I am currently tinkering around with a real-time computer vision project for vehicles and visualization is powered by Babylon.JS within an Electron App. Communication of data via TCP sockets. All is still a bit raw, but work in progress ; ).
The algo is a semantic segmentation with a simple 3D reconstruction based on a flat world assumption. Also camera calibration is done manually for this recording. So might not work too well when you put the camera in your car :).
Here are some resources for the interested:
Visualization: GitHub - j-o-d-o/OpenSense-Visu: Visualization for automotive sensor detections
C++ “Backend”: GitHub - j-o-d-o/OpenSense-App: C++ app to run automotive computer vision algos
Tensorflow Models: GitHub - j-o-d-o/computer-vision-models: Various computer vision models for Tensorflow