• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 97 subscribers
  • Views 477 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

PROCESSOR-SDK-J721E: Clarification required to get Video pipeline form Thermal camera using Vision apps Single Camera Application

Kaviraj S
Intellectual 320 points

Hi,

We are using a Thermal camera that will give 8-bit Monochrome as output, Our Target is to get a video pipeline from the camera using the Single-Camera application, the flow diagram is attached above.

At this stage we are trying to get the image from the sensor and show it in the display,In the above diagram the CSI2-Rx acts as a interface between the VISS node and the Fusion board, we need the following clarifications:  

  1. Does the CSI2-RX support 8 bit Monochrome as input from the camera and provide the 8-bit Monochrome output?
  2. According to the source code of the Single camera application the VISS node process, the RAW data from the camera and converts it into YUV format for display. Since the camera sensor is Thermal( monochrome) and we don't need YUV (color) output, can we bypass the VISS node and give the Monochrome output of CSI2 to DSS(Display Subsystem) node?..  (or) Should we retain the VISS node and eliminate LDC and MSC?   
  3. Is there an alternative approach ?.

Our goal is to setup the video pipeline.

Thanks,

Kaviraj

  • 0
    TI__Guru**** 501795 points

    Hi Kaviraj,

    Please find answers to your questions below,

    1, Yes, CSIRX can capture 8bit RAW data, it can be monochrome data.  As such, display supports YUV and RGB formats, but we have a way to display RAW8 data in the display node. So you could directly connect capture node to the display for the RAW8 data format.

    2, Yes, VISS and LDC can be bypassed and capture node output can be directly connected to the display. You could refer to the example ti-processor-sdk-rtos-j721e-evm-07_01_00_11\tiovx\kernels_j7\hwa\test\test_capture_display.c\. This example does exactly what you are looking for. It connects capture to the display and displays RAW data. Can you refer to it and make similar changes in single cam example or just move and build this example in vision apps?

    Regards,

    Brijesh

  • 0 in reply to Brijesh Jadav
    Intellectual 320 points

    Hi Brijesh,

    Thanks for the reply, we are facing one issue

    We have done modifications in the 'single-camera application' available in Vision App and built it. Then we followed the steps given in the link below to copy the executables from RTOS SDK to the SD card :

    https://software-dl.ti.com/jacinto7/esd/processor-sdk-rtos-jacinto7/latest/exports/docs/vision_apps/docs/user_guide/ENVIRONMENT_SETUP.html

    After copying, the external HW device (deserializer) which was connected to I2C channel 9 was not accessible. 
       
    As per the document, it copies uEnv.txt in boot partition to select .dtbo which is required to run the Vision App demo.
    Not clear why copying should change the previous functionality of deserializer access through I2C Channel 9.
    Request your help..
    Thanks, 
    Kaviraj
  • 0 in reply to Kaviraj S
    TI__Guru**** 501795 points

    Hi Kaviraj,

    Yes, we need to copy vision apps dtb file, in order to run vision apps usecase.

    But from your reply, it seems you are using i2c instance 9, please confirm the same.

    By default, in vision apps, we use i2c instance 6 for configuring sensors and serdes. If you are using i2c instance 9, we need to make changes in the vision apps dtb file for this instance and then we should be able to access i2c9 from r5f.

    Rgds,

    Brijesh