Installation

Contents:

Install Asilla SDK Client package

We provide Asilla SDK Client package as a docker image for both of Nvidia Jetson devices and on-premise server.

1. Preparation
For Nvidia Jetson Device
System requirements
  • Jetpack version: 4.4.1, 4.5
  • Docker: by default, docker has been installed with Jetpack and docker container location is on OS partition. Our docker image requires at least 5GB free space. If your OS partition is not enough space, you should change docker container location to another partition by editing file /etc/default/docker at below line:
    DOCKER_OPTS="-g /path/to/directory"

For On-premise server
System requirements

For a comfortable execution, your system must at least meet the following requirements.

Our docker image is based on NVIDIA CUDA 10.2.89, which requires NVIDIA Driver release 440.33.01

CPU Intel or AMD
RAM At least 8GB
GPU At least 8GB
Now our built docker image support GTX 1080, GTX 1080Ti, RTX 2080, RTX 2080Ti, RTX 2080 SUPER For other GPU type, please contact us.
Storage At least 30GB
OS Ubuntu 18.04 LTS
Nvidia driver At least version 440.33.01 (This version is usually installed with CUDA Version 10.2)
Prerequisites
2. Installation steps
Step 1: Pull and run docker image

To see command of docker image pulling and get free license, please go to the download page.

Step 2: Access the docker container

Check the container ID of the asilla_sdk_client container. Then use that container ID to access asilla_sdk_client container.
sudo docker exec -it [Container Id] /bin/bash

Then, you can activate the Asilla_SDK client package on your computer.

Step 3: Activation

Run the following command in order to activate the SDK client package.
asdk activate

Then enter the Client ID, Licence ID and then press Enter.

Once you have completed this step, please select "About SDK" from contact us and inquire about ClientID and LicenseID. We will invite you to our support Slack workspace and issue a license for you

Congratulations, AsillaSDK client package activation is done successfully.

Step 4: Run AsillaSDK Client

Note that below commands should be runned in docker container.

After configuration settings, the system is ready to execute AsillaSDK client package’s applications. Use the following command for setting up cameras:
asdk set_camera_url [cam_id] [camera_url]

Parameters explanation:
  • cam_id: Camera ID. Use this API to get the available Camera ID.
  • camera_url: The camera URL of your IP camera or the device path of your USB camera. Sample camera URL:
    • rtsp://<ip_address>/media/video1 for IP camera
    • /dev/video0 for USB camera

(Optional) After setting up all cameras, run the command below to capture alert background for each camera.
asdk capture [cam_id]

(Recommended) (Jetson Only) Before running the SDK system, users are recommended to use the maximum Power Management Mode prepared by Nvidia.

  • For Jetson Xavier NX, run sudo /usr/sbin/nvpmodel -m 2
  • For Jetson Nano, run sudo /usr/sbin/nvpmodel -m 0

If you want to use more than 4 cameras, please run the command asdk add_camera [cam_id] before you run asdk set_camera_url[cam_id] [camera_url] to configure the camera.

For example, if you want to add the 4th camera, you would run asdk add_camera 4 and asdk set_camera_url 4 [camera_url]

Then simply start and stop the SDK client package apps using the following commands.
asdk run_client
asdk stop_client

Use the following command to train the abnormal behavior recognition application (Anolla) independently for each camera or for multiple cameras at the same time.
asdk train 1,2,3
Numbers 1,2,3 are the camera ids that you want to train.

After start training, you may need to confirm the status of each camera, whether a specific camera is at data collecting mode, training mode or running mode. You always can check the status of each camera through the API provided by Asilla. Call the endpoint getCameraList or getSpecificCamConfig and it will display all parameters and the values related to a certain camera. Thereafter, you may check the value of cam_status to confirm the current status of the camera.

cam_status = "Running" / "Collecting"/ "Training"

See more information about training process, please refer this sample.

Once you start the training process, you must wait for training to finish, it will be changed to "running" mode automatically. In case if you want to cancel the training process, you may stop the application using following command:
asdk stop_client

By default, Line-cross (LC) function is disabled due to the specific configuration of the observation area. You need to set up some specific configurations to enable this feature. See this sample for more information.

Step 5: View realtime output streaming

To view realtime output streaming from Asilla SDK Client, you can use streaming viewer application (example: VLC, ffplay) with below RTSP:

  • Camera 01: rtsp://{SDK_IP_ADDRESS}:8091/cam01.sdp
  • Camera 02: rtsp://{SDK_IP_ADDRESS}:8091/cam02.sdp
  • Camera 03: rtsp://{SDK_IP_ADDRESS}:8091/cam03.sdp

Step 6: Advanced configuration settings.

For advanced configuration settings, please go to API & references. Configuration settings can be done through the API provided by us.


Upgrading from the Trial plan to the Paid plan

With the free version of Asilla SDK client V1.0, you are bound to a limited number of cameras as follows:

On-Premises Server 3
Nvidia Jetson NX 3
Nvidia Jetson Nano (4GB) 1

If you need to use Asilla SDK client v1.0 with more cameras, you are kindly requested to get a paid version. Please contact us with the required number of cameras and the type of the device you need to utilize them with. You will be provided with a new license enabling you to run with additional cameras. Follow the instructions below to upgrade your Asilla SDK client version with the newly issued license,

Please stop the system first if it is running,
asdk stop_client

And re-activate using the following command,
asdk activate

Then simply add the camera with a camera ID following the command below,
asdk add_camera [cam_id]

Thereafter, set the camera urls of the cameras that you want to add,
asdk set_camera_url [cam_id] [camera_url]


Install Asilla SDK Server package

Coming soon