Librealsense Overview, Examples, Pros and Cons in 2026

RealSense SDK 2.0 is a cross-platform library for RealSense depth cameras. The SDK allows depth and color streaming, and provides intrinsic and extrinsic calibration information.

Use Cases

Below are some of the many real-world applications powered by RealSense technology:

Robotics	Depth Sensing	3D Scanning

Drones	Skeletal and People Tracking	Facial Authentication

Why RealSense?

High-resolution color and depth at close and long ranges
Open source SDK with rich examples and wrappers (Python, ROS, C#, Unity and more...)
Active developer community and defacto-standard 3D stereo camera for robotics
Cross-platform support: Windows, Linux, macOS, Android, and Docker

Product Line

RealSense stereo depth products use stereo vision to calculate depth, providing high-quality performance in various lighting and environmental conditions.

Here are some examples of the supported models:

Product	Image	Description
D555 PoE		The RealSenseâ¢ Depth Camera D555 introduces Power over Ethernet (PoE) interface on chip, expanding our portfolio of USB and GMSL/FAKRA products.
D457 GMSL/FAKRA		The RealSenseâ¢ Depth Camera D457 is our first GMSL/FAKRA high bandwidth stereo camera. The D457 has an IP65 grade enclosure protecting it from dust ingress and projected water.
D455		The RealSense D455 is a long-range stereo depth camera with a 95 mm baseline, global-shutter depth sensors, an RGB sensor, and a built-in IMU, delivering accurate depth at distances up to 10 m..
D435if		The D435if is one of RealSenseâ¢ Depth Camera with IR pass filter family expanding our portfolio targeting the growing robotic market. The D400f family utilizes an IR pass filter to enhance depth quality and performance range in many robotic environments.
D405		The RealSenseâ¢ Depth Camera D405 is a short-range stereo camera providing sub-millimeter accuracy for your close-range computer vision needs.

ðï¸ Explore more stereo products

Getting Started

Start developing with RealSense in minutes using either method below.

1ï¸. Precompiled SDK

This is the best option if you want to plug in your camera and get started right away.

Download the latest SDK bundle from the Releases page.
Connect your RealSense camera.
Run the included tools:
- RealSense Viewer: View streams, tune settings, record and playback.
- Depth Quality Tool: Measure accuracy and fill rate.

Setup Guides - precompiled SDK

$Linux\Jetson Guide$

Note: For minor releases, we publish Debian packages as release artifacts that you can download and install directly.

2ï¸. Build from Source

For a more custom installation, follow these steps to build the SDK from source.

Clone the repository and create a build directory:

git clone https://github.com/realsenseai/librealsense.git
cd librealsense
mkdir build && cd build

Run CMake to configure the build:
```
cmake ..
```
Build the project:
```
cmake --build .
```

Setup Guides - build from source

Python Packages

Which should I use?

Stable: pyrealsense2 â validated releases aligned with SDK tags (Recommended for most users).
Beta: pyrealsense2-beta â fresher builds for early access and testing. Expect faster updates.

Install

pip install pyrealsense2 # Stable
pip install pyrealsense2-beta # Beta

Both packages import as pyrealsense2. Install only one at a time.

Ready to Hack!

Our library offers a high level API for using RealSense depth cameras (in addition to lower level ones). The following snippets show how to start streaming frames and extracting the depth value of a pixel:

C++

#include <librealsense2/rs.hpp>
#include <iostream>

int main() {
    rs2::pipeline p;                 // Top-level API for streaming & processing frames
    p.start();                       // Configure and start the pipeline

    while (true) {
        rs2::frameset frames = p.wait_for_frames();        // Block until frames arrive
        rs2::depth_frame depth = frames.get_depth_frame(); // Get depth frame
        if (!depth) continue;

        int w = depth.get_width(), h = depth.get_height();
        float dist = depth.get_distance(w/2, h/2);         // Distance to center pixel
        std::cout << "The camera is facing an object " << dist << " meters away\r";
    }
}

Python

import pyrealsense2 as rs

pipeline = rs.pipeline() # Create a pipeline
pipeline.start() # Start streaming

try:
    while True:
        frames = pipeline.wait_for_frames()
        depth_frame = frames.get_depth_frame()
        if not depth_frame:
            continue

        width, height = depth_frame.get_width(), depth_frame.get_height()
        dist = depth_frame.get_distance(width // 2, height // 2)
        print(f"The camera is facing an object {dist:.3f} meters away", end="\r")

finally:
    pipeline.stop() # Stop streaming