Generating smaller point cloud dataset from a big point cloud dataset via subsampling

I planned to develop a machine learning program which can perform tree segmentation task. However, currently I have only one point cloud data of a big area oil palm terrain. I want to find someone to help generate smaller datasets by subsampling from this big area oil palm terrain. In other words, divide the big point cloud data into smaller point cloud data.