skills.rest
NVIDIA-Omniverse
NVIDIA-OmniverseOFFICIAL
Explore all skills by this author →
Website𝕏@nvidiaomniverseGitHub

reading-sensor-pointclouds

Read lidar and radar PointCloud tensors and map them to CPU or CUDA memory.

201|25|Updated Dec 18, 2025
Search Tags:#mapping#tensor#cuda#radar#lidar#pointcloud#cpu-memory
One-click install
npx skills add https://github.com/NVIDIA-Omniverse/ovrtx --skill reading-sensor-pointclouds
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: reading-sensor-pointclouds
Source: https://github.com/NVIDIA-Omniverse/ovrtx/tree/main/.agents/skills/reading-sensor-pointclouds
Command: npx skills add https://github.com/NVIDIA-Omniverse/ovrtx --skill reading-sensor-pointclouds

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Access and read PointCloud outputs from lidar or radar sensors, mapping composite render-var tensors such as Coordinates, Counts, Intensity, RCS, RadialVelocityMs, and TimeOffsetNs to usable memory spaces, and handling per-point validity and memory lifetime. For channel meanings and units, see the related interpreting-lidar-pointclouds and interpreting-radar-pointclouds skills.

Core Features & Use Cases

  • Read PointCloud channels (Coordinates, Counts, Intensity, TimeOffsetNs) and map them to CPU or CUDA memory.
  • Slice valid entries using Counts and Flags and handle per-point validity for downstream processing or visualization.
  • Use with Python or C/C++ examples to access and interpret PointCloud data in sensor pipelines.

Quick Start

Load a lidar PointCloud, map it to CPU memory, and print coordinates and counts.

Frequently Asked Questions about reading-sensor-pointclouds

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I read lidar and radar PointCloud data and map it to CPU or CUDA memory?▼

To read PointCloud data, you identify composite render-var tensors from lidar and radar sensors and map per-point channels like Coordinates, Counts, and TimeOffsetNs directly to CPU or CUDA memory spaces for analysis.

How do I handle per-point validity when processing lidar PointCloud tensors?▼

You handle per-point validity by slicing valid entries using the Counts and Flags tensors, ensuring only valid data passes to downstream processing or visualization tasks.

Can I access radar PointCloud channels like RadialVelocityMs and RCS in Python and C/C++ workflows?▼

Yes, you can access radar and lidar PointCloud channels including RadialVelocityMs, RCS, Intensity, and TimeOffsetNs using provided Python and C/C++ examples in your sensor pipelines.

What is the best way to interpret cross-sensor channel semantics for PointCloud data?▼

Interpreting PointCloud channel semantics requires reading composite render-var tensors and understanding per-point channel meanings, often referencing related interpreting-lidar-pointclouds and interpreting-radar-pointclouds skills for unit details.

Why do I need to manage memory lifetime when reading PointCloud outputs from sensors?▼

Managing memory lifetime is necessary because PointCloud data mapped to CPU or CUDA memory must remain valid during downstream processing or visualization to prevent access violations and data corruption.

Related Skills

Explore More Software Engineering Skills→
PubCyBerryPubCyBerry

reading-sensor-pointclouds

Map and read lidar or radar PointCloud tensors for sensor data processing.

Community
Intermediate
NVIDIA-OmniverseNVIDIA-Omniverse
201

interpreting-radar-pointclouds

Interpret radar PointCloud tensors by mapping channel meanings and validating detections.

Official
Intermediate
PubCyBerryPubCyBerry

interpreting-lidar-pointclouds

Explain lidar PointCloud tensor channels, units, validity flags, and coordinate frames.

Community
Advanced
NVIDIA-OmniverseNVIDIA-Omniverse
201

configuring-lidar-sensors

Automate OmniLidar prim configuration and PointCloud channel mapping in USD scenes.

Official
Intermediate