This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

SK-TDA4VM: SK-TDA4VM: Enable highest resolution of IMX219 Camera

Part Number: SK-TDA4VM

Hi All,

I am working on the edge-ai-gst demo application provided in the PSDK. while using IMX219 camera as input source for the demo application (object detection) with 1920x1080 30fps, the demo is working fine and able to view the camera image in the full HD monitor.

I am Not able to view the camera capture image with Camera's Highest resolution of 8MPix 15fps mode which supports in my driver.

I have shared the configuration used for changing 4k capture below,

title: "Object Detection"
log_level: 2
inputs:
input0:
source: /dev/video-rpi-cam0
format: rggb
width: 3280
height: 2464
framerate: 15

models:
model0:
model_path: /opt/model_zoo/TVM-OD-5120-ssdLite-mobDet-DSP-coco-320x320
viz_threshold: 0.6
model1:
model_path: /opt/model_zoo/TFL-OD-2020-ssdLite-mobDet-DSP-coco-320x320
viz_threshold: 0.6
model2:
model_path: /opt/model_zoo/ONR-OD-8200-yolox-nano-lite-mmdet-coco-416x416
viz_threshold: 0.6
model3:
model_path: /opt/model_zoo/ONR-OD-8420-yolox-s-lite-mmdet-widerface-640x640
viz_threshold: 0.6
outputs:
output0:
sink: kmssink
width: 1920
height: 1080
overlay-perf-type: graph

flows:
flow0: [input0,model2,output0,[0,0,1920,1080]]

Kindly need your support.

Thanks and Regards,
Dinesh kumar S

  • Hi Dinesh,

    Are you facing any error during the execution of the demo with 8M at 15 fps?

    Could you share the logs for the same?

    Regards,

    Nikhil

  • Hi Nikhil,

    I am not getting any error while running the demo but not able to visualize the video streaming in the display.

    I have attached the object detection cpp application log and the screenshot of the output.

    I  would also like to know how to enable the gstreamer debug log prints in edgeai-gst-plugins, so that i can debug the code easily. I have tried changing the main meson.build from,

    cdata.set('GST_LEVEL_DEFAULT', 'GST_LEVEL_NONE') to cdata.set('GST_LEVEL_DEFAULT', 'GST_LEVEL_INFO')

    yet not able to view the logs in debug consle.

    /resized-image/__size/320x240/__key/communityserver-discussions-components-files/791/1000017097.jpg

    Log output of object detection cpp application:

    +-----------------------------------------------------------------+
    | Object Detection |
    +-----------------------------------------------------------------+
    +-----------------------------------------------------------------+
    | Input Source: /dev/video-rpi-cam0 |
    | Model Name: ONR-OD-8200-yolox-nano-lite-mmdet-coco-416x416 |
    | Model Type: detection |
    +-----------------------------------------------------------------+
    | dl-inference : 4.03 ms from 1622 samples |
    | total time : 32.90 ms from 1621 samples |
    | framerate : 30.50 fps from 1621 samples |
    +-----------------------------------------------------------------+
    581.284120 s: VX_ZONE_INIT:[tivxHostDeInitLocal:110] De-Initialization Done for HOST !!!
    581.288727 s: VX_ZONE_INIT:[tivxDeInitLocal:193] De-Initialization Done !!!
    APP: Deinit ... !!!
    REMOTE_SERVICE: Deinit ... !!!
    REMOTE_SERVICE: Deinit ... Done !!!
    IPC: Deinit ... !!!
    IPC: DeInit ... Done !!!
    MEM: Deinit ... !!!
    DDR_SHARED_MEM: Alloc's: 91 alloc's of 217709047 bytes
    DDR_SHARED_MEM: Free's : 91 free's of 217709047 bytes
    DDR_SHARED_MEM: Open's : 0 allocs of 0 bytes
    MEM: Deinit ... Done !!!
    APP: Deinit ... Done !!!

    Regards,

    Dinesh kumar s

  • Hello Dinesh,

    Have you changed the format for imx219 in setup_cameras.sh script in /opt/edgeai-gst-apps/scripts/ directory? For 4K (3280x2464) resolution, make the following change to setup_imx219():

    #IMX219_CAM_FMT='[fmt:SRGGB10_1X10/1640x1232]'
     IMX219_CAM_FMT='[fmt:SRGGB8_1X8/3280x2464]'

    Run the script after making this change and try to run the app again.

    To enable GStreamer debug messages, add GST-DEBUG=3 to the pipeline generated and output to terminal from running your application. More information about debugging on GStreamer can be found here: Basic tutorial 11: Debugging tools (gstreamer.freedesktop.org)

    Thanks,

    Fabiana

  • Hi Fabiana,

    Thanks for the reply.

    I have changed the setup camera script as mentioned and got worked but the image is looking as fish eye lens capture. Need to correct the lens distortion values. 

    How to adjust the camera LDC values in the script. Kindly need your support.

    I have attached the screenshots for your reference.

    Thanks,

    Dinesh

  • Hi Dinesh,

    Could you change your configuration to the following and try again?

    inputs:
    input0:
    source: /dev/video-rpi-cam0
    format: rggb
    width: 3280
    height: 2464
    framerate: 15
    subdev-id: /dev/v4l-rpi-subdev0

    Thank you,

    Fabiana

  • Hi Fabiana,

    I have tried the above configuration but the issue still persists. Have tried with enabling the ldc flag as below, but the dcc_ldc.bin file is not present in /opt/imaging/ directory.

    inputs:
    input0:
    source: /dev/video-rpi-cam0
    format: rggb
    width: 3280
    height: 2464
    framerate: 15

    ldc: true
    subdev-id: /dev/v4l-rpi-subdev0

    Regards,

    Dinesh kumar S

  • Hi Dinesh,

    Could you share the output of ./init_script as shown here: 5. Configuring applications — Processor SDK Linux for SK-TDA4VM Documentation ?

    Thank you,

    Fabiana