Inside the KPU-Net Workflow, Part 3 When Training Breaks, Lies, and Teaches You What Actually Matters In the prior post, I walked through the early stages of training an AI model to identify ground points in LiDAR across Virginia. The pipeline was running, the model was learning, and the metrics were moving in what looked … Read more
This is a follow up to my prior post on AI LiDAR Ground Segmentation. If you have not read it, go here. I want a well trained model and so I will be feeding it constantly. To start we are training on the Commonwealth of Virginia. I need data so I wrote a script that … Read more
LiDAR is one of the most powerful tools in GIS, Surveying, & Mapping, but one task has always been tricky: Separating ground from everything else.Trees, cars, buildings, utilities, and clutter, LiDAR sees it all, and so do we. When training AI ground segmentation LiDAR models, traditional methods like PMF or cloth simulation still struggle in … Read more