Kneron PLUS - Python (Establishing device connection)

Gopinath

Hi,

What is the best way to establish connection to the hardware? I have tried Kneron PLUS - Python software with USB connection. However, the device is not recognized by the host.

Steps followed:

1. connected 5V DC Jack
2. Turned on the power btn & the reset btn
3. connected USB to host and tried scanning the device.

Am I missing something?

Scan Output :

scanning kneron devices ...

number of Kneron devices found: 0

Maria Chen

Hi Gopinath,

Could you specify what OS you're using? If you were using Windows, you would need to install Zadig, and if you were using Ubuntu/Raspberry Pi, you would need to configure USB permission.

For more information in regards to installing, you could refer to this link: Install Dependency - Document Center (kneron.com)

Gopinath

Hi Maria,

I am using windows. Yes, I have installed Zadig as well. It does not recognize the KL520 device.

Maria Chen

Hi Gopinath,

Thank you for answering the questions. So even if you clicked the Options tab -> List All Devices in Zadig, it wouldn't come up as "Kneron KL520"? Is the USB dongle detected in your Device Manager?

Gopinath

Thanks, Maria! I was able to connect the device successfully.

Could you please point me to a document or an example that explains the usage of camera on KL520 for real time inference and result processing?

Also, is it possible to post process the inference result on chip and display the output on screen?

Thank you.

Maria Chen

Hi Gopinath,

I'm glad I could help! And sure, you could download Kneron PLUS from here Developers | Kneron – Full Stack Edge AI to do inferencing with your camera on your laptop, or with a USB camera connected to your computer.

If you use C, you could refer to this part of the documentation: Run Examples - Document Center (kneron.com)

If you use Python, you could refer to this part: Inference with Drop Frame - Document Center (kneron.com)

There's also a section with Kneron APIs to see how to use these functions: kp.inference - Document Center (kneron.com)

Please feel free to ask us if you need anything else.

Gopinath

Hi Maria,

I was able to get the inference from NPU with the help of Kneron PLUS.

However, KL520 SOC comes with a camera set up and a display screen. We want to use that camera to process real world images and display the inference on the screen. How do we program to make use of the camera on KL520 soc and post -process the results on device using it's on board cpu?

Could you share any document or links that would help us use achieve this?

Thanks!

Maria Chen

Hi Gopinath,

Are you talking about the KL520 series AI SoC Developement Kit? If so, there is an Assembly Instruction inside.

The board should be set up like the photo below. If this wasn't what you were asking about, please let us know what KL520 SoC you've got.

After setting the board up, you could refer to this link to load the firmware and model as a bin file onto the board's chip: Appendix - Document Center (kneron.com)

To load the firmware and model, you would need the KL520 SDK, which you could download from here: Developers | Kneron – Full Stack Edge AI

Inside the KL520 SDK example_projects, you could use tiny_yolo_v3_host_usbout to do inferencing with the KL520 SoC Development Kit's camera.

Gopinath

Hi Maria,

Thanks for your response!

Is there a python example for testing a customized model that is utilizing both camera and display board on KL520 series AI SoC Developement Kit?

Maria Chen

Hi Gopinath,

No problem!

I don't think there is. The examples for AI SoC Development Kit only uses the Tiny Yolo V3 model.

What you could do is use Kneron PLUS' Generic Inference example in Python to inference with your own model: Run Examples - Document Center (kneron.com) / Generic Inference - Document Center (kneron.com)

Then, you could refer to KL520 SDK's customized inference when you put your own model's post-process function inside the NCPU. You would need to edit the KL520 SDK firmware (SCPU and NCPU) by using Keil, which you'd need to purchase a license to build the firmware: Introduction - Document Center (kneron.com) / v2.0.x - Document Center (kneron.com)

When you've confirmed that the above flow work on Kneron PLUS, you could take a look at the tiny_yolo_v3_host example in the KL520 SDK to figure out how to control the screen and camera, then edit the tiny_yolo_v3_host example to fit your model in.