Semantic Segmentation for LiDAR Point Clouds, Applied
From unsupervised clustering to a trained PointNet model. 15 lessons that take you through the three families of segmentation: unsupervised, supervised machine learning, and deep learning.
See semantic segmentation in action
From a raw LiDAR scan to ground, buildings, vegetation, and objects, each point labeled by code you wrote.
Points per workflow. Built for real LiDAR datasets, not toy samples.
Students trained worldwide across 80 countries.
Production experience distilled into structured, repeatable workflows.
Your LiDAR scan is beautiful. And useless.
You captured millions of points. The client opens the file, rotates it once, and asks the only question that matters, “Which points are the ground? Where are the buildings? How much vegetation encroaches on the power line?” Without semantic labels, you have no answer. The scan is geometry without meaning.
Semantic segmentation is the layer that turns coordinates into answers. And there is not one way to do it, there are three: unsupervised clustering when you have no labels, supervised machine learning when you have some, and deep learning when you need scale. This course teaches you all three, and when to use each.
Most tutorials teach you a single algorithm and call it a day. Real projects do not cooperate: one dataset has labels, the next does not, the third is too large for your trained model. The engineer who masters the full decision tree, from K-Means to PointNet, is the one who never gets stuck.
What you’ll build
Three working segmentation systems on real LiDAR data.
Unsupervised segmentation engine
Label point clouds with zero training data. K-Means clustering on geometric features, applied to raw LiDAR scans.
Supervised ML workflow
The complete supervised learning loop. Feature engineering, training set design, classifier selection, and evaluation on held-out LiDAR tiles.
PointNet from scratch
Implement, train, and evaluate PointNet in PyTorch. Architecture, data preparation, training, and inference.
Detection vs. classification vs. segmentation
The conceptual map. Know exactly which problem you are solving and which family of methods applies before writing a line of code.
Reusable Python pipelines
Every lesson produces a script you can point at your own data. Data preparation, training, and inference stages cleanly separated.
System design thinking
The closing lesson steps back from algorithms to architecture. How to design a 3D deep learning system that survives contact with production data.
Semantic segmentation is where LiDAR becomes truly useful. I built this course around the classification workflows I run in production, the same decision tree I walk through every time a new dataset lands on my desk. You will not just learn algorithms, you will learn which one to reach for.
How this course works
Hands-on, Python-first, and built around real LiDAR datasets.
100% asynchronous
Access everything 24/7 on the LMS. Self-paced. No live sessions required.
Three methods, one dataset logic
Unsupervised, supervised, and deep learning approaches applied side by side, so you see exactly what each buys you.
Real LiDAR datasets
Aerial and terrestrial scans with millions of points. The messiness is part of the lesson.
Production patterns
Training set design, evaluation metrics that matter, and the inference patterns you need for scan-scale workflows.
Lifetime access
One payment, permanent access. Every future update included.
Foundation for 3D AI
Segmentation is the gateway problem of 3D machine learning. Master it here and every downstream architecture makes sense.
Anyone can copy a PointNet tutorial. The skill that gets you hired is knowing when a 30-line K-Means script beats a neural network, and when it does not. This course trains that judgment on real data, lesson after lesson.
The Curriculum
15 lessons. From segmentation fundamentals to a trained PointNet model.
Prerequisites
This course is for engineers, GIS professionals, and researchers who want to classify LiDAR point clouds with machine learning, end to end.
- Python (beginner+): comfortable with loops, functions, and basic NumPy operations
- Basic 3D knowledge: you know what a point cloud is and have opened one in CloudCompare or a similar tool
- Hardware: 16 GB RAM recommended. A GPU helps for the PointNet module but is not required
- Software: Python, NumPy, scikit-learn, PyTorch. All free and open-source
No prior machine learning experience required. The supervised and deep learning workflows are built from first principles.
The conceptual map. Detection, classification, and segmentation: three different problems that get confused daily. Know which one you are solving.
Label points with no annotations at all. K-Means clustering on geometric features, and where unsupervised methods shine.
The full supervised loop on LiDAR data. Training set design, classifier selection, evaluation, and the ML solutions that still beat deep learning on small datasets.
The heart of the course. Implement PointNet from scratch: environment, architecture, data preparation, training, and inference.
The step most courses skip. How to assemble models, data flows, and evaluation into a 3D deep learning system that holds up in production.
Your instructor
Dr. Florent Poux
I’ve spent 12+ years in 3D geospatial: from field surveys with total stations to building AI systems for Fortune 500 companies. I published the O’Reilly book on 3D Data Science with Python. I’ve advised startups valued at over 15M EUR. I’ve held a professorship, taught at university, and led R&D for some of the largest organizations in the space.
I don’t teach syntax. I teach judgment. Every module is built around real decisions I’ve faced in production. Which neural renderer fits an industrial inspection job. How to architect a semantic pipeline that doesn’t choke on 500M points. When to use algorithmic methods and when to switch to deep learning.
What students say
Engineers, GIS professionals, and researchers from 80 countries.
“Our team processed 200M points for a highway survey. Before this course, we were stuck at 10M with crashes. The memory management module saved us weeks of work.”
“I’m a surveyor with 20 years of field experience. This gave me the Python and AI skills to modernize our entire workflow. Best investment I’ve made in my career.”
“RANSAC plus DBSCAN on a 120M-point mining dataset — segmented, volumetric change detected, and report ready in one afternoon. That used to take a week.”
“This course is the foundation I wish I had when I started with LiDAR. The feature extraction module alone reshaped how I approach every project.”
“I came in fluent in Python but lost on 3D. After Module 2 I was reading PointNet++ papers like a normal human. The intuition Florent builds is the missing link.”
“We replaced a brittle classical pipeline with a sparse-conv network thanks to this course. Production accuracy jumped from 78 to 94 percent on the same data.”
“The training loop debugging chapter was worth the price by itself. I now know why my models stop learning instead of guessing at hyperparameters.”
Get lifetime access
One payment. Every lesson, every update, every segmentation workflow.
Applied Semantic Segmentation
Complete segmentation curriculum + Python source code + real LiDAR datasets + lifetime updates
- 15 lessons (12+ hours, 5 modules)i
- Complete Python source code + datasets
- Unsupervised + supervised + deep learning workflows
- PointNet implementation in PyTorch
- Lifetime access + all future updatesi
- 90-day results guaranteei
Zero-risk guarantee: If you don’t see real results within 90 days, I’ll refund you in full. No questions.
The complete ecosystem
3D AI Architect Program
The complete spatial AI curriculum, delivered in 3 tiers. Pick the depth that matches where you are — Foundations to get moving, Professional for the full OS stack, Ultimate for live access and priority support.
- 3D AI Acceleratori: 17 episodes in 6 acts
- 3D Course Libraryi: 24+ standalone courses
- All 4 OS courses (Professional & Ultimate tiers)
- Neurones 3D software access
- Monthly drop-in sessions with Dr. Poux (Ultimate)
- Spatial AI job and market intel
- Priority support + services access (Ultimate)
- 300+ hours of content
What you’re getting access to
Everything I’ve built over 12+ years, from land surveying in the field to advising 15M EUR startups, compressed into one curriculum you can start today. Delivered by the first QUALIOPI-certified 3D geospatial academy.
Every pipeline was battle-tested on Fortune 500 projects processing billions of points. You’re getting the real playbook, not theory.
Methods validated by peer-reviewed publications, the ISPRS scientific community, and 1,500+ academic citations. Not guesswork.
Built by someone who surveyed in the field, defended a PhD, advised funded startups, and shipped products to Fortune 500 clients.
I share more free content than most people put behind a paywall. That’s intentional. I want you to know exactly what you’re getting before you invest. This course is the concentrated, structured version of everything I know. No fluff. No filler. Just the production path.
Find the right path for you
From single courses to the complete ecosystem.
| Feature | Standalone Course | Applied Semantic Segmentation | Course Library | 3D AI Architecti | Enterprise |
|---|---|---|---|---|---|
| Courses included | 1 topic | 5 modules | Full catalogi | 3 OS courses + Library (tiered) | Custom |
| Hours of content | 2-8h | 12+ hours | 150+ hours | 300+ hours (tiered) | Custom |
| Production source code | ✓ | ✓ | ✓ | ✓ | ✓ |
| Lifetime access | ✓ | ✓ | – | ✓ | ✓ |
| 3D AI Accelerator Tracki | – | – | – | ✓ | ✓ |
| Neurones 3D softwarei | – | – | – | ✓ | ✓ |
| Spatial AI job & market inteli | – | – | – | ✓ | ✓ |
| Monthly drop-in sessionsi | – | – | – | ✓ | ✓ |
| Priority support + services accessi | – | – | – | ✓ tiered | ✓ |
| Custom onboardingi | – | – | – | – | ✓ |
| Team licensing | – | – | – | – | ✓ |
| Price | €97 – €497 | €297 | €1,297 | Starts at €1,999 | On request |
Straight answers
Do I need prior machine learning experience?
No. The course builds the supervised and deep learning workflows from first principles. You need Python basics and a rough idea of what a point cloud is.
What software do I need?
Python, NumPy, scikit-learn, and PyTorch. All free and open-source. No paid licenses required.
Do I need a GPU?
Not strictly. The unsupervised and supervised modules run on any laptop. For the PointNet module a GPU speeds up training, but the datasets are sized so a CPU still works.
How long do I have access?
Lifetime. One payment, permanent access. Every future update is included.
What’s the refund policy?
90 days. Segment your own LiDAR data with the workflows. If you don’t see results, email me for a full refund.
How is this different from the 3D Deep Learning Foundations course?
3D Deep Learning Foundations (35411) goes deeper on the neural network side. This course covers the full segmentation decision tree, unsupervised and supervised machine learning included, with PointNet as the deep learning capstone. If you want the complete method map for LiDAR classification, start here.
Will this work on my own LiDAR data?
Yes. Every workflow is written to be re-pointed at your own LAS or PLY files. Adapting the pipelines to your data is part of the intended path.
Can I upgrade to an Operating System later?
Yes. When you are ready for the full production stack, contact me for credit toward the 3D Deep Learning OS or the Spatial OS.
Not sure if this course fits?
If you have specific questions about how the curriculum applies to your role, your team’s needs, or your technical background, I’m happy to help you figure it out before you commit.
Book a 15-min call