ViduSdk

1. System Requirement

Ubuntu 20.04 or later Ubuntu LTS version x86-64bit system; For Windows system, please checkout windows branch.

2. Installation

You could install ViduSdk using package manager, execute sudo dpkg -i Vidu_SDK_Api_x_x(_x)_amd64.deb or sudo apt install ./Vidu_SDK_Api_x_x(_x)_amd64.deb.
When you want to uninstall ViduSdk package, just execute sudo dpkg -r vidusdkapi or sudo apt remove vidusdkapi.

If you want to obtain accurate metadata, linux driver patch is required, please execute the following commands:
Before you execute, you need to disconnect the connected camera.

./camdriver/patch-okulo-ubuntu-lts-hwe.sh

PS: If your system has been updated, it needs to be patched again.

3. Start

The ViduSdk provides the following executable tools for you to have a quick taste of the functionality of Okulo P1 Camera and Okulo C1 Camera:

Then you could execute vidu_viewer or click Ubuntu Applications Vidu viewer to run vidu_viewer

• vidu_viewer
  • show the origin camera stream, point cloud and control the video stream;

Run python demo, you could execute python3 python/vidu_test.py

• python/vidu_test.py
  • python example, which invokes the SDKpythonLib/pyvidu.cpython-xx-x86_64-linux-gnu.so.

4. Development

Below we briefly introduce some important folders to the developers. The comprehensive guidance on the usage of the software development for the Okulo camera is here

• Vidu_SDK_Api/SDKcode/GenTL provides C headers by which the developer can directly invoke the Vidu_SDK_Api/SDKlib/libDYVGenTL.cti using C interface according to the GenTL standard.
• Vidu_SDK_Api/SDKcode/GenTLwrapper provides a C++ wrapper of the Vidu_SDK_Api/SDKlib/libDYVGenTL.cti, which makes the coding with okulo camera more conciser and C++ tone.
• Vidu_SDK_Api/SDKcode/common provides some debug or timing code fragment.
• example provides examples on how to get data from the camera. Please execute the following commands to compile those examples on your machine:

You can build example in root directory of SDK package(need opencv-dev), you can install by sudo apt-get -y install libopencv-dev python3-opencv.

cd ViduSdk/
mkdir build && cd build
cmake ../
make -j
./example/streamShow/streamShow

Also you can build your code or move the example outside of the package by adding

find_package(OpenCV REQUIRED)
find_package(Vidu_SDK_Api REQUIRED)

and

target_link_libraries(${PROJECT_NAME} Vidu_SDK_Api::GenTLwrapper)

in the CMakeLists.txt file.

5. Addendum

If unable to install or do not want to install the deb package, you can follow the following steps to configure it:

• execute bash install.sh install dependencies, which also sets the environment variable DYVCAM_GENTL64_PATH
• execute source ~/.bashrc make environment variables effective

Now, you can use vidu_viewer to open the camera, if you also want to use Python libraries, you also need to specify the PYTHONPATH environment variable to the Vidu_SDK_Api/SDKLib/SDKPythonLib directory, which contains the exported Python library files.

In this case, when using find_cackage, it is necessary to specify the specific directory of Vidu_SDK_Api's cmake, for example Vidu_SDK_Api/cmake. Just like:

find_package(Vidu_SDK_Api REQUIRED HINTS $ENV{DYVCAM_GENTL64_PATH}/../cmake)