 stars 148 forks source link

Fresh Ubuntu install - tcam-capture not working (or any other code) #246

Closed rrittgarn closed 5 years ago

rrittgarn commented 5 years ago

Hello, working with a DMK 33GX178e that works properly from windows, however linux code is not working.

Installed a fresh image of Ubuntu 18.04 on my laptop, installed required packages from readme. Installed OpenCV from source with extras and when running the CPP example Saving images using OpenCV I get these errors:

https://pastebin.com/z8WzZfup

I have manually set the GST Plugins directory (which resolved errors about gst binaries not found). Attempted both packaged version (per README) and built GStreamer from source. I am getting issues with properties being read from the camera in all applications, including tcam capture.

tcam-capture errors: https://pastebin.com/WchYC7uK

Guessing there is a missing dependency or required version of something missing from documentation as I had this issue before re-installing Ubuntu.

TIS-Stefan commented 5 years ago

Hello Ryan

I am sorry for asking, but did you follow the the building steps documented on our Readme at https://github.com/TheImagingSource/tiscamera?

Stefan

rrittgarn commented 5 years ago

Yes, as stated in original post, I used the readme. I copied and pasted all required packages. I believe that list is incomplete because I know for some of the examples you require GTK-3.0, however that library is not included in your required packages.

Please let me know what information i can provide for further troubleshooting. Again, this was a fresh install of Ubuntu that I copied and pasted the instructions from the readme and it does not work.

Thanks

TIS-Stefan commented 5 years ago

Ryan

one test please Does gst-launch-1.0 tcambin ! videoconvert ! ximagesink

work on the terminal?

Stefan

rrittgarn commented 5 years ago

No, i get a blank output with the following messages in console:

$ gst-launch-1.0 tcambin ! videoconvert ! ximagesink Setting pipeline to PAUSED ... Pipeline is live and does not need PREROLL ... INFO: Working with src caps: video/x-raw, framerate=(fraction)19/2, width=(int)3072, height=(int)2048, format=(string)GRAY16_LE Setting pipeline to PLAYING ... New clock: GstSystemClock ERROR: from element /GstPipeline:pipeline0/GstTcamBin:tcambin0/GstTcamSrc:tcambin-source: Device lost Additional debug info: /home/ryan/projects/tiscamera/src/gstreamer-1.0/gsttcamsrc.cpp(1045): gst_tcam_src_device_lost_callback (): /GstPipeline:pipeline0/GstTcamBin:tcambin0/GstTcamSrc:tcambin-source Execution ended after 0:00:21.265382708 Setting pipeline to PAUSED ... Setting pipeline to READY ... Setting pipeline to NULL ... Freeing pipeline ...

TIS-Stefan commented 5 years ago

Which computer model do you use? Does the green LED light on the camera?

rrittgarn commented 5 years ago

The Camera works from windows using your IC Capture app. It is a GigE camera, so the PC is irrelevant.

It is an XPS13 Developer edition running Ubuntu 18.04.

TIS-Stefan commented 5 years ago

Please excuse me asking. I asked, because the reported error happens, if the camera is connected to a 100mBit network. Therefore, I asked for the green LED, that indicates 1GB speed.

rrittgarn commented 5 years ago

Yes, it is plugged into a juniper EX-2200-24P

It shows green currently.

TIS-Stefan commented 5 years ago

ok, ruled one possible problem out.

You should not get a device lost, when starting the live video, obviously, even if 16 bit video formats are used. So next step....

gst-launch-1.0 tcambin ! video/x-raw, format=BGRx, width=640, height=480, framerate=30/1 ! videoconvert ! ximagesink Makes a small colored video format. I am not sure, whether BGRx can be exchanged against GRAY8.

Please let me know the result.

Sorry for all this guessing around.

rrittgarn commented 5 years ago

small blank box - no image.

Here's the console output: https://pastebin.com/ivubGQFi

Thanks

rrittgarn commented 5 years ago

I was able to get a C++ program to at least give me additional information:

double property 'DeviceTemperature': value = 58.500000 default = 58.500000 min = -179769313486231570814527423731704356798070567525844996598917476803157260780028538760589558632766878171540458953514382464234321326889464182768467546703537516986049910576551282076245490090389328944075868508455133942304583236903222948165808559332123348274797826204144723168738177180919299881250404026184124858368.000000 max = 179769313486231570814527423731704356798070567525844996598917476803157260780028538760589558632766878171540458953514382464234321326889464182768467546703537516986049910576551282076245490090389328944075868508455133942304583236903222948165808559332123348274797826204144723168738177180919299881250404026184124858368.000000 step_size = 0.001000

I believe this is what tcam-capture is choking on, but I still don't have image output.

Thanks

rrittgarn commented 5 years ago

Console output on your next suggestion: https://pastebin.com/HecWpx27

Still blank box.

To verify that my gstreamer is working, I tested with my webcam and was successful with this pipeline: gst-launch-1.0 v4l2src ! videoconvert ! autovideosink

Tried replacing ximagesink with autovideosink and no luck there either.

TIS-Edgar commented 5 years ago

There seems to be a negotiation error in the tcambin. This prevents GRAY8/GRAY16_LE -> BGRx, which is something tcam-capture implicitly expects to work.

gst-launch-1.0 tcamsrc ! video/x-raw, format=GRAY16_LE, width=640, height=480, framerate=30/1 ! videoconvert ! ximagesink

This pipeline should work. If not I would require the output of 

export TCAM_LOG=DEBUG
gst-launch-1.0 --gst-debug=tcamsrc:5 tcamsrc ! video/x-raw, format=GRAY16_LE, width=640, height=480, framerate=30/1 ! videoconvert ! ximagesink

With a 33 mono camera all auto-adjustmenst are in-camera so the bin offers little/to no advantage for you. As a potential workaround you could try the development branch of the upcoming 0.11.0 release, as this issue seems to be resolved there.

rrittgarn commented 5 years ago

top pipeline did not work.

Without Sudo getting permissions errors on the gige service pid file: https://pastebin.com/YzaSPL4z

With Sudo: https://pastebin.com/wmuhYqgZ

Is there anything special I need to do to try out the development branch?

TIS-Edgar commented 5 years ago

Simply call git checkout development. Please uninstall tiscamera before doing that. It is also recommended to use a new build folder.

Can you give me your procedure for building gstreamer? All our tests show that a GRAY16 pipeline should be running, and with gstreamer installed from the repositories it definitely runs. The custom gstreamer may cause the issues, but I would like to verify that before saying anything with certainty.

rrittgarn commented 5 years ago

I actually re-installed Ubuntu again thinking the same thing. I only have Gstreamer from packages now - no custom build. OpenCV is installed and working (as far as i can tell thus far).

I did the following:

git checkout development git pull cd tiscamera/build sudo make uninstall cd .. rm -rf build mkdir build cd build cmake -DBUILD_ARAVIS=ON -DBUILD_GST_1_0=ON -DBUILD_TOOLS=ON -DBUILD_V4L2=ON -DCMAKE_INSTALL_PREFIX=/usr .. make sudo make install

then trying the gst-launch-1.0 tcamsrc ! video/x-raw, format=GRAY16_LE, width=640, height=480, framerate=30/1 ! videoconvert ! ximagesink Pipeline I am still seeing errors, but I see video for the first time in this whole lovely ordeal.

I still cannot launch tcam-capture successfully - complains of disconnect.

TIS-Edgar commented 5 years ago

Good to hear, that at least a basic stream is working,

tcam-capture is a bit aggressive when it comes to device-lost events. Those are typically triggered by errors reported by the backend such as too many missing images, missing heartbeats, etc.

A common cause may be a low mtu. You can change it with sudo ip link set dev eth0 mtu 9000 where eth0 should be replaced with your interface in question.

If that does not resolve the problems I would require a new log of the current situation.

export TCAM_LOG=DEBUG
gst-launch-1.0 --gst-debug=tcamsrc:5 tcamsrc ! video/x-raw,format=GRAY16_LE,width=640,height=480,framerate=30/1 ! videoconvert ! ximagesink

tcam-capture should give a somewhat identical log, but stop the stream after the first error.

rrittgarn commented 5 years ago

Looking into MTU, my switch is not configured for jumbo frames, and I was not aware it was a requirement of the GIGE cameras. I have verified proper MTU and am using a different USB NIC on my laptop successfully for tcam-capture only on the Development Branch though. master does not yield images.

rrittgarn commented 5 years ago

Additionally, now trying to build the latest linux examples, i am getting YAML related errors.

What version of yaml-cpp should be installed?

rrittgarn commented 5 years ago

ryan@ryan-XPS-13-9343:~/projects/Linux-tiscamera-Programming-Samples/Simple Snap Image in C++/build$ make [ 25%] Linking CXX executable simple-snapimage CMakeFiles/simple-snapimage.dir/home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp.o: In function gsttcam::TcamCamera::store_device_state(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >)': /home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp:690: undefined reference toYAML::operator<<(std::ostream&, YAML::Node const&)' CMakeFiles/simple-snapimage.dir/home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp.o: In function gsttcam::TcamCamera::restore_device_state(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >)': /home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp:697: undefined reference toYAML::LoadFile(std::cxx11::basic_string<char, std::char_traits, std::allocator > const&)' CMakeFiles/simple-snapimage.dir/home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp.o: In function YAML::detail::memory_holder::create_node()': /usr/include/yaml-cpp/node/detail/memory.h:39: undefined reference toYAML::detail::memory::create_node()' CMakeFiles/simple-snapimage.dir/home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp.o: In function YAML::detail::node_ref::mark_defined()': /usr/include/yaml-cpp/node/detail/node_ref.h:28: undefined reference toYAML::detail::node_data::mark_defined()' CMakeFiles/simple-snapimage.dir/home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp.o: In function YAML::detail::node_ref::set_null()': /usr/include/yaml-cpp/node/detail/node_ref.h:33: undefined reference toYAML::detail::node_data::set_null()' CMakeFiles/simple-snapimage.dir/home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp.o: In function YAML::detail::node_ref::set_scalar(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)': /usr/include/yaml-cpp/node/detail/node_ref.h:34: undefined reference toYAML::detail::node_data::set_scalar(std::cxx11::basic_string<char, std::char_traits, std::allocator > const&)' CMakeFiles/simple-snapimage.dir/home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp.o: In function YAML::detail::node_ref::begin()': /usr/include/yaml-cpp/node/detail/node_ref.h:43: undefined reference toYAML::detail::node_data::begin()' CMakeFiles/simple-snapimage.dir/home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp.o: In function YAML::detail::node_ref::end()': /usr/include/yaml-cpp/node/detail/node_ref.h:48: undefined reference toYAML::detail::node_data::end()' CMakeFiles/simple-snapimage.dir/home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp.o: In function YAML::Node::Scalar[abi:cxx11]() const': /usr/include/yaml-cpp/node/impl.h:152: undefined reference toYAML::detail::node_data::empty_scalar[abi:cxx11]' CMakeFiles/simple-snapimage.dir/home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp.o: In function YAML::Node::AssignData(YAML::Node const&)': /usr/include/yaml-cpp/node/impl.h:250: undefined reference toYAML::detail::memory_holder::merge(YAML::detail::memory_holder&)' CMakeFiles/simple-snapimage.dir/home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp.o: In function YAML::as_if<bool, void>::operator()() const': /usr/include/yaml-cpp/node/impl.h:116: undefined reference toYAML::convert::decode(YAML::Node const&, bool&)' CMakeFiles/simple-snapimage.dir/home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp.o: In function YAML::detail::node& YAML::detail::node_data::get<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, boost::shared_ptr<YAML::detail::memory_holder>)': /usr/include/yaml-cpp/node/detail/impl.h:112: undefined reference toYAML::detail::node_data::convert_to_map(boost::shared_ptr)' /usr/include/yaml-cpp/node/detail/impl.h:126: undefined reference to YAML::detail::node_data::insert_map_pair(YAML::detail::node&, YAML::detail::node&)' CMakeFiles/simple-snapimage.dir/home/ryan/projects/tiscamera/examples/cpp/common/tcamcamera.cpp.o: In functionYAML::detail::node& YAML::detail::node_data::convert_to_node<std::cxx11::basic_string<char, std::char_traits, std::allocator > >(std::cxx11::basic_string<char, std::char_traits, std::allocator > const&, boost::shared_ptr)': /usr/include/yaml-cpp/node/detail/impl.h:171: undefined reference to `YAML::detail::memory_holder::merge(YAML::detail::memory_holder&)' collect2: error: ld returned 1 exit status CMakeFiles/simple-snapimage.dir/build.make:161: recipe for target 'simple-snapimage' failed make[2]: [simple-snapimage] Error 1 CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/simple-snapimage.dir/all' failed make[1]: [CMakeFiles/simple-snapimage.dir/all] Error 2 Makefile:129: recipe for target 'all' failed make: *** [all] Error 2

TIS-Edgar commented 5 years ago

The package libyaml-cpp-dev should contain all required dependencies for that example.

rrittgarn commented 5 years ago

That is the package I have installed and am now getting the YAML issues and cannot compile them. Is there a bug related to this in the development branch?

TIS-Edgar commented 5 years ago

It is not. The example compiles as expected on Ubuntu 18.04. Could you try a new build directory for the example? I am currently suspecting a caching issue in cmake.

rrittgarn commented 5 years ago

Followed the same steps as above for cleaning and recreating - i cannot build the example files without a YAML related error message

TIS-Edgar commented 5 years ago

I found the problem. It was that the example failed to link against yaml-cpp. The fixes have been pushed to the development branch of the Samples repository.

cd ~/projects/Linux-tiscamera-Programming-Samples/Simple Snap Image in C++/build
git pull
git checkout development
make