Fixing the Tensorflow error: could not load dynamic library 'libcudart.so.11.0'

I tried to install and run Tensorflow on a Ubuntu 20.04 laptop with a Nvidia GPU but I encountered the "could not load dynamic library 'libcudart.so.11.0'" error message, as shown in the screenshot below.

To resolve the issue, I had to install the Nvidia kernel and Cuda 11 libraries from the Nvidia repository. The steps are outlined below.

1. On the Ubuntu machine, open a Terminal. Type in the following commands to add the Nvidia ppa repository:
   
   $ wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/cuda-ubuntu2004.pin
   
   $ sudo mv cuda-ubuntu2004.pin /etc/apt/preferences.d/cuda-repository-pin-600 && sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/7fa2af80.pub
   
   $ sudo add-apt-repository "deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/ /"
   
   
2. In the Terminal, type in the commands to install the Nvidia kernel.
   
   $ sudo apt-get update && sudo apt-get install -y nvidia-kernel-source-460
   
   
3. Finally, install Cuda with the following command.
   
   $ sudo apt-get -y install cuda
   
   

Subsequently when importing the Tensorflow library, the error message no longer appears.

Note: Download and install any additional missing libraries from https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/ if necessary.

Export Velodyne PCAP to LAS files in VeloView

VeloView is an open source application for working with Velodyne's popular LiDAR sensors (HDL-64E, HDL-32E, VLP-16 etc.). It is available from this website https://www.paraview.org/veloview/.

VeloView has an export to LAS file format command but you would not know it as it is not available from the graphical user interface. To generate LAS files, it is necessary to use the Python command line interface. The steps below show how it can be done.

1. Run VeloView and load a Velodyne PCAP file.
   
   
2. Choose Tools | Python Console.
   
   The Python Console appears.
   
3. To export out the current frame, type in the following command at the Python prompt.
   
   >>> vv.saveLASCurrentFrame('c:/path/to/output.las', 0)
   
   Note 1: the last argument indicates the coordinate transformation mode. 0 = raw, 1 = absolute, 2 = relative.
   Note 2: Replace c:/path/to/ with your actual file system path
   
4. To export a range of frames by frame numbers, type in the following command at the Python prompt.
   
   >>> vv.saveLASFrames('c:/path/to/output.las', 10, 20, 0)
   
   Note: in this example, 10 is the starting frame, 20 is the ending frame, and 0 is the raw transformation mode.
   
5. To export a list of frames by time steps, type in the following command at the Python prompt.
   
   >>> vv.saveLAS('c:/path/to/output/las', timesteps, 0)
   
   Note: timesteps is an array of integers specifying the time steps e.g. [0,1,2,...10] and 0 is the raw transformation mode.
   
   To get the available time steps in the PCAP file, type in the following command at the Python prompt.
   
   >>> timesteps = vv.getCurrentTimesteps()
   
6. Optional. Display the resultant LAS file in a viewer.
   
   

Resolving Python Module 'numpy' has no 'xxxx' member error message in Visual Studio Code

While writing some Python code in Visual Studio Code that calls some Numpy classes, Pylint error messages appear complaining about non-existing Numpy members, e.g. the log2 method, and red wavy lines underline the offending code, as shown in the screenshot below.


One way to clear this up is to get Pylint to white-list Numpy using Visual Studio Code's Settings.

1. In Visual Studio Code, select File | Preferences | Settings.
   
   The settings.json file is displayed in the editor.
   
   
2. Click the Workspace Settings tab.
   
   The settings appear in the editor.
   
   
3. In the editor, type in the following and save the file.
   
   { "python.linting.pylintArgs" : [ "--extension-pkg-whitelist=numpy" ] }
   
   
   
   
4. Close and reopen the folder in Visual Studio Code.
   
   The missing member messages and the red wavy underlines no longer appear.
   

Resolving the ImportError: No module named paraview.simple when running VeloView

VeloView is an open source viewer for displaying point cloud data from a Velodyne laser sensor, you know the one used in driverless Google cars. After downloading and compiling the VeloView source code on an Ubuntu computer, I encountered some error messages when running the VeloView application, as shown below.
"ImportError: No module named paraview.simple"

Even though the application's main window appears, none of the commands work. And the reference grid is not displayed in the 3D view.

I figured the problem has something to do with Python being unable to locate required modules, and I added all the locations of all the other VeloView compiled packages - Paraview, VeloView, VTK into the PYTHONPATH environment variable, as shown below.

1. Open up the .profile file in the user's home directory in a text editor, e.g.
   
   $ cd
   $ vi .profile
2. Add in the following line.
   
   export PYTHONPATH=$PYTHONPATH:/usr/local/lib/paraview-4.3/site-packages:/usr/local/lib/veloview-3.1/site-packages:/usr/local/lib/paraview-4.3/site-packages/vtk
   
   A part of the .profile file is shown in the screenshot below.
   
   
   
3. Save and close the .profile file.

Now when the VeloView application is run, the ImportError message no longer appears.