- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello,
I'm trying to run one of the samples - security_barrier_camera_demo on Raspberry Pi with OpenVINO 2019.1.
The demo runs properly only with the single model specified :
/home/pi/openvino/build/armv7l/Release/security_barrier_camera_demo -i /home/pi/openvino/build/models/car_1.bmp -m /home/pi/openvino/build/models/vehicle-license-plate-detection-barrier-0106-fp16.xml -d MYRIAD
But after running it with the two or three additional models I'm getting an error :
[ INFO ] InferenceEngine:
API version ............ 1.6
Build .................. 22443
[ INFO ] Parsing input parameters
[ INFO ] Capturing video streams from the video files or loading images
[ INFO ] Files were added: 1
[ INFO ] /home/pi/openvino/build/models/car_1.bmp
[ INFO ] Number of input image files: 1
[ INFO ] Number of input video files: 0
[ INFO ] Number of input channels: 1
[ INFO ] Display resolution: 1920x1080
[ INFO ] License Plate Recognition detection DISABLED
[ INFO ] Loading plugin MYRIAD
API version ............ 1.6
Build .................. 22443
Description ....... myriadPlugin
[ INFO ] Loading plugin MYRIAD
[ ERROR ] Cannot find plugin for device: Default
The commads are as follows (created based on the Windows version) :
/home/pi/openvino/build/armv7l/Release/security_barrier_camera_demo -i /home/pi/openvino/build/models/car_1.bmp -m /home/pi/openvino/build/models/vehicle-license-plate-detection-barrier-0106-fp16.xml -m_va /home/pi/openvino/build/models/vehicle-attributes-recognition-barrier-0039-fp16.xml -d MYRIAD -d_va MYRIAD
or
/home/pi/openvino/build/armv7l/Release/security_barrier_camera_demo -i /home/pi/openvino/build/models/car_1.bmp -m /home/pi/openvino/build/models/vehicle-license-plate-detection-barrier-0106-fp16.xml -m_va /home/pi/openvino/build/models/vehicle-attributes-recognition-barrier-0039-fp16.xml -m_lpr /home/pi/openvino/build/models/license-plate-recognition-barrier-0001-fp16.xml -d MYRIAD -d_va MYRIAD -d_lpr MYRIAD
It seems like the demo continues to load the plugin multiple times and that's not the case under Windows :
[ INFO ] InferenceEngine:
API version ............ 1.6
Build .................. 22443
[ INFO ] Parsing input parameters
[ INFO ] Capturing video streams from the video files or loading images
[ INFO ] Files were added: 1
[ INFO ] F:\Program Files (x86)\IntelSWTools\openvino\deployment_tools\demo\\car_1.bmp
[ INFO ] Number of input image files: 1
[ INFO ] Number of input video files: 0
[ INFO ] Number of input channels: 1
[ INFO ] Display resolution: 1920x1080
[ INFO ] Loading plugin MYRIAD
API version ............ 1.6
Build .................. 22443
Description ....... myriadPlugin
[ INFO ] Loading network files for VehicleDetection
[ INFO ] Batch size is forced to 1
[ INFO ] Checking Vehicle Detection inputs
[ INFO ] Checking Vehicle Detection outputs
[ INFO ] Loading Vehicle Detection model to the MYRIAD plugin
[ INFO ] Loading network files for VehicleAttribs
[ INFO ] Batch size is forced to 1 for Vehicle Attribs
[ INFO ] Checking VehicleAttribs inputs
...etc
Of course it is pretty possible that I missed something from readmes or docs...
Any ideas how to run the demo with all the models?
Thanks in advance,
Tomasz Szulc
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
For those who care: after looking into cpp,hpp I found a little glitch/difference in interpreting input args/strings by Windows and RPi compilers
Running the demo with -d_va "MYRIAD" -d_lpr "MYRIAD" solves the problem.
Otherwise the "target device" for remaining options is ignored and resolves to "Default" which is treated only a placeholder in this case.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Dear Szulc, Tomasz,
Thank you for reporting back and of course we care about your discoveries ! Can you be more specific about the discrepancy you found ? According to the documentation https://docs.openvinotoolkit.org/latest/_inference_engine_samples_security_barrier_camera_demo_README.html
-m "<path>" Required. Path to the Vehicle and License Plate Detection model .xml file.
-m_va "<path>" Optional. Path to the Vehicle Attributes model .xml file.
-m_lpr "<path>" Optional. Path to the License Plate Recognition model .xml file.
You should be passing in 3 different IR *.xml files to three different -m switches (-m, -m_va and -m_lpr). And looking back at your command-lines, that's exactly what you did !
So I'm confused as to what your issue was, or what the discrepancy you found was.
Thanks for using OpenVino !
Shubha
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hello,
The problem occured when I ran the demo under Raspbian without the -d_va,-d_lpr command line parameters specified within "" characters ie : -d_va MYRIAD -d_lpr MYRIAD instead of running it like this : -d_va "MYRIAD" -d_lpr "MYRIAD".
Under Windows/(VisualStudio), where I was debugging (ie. exploring) this application there was no need for the "" charactes in the commad line for the -d,-d_va,-d_lpr options. I was probably to quick assuming that I can transfer the command line -d_va,-d_lpr parameters one to one to Raspbian.
May be it should be verified if the function CommandLineFlagParser::ParseNewCommandLineFlags (gflags.cc) works equally in both cases.
But nevertheless , now I know how to pass the command line parameters ;-)
My Best Regards
Tomasz Szulc
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Dearest Tomasz Szulc,
Thank you kindly for your explicit description of the issue. IMHO, this is a bug.CommandLineFlagParser::ParseNewCommandLineFlags should not behave differently among the different devices.
Thanks for using OpenVino !
Shubha

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page