I am currently looking into LiDAR data, and I was wondering are there any Computer Vision only techniques for object recognition/reconstruction for this kind of data? All the methods I have looked into have some form of Machine Learning/Deep Learning involved. I am still a novice so pardon me if my question seems weird.
It's true that your question looks very weird. Maybe you're searching for the tradition approach (hand-designed feature, ransac, etc). If it's the case then pcl_tutorials (old but good) or open3d (new and good too) is your friend here. Go there and find the method you need.
To be honest, you need to have a brief knowledge of point cloud data processing first. Before lurking into the above document, I recommend reading the matlab point cloud data processing or better grabbing some book about point cloud data processing.
Good luck