3D Point Cloud Feature Extraction Tutorial for Interactive Python App Development

Подписаться 2,8 тыс.

Просмотров 5 тыс.

50% 1

This tutorial is for Python enthusiasts and 3D Innovators! We dive into the exciting world of 3D LiDAR point cloud feature extraction using Python. If you're interested in creating interactive Python Apps to handle 3D LiDAR data, then this video is for you! We'll be covering everything from Environment Setup to feature extraction and its base components, so whether you're a beginner or an experienced Python programmer, there's something here for you. Grab your favorite beverage ☕, and let's get started!
I go through 6 phases that follow the different chapters.
Download the LiDAR dataset used: drive.google.com/drive/folder...
🍿 NEXT STEPS:
Code a 3D Point Cloud Segmentation Solution with Python: • 3D Point Cloud Segment...
Finish the 3D Tutorial Series: learngeodata.eu/3d-tutorials/
Dive in Expert articles: / florentpoux
Become a 3D Data Science Expert: learngeodata.eu
🙋 FOLLOW ME
Linkedin: / florent-poux-point-cloud
Github: github.com/florentPoux
Research: scholar.google.com/citations?...
WHO AM I?
If we haven’t yet before - Hey 👋 I’m Florent, a professor-turned-entrepreneur, and I’ve somehow become the world’s most-followed 3D expert. Through my videos here on this channel and my writing, I share evidence-based strategies and tools to help you be better coders and 3D innovators.
📄 CHAPTERS
[00:00:00]: Introduction: LiDAR Point Cloud Vectorization
[00:03:09]: Download the 3D LiDAR Dataset
[00:04:25]: 3D Environment Setup
[00:06:30]: 3D Data I/O and Fundamentals (PyVista)
[00:10:55]: 3D Data Structure Creation
[00:11:11]: kD-tree for 3D Point Clouds Explained
[00:18:27]: PCA (Principal Component Analysis) for 3D Explained
[00:20:40]: Point Cloud Feature Extraction with PCA
[00:27:54]: Feature Extraction: Neighborhood Definition
[00:30:21]: Relative Feature Extraction
[00:32:00]: Conclusion on 3D Point Cloud Feature Extraction

Наука

Опубликовано:

25 июн 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 17

@user-yk3gu2kt6s 4 месяца назад

Thank you Florent for these clear explanations !

@FlorentPoux 4 месяца назад

Glad it was helpful!

@johnrogers3315 5 месяцев назад

Great tutorial, many thanks, Florent

@rahulbarman9861 26 дней назад

thank you sir

@FlorentPoux 20 дней назад

Most welcome

@user-oq2se6ek2e 3 месяца назад

Thanks for your excellent tutorial!

@FlorentPoux 3 месяца назад

Glad it was helpful! Thanks for the kind words!

@1bhogan Месяц назад

Thank you so much for all the wonderful content, Florent! I am wondering if you have any advice for handling larger datasets. For example, I am trying to handle a point cloud of around 14 million points at the moment doing some feature extraction. I often have problems with memory in Python (running locally on my laptop). I am wondering if you have any advice for overcoming those kinds of issues related to working with large datasets like batching (but in this case, wouldn't that create issues with the nearest neighbor calculation at the edges of your batches?) or cloud services?

@FlorentPoux 20 дней назад

Thanks a lot for the kind words! To handle large datasets, if you have memory error, maybe you can try working out of core with memory mapping mechanism. Numpy has some capabilities in this domain, or switching to laspy you have also the ability to load by chunks, which is very helful. And finally, you have data format like parquet that may be a bit lighter on the memory side of things. In production, I tend to use tiling and out of core processing.

@ABSOLS92 3 месяца назад

Could you please create a video or give me hints about generating photo-realistic scenes out of point clouds, something similar to Gaussian Splatting but with point cloud priors with reduced computation

@FlorentPoux 3 месяца назад

This is definitly feasible indeed!

@ymanebelahsen1222 3 месяца назад

Thank you Florent! can you please do for us a tutorial about point cloud classification with deep learning and machine learning algorithms to clarify more the process starting from point cloud generation from agisoft and doing the segmenation with DL ML algorithms

@FlorentPoux 3 месяца назад

Great suggestion! While I work on it, if you want you have courses on it here: learngeodata.eu/ (3D Deep Learning Course, and 3D Segmentation Deck with the 3D Machine Learning Short Course)