Lane Detection Optimized With GPU Coder for Jetson TX2

Hallo! In the project, we use camera of the type DFK23UP1300 and the Jetson TX2 for image processing. I want to make a self drive mini car. I think that I can use this simple:

https://de.mathworks.com/help/gpucoder/examples/lane-detection-optimized-with-gpu-coder.html?searchHighlight=GPU%20coder&s_tid=doc_srchtitle

I checked it works on Jetson TX2, but I don't know how to pretraine AlexNET to my road.

I have to use these recommendations: https://www.mathworks.com/help/nnet/ref/alexnet.html

Can I use Ground-Truth Labeling for pretraine AlexNET?

Maybe you can give advice?

Thanks a lot.

Answer 1

[Girish Venkataramani]

Hi Vasyl

We published an blog article that describes the training:

https://devblogs.nvidia.com/parallelforall/deep-learning-automated-driving-matlab/

Let me know if this helps. Girish

[Vasyl Varvolik]

I use Ground Truth Labeler on the MATLAB and I don't understand how to get redressionOutputs with a,b,c coefficients

I don't understand why the right and left lines are superimposed