Hi,
The TIDL examples written for OpenCV GUI output found at
(usr/share/ti/examples/tidl) function on the BeagleBone AI with image and
video inputs but does not function with camera input. Specifically I am
looking at the ssd_multibox demo which tidl-api documentation here states
should function with camera (live video) input (
http://downloads.ti.com/mctools/esd/docs/tidl-api/example.html)
debian@beaglebone:/usr/share/ti/examples/tidl/ssd_multibox$ sudo
./ssd_multibox -h
Usage: ssd_multibox
Will run partitioned ssd_multibox network to perform multi-objects
detection
and classification. First part of network (layersGroupId 1) runs on EVE,
second part (layersGroupId 2) runs on DSP.
Use -c to run a different segmentation network. Default is jdetnet_voc.
Optional arguments:
-c <config> Valid configs: jdetnet_voc, jdetnet
-d <number> Number of dsp cores to use
-e <number> Number of eve cores to use
-i <image> Path to the image file as input
Default are 9 frames in testvecs
* -i camera<number> Use camera as input*
* video input port: /dev/video<number>*
-i <name>.{mp4,mov,avi} Use video file as input
-l <objects_list> Path to the object classes list file
-f <number> Number of frames to process
-w <number> Output image/video width
-p <number> Output probability threshold in percentage
Default is 25 percent or higher
-v Verbose output during execution
-h Help
I am executing demo on a monitor using a C1 Pro camera (
https://www.kurokesu.com/shop/CAMERA_C1_PRO) as input which has MJPG/YUYV
(video1) and h.264 (video2) output.
debian@beaglebone:/usr/share/ti/examples/tidl/ssd_multibox$ v4l2-ctl
--list-devices
vpe (platform:vpe):
/dev/video0
Kurokesu C1 PRO SN000163: Kurok (usb-xhci-hcd.1.auto-1.4):
/dev/video1
/dev/video2
Running the demo throws the following error (tried both video1/video2 with
same output):
debian@beaglebone:/usr/share/ti/examples/tidl/ssd_multibox$ sudo
./ssd_multibox -i /dev/video1
Input: /dev/video1
terminate called after throwing an instance of 'std::out_of_range'
what(): basic_string::substr: __pos (which is 140) > this->size() (which
is 0)
I understand this demo was initially written to run on the am572x evm with
camera module but I'd like to get it running on the Beaglebone AI in lieu
of the in progress mjpg-streamer version in cloud9
Any assistance would be much appreciated.
Thanks.
Rob
--
For more options, visit http://beagleboard.org/discuss
---
You received this message because you are subscribed to the Google Groups
"BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/beagleboard/27302220-d659-4722-85ad-9fedfaa88bc3%40googlegroups.com.
