Deep learning function for LiDAR and point cloud data.

I want to use deep learning to find people using a 3D lidar. The goal is to notify me if people are detected in live data comming from a SICK MRS6000 with ROS.

In the video "training and validating detectors" we see the presenters use the functrion "TrainACFobjectdetector" and it works great for the video data. I have used the MATLAB ground truth labeler on point cloud sequence data and with the "Point Cloud Temporal interpolator" automation algorithm, i can track my objects and it works extreamly well. I have recorded more data and i want to use deep learning to tell me if it's a person in the data.

How can i train AI on point cloud data?

Is it any function that is as easy to use?

Any help is appreciated!

Answer 1

[Srivardhan Gadila]

You can refer to the following: Getting Started with Point Clouds Using Deep Learning, Lidar Labeling, Segmentation, and Detection and Examples of Labeling, Segmentation, and Detection on Lidar.