Sensors · Point clouds · Lab 01 · Beginner

Understand & visualize a point cloud

Your robot has just received its first LiDAR point cloud.

The sensor has returned thousands of XYZ measurements, but right now they are just numbers. Your job is to figure out what those numbers represent and understand the robot's surroundings.

Your task

Your first three LiDAR points

Colour

Tap or click a point to inspect it.

meet.py Loading Python…

Output

Ctrl+Enter runs

What Python can I use?

Plain Python, math, and a browser-sized subset of NumPy: np.array, indexing like points[0], points[:, 0], points[-1] and masks; + − * / **; .min() .max() .sum() .mean() .argmin() .argmax() (with axis=); np.sqrt, np.sin, np.cos, np.arctan2, np.degrees, np.linalg.norm, np.column_stack. Code written here also runs with real NumPy.

Hints

What is a point cloud?

A LiDAR spins and fires laser pulses. For every pulse it measures how far away the surface is and in which direction it fired. Each measurement becomes one point (x, y, z), and one sweep produces thousands of them: a point cloud. Robots use point clouds to see obstacles, build maps and localize themselves.

This lab starts at the very beginning: what the three numbers mean, how the sensor produces them, how to look at them in 3D, and how to answer basic questions about a scan with a few lines of Python. Everything runs in your browser; your code never leaves this page.

What you will be able to do

Explain what x, y and z mean in a robot's sensor frame (X forward, Y left, Z up).

Index a point cloud: single points, the last point, and whole X, Y and Z columns.

Convert LiDAR range, azimuth and elevation to x, y and z.

Recognise the ground, walls, objects and shadows in a 3D scan.

Compute the point count, the extent of a cloud and its closest point, with its distance and bearing.

The point-cloud sequence

Lab 01: Understand & visualize a point cloud: this lab.

Lab 02: Process & transform a point cloud (coming soon).

Lab 03: Build a LiDAR obstacle detection pipeline (coming soon).