Computer vision in Robotics : Autonomous Mobile Robots, Robotic Arms and articulated robots, Humanoid robots and more

A revolution in how robots see the world.

Lightweight, low power and easy‑to‑use depth cameras that give robots the ability to navigate landscapes, avoid obstacles and recognize objects, people and more.

Computer vision in Robotics

The world of computer vision in robotics comes into focus.

When robots can see, measure, analyze, and respond to their environment, a whole new world of possibilities opens up. To bring these possibilities to life, the robotics industry is increasingly adopting computer vision technology which allows robots to sense depth, navigate landscapes, and recognize objects, people, and scenes. In the last several years alone, the demand for computer vision technology that is reliable and performs across a variety of conditions has grown rapidly.

Depth driven robotics.

All kinds of robots use Intel RealSense depth cameras. Here are some of the main categories where you can find robots using depth.

Articulated robots

Articulated robots, also known as robotic arms, are robots that can pick up and move objects. They can either be mounted on a static base or be included as part of an AMR. They are found in many industries, picking fruit, constructing objects or sorting and packaging items for delivery. Using one or more Intel RealSense depth cameras as part of a robotic arm allows the robots to accurately size and grasp many different objects. It is also crucial for robotic arms to be able to differentiate between objects that may appear very similar in a 2D photograph. The use of depth means that the robot can distinguish and understand the shapes and sizes of multiple objects.

Autonomous Mobile Robots

Many different industries today use robots that can broadly be categorized as autonomous mobile robots, from delivery robots to those found in warehouses, factories, hospitals and more. An autonomous mobile robot is one which can navigate and move itself around spaces independently in service of its tasks. Once given appropriate tasks, the robot can act without the need for human guidance, finding its path and avoiding obstacles safely. AMRs require a solid understanding of their environment in real time to act – this is where Intel RealSense depth cameras come in. With high precision depth cameras that can operate in any environment, AMRs can get where they need to go, every time.

Humanoid robots

Humanoid robots can technically fall into either the AMR or Robotic arm categories – they are robots that are roughly humanoid in shape and can move independently but can often also have articulated arms for performing tasks. Humanoid robots are most often used in places where they interact directly with people – the humanoid appearance is often friendly and welcoming especially in industries like hospitality, medicine and retail. Intel RealSense depth cameras can be used to navigate and hold objects, just as with AMR and robotic arms. They can also be used to locate people, track and understand gestures or accept directions from users.

Wide variety of use cases.

The high-resolution imaging and depth sensing technology of the Intel RealSense cameras allow them to deliver a full range of computer vision capabilities specifically targeted for robotics developers. For high‑precision applications, choose the L515 or D415. If your application is fast‑moving or outdoors, select either the D435 or D435i, and for longer range applications, the D455 is perfect.

Collision Avoidance

Collision Avoidance
Collision avoidance

High frame rate and depth accuracy provides the data needed, when it’s needed, to avoid obstructions. The computer vision solution of choice for robot cleaners and other automated industrial use cases that require collision avoidance. Pick and place robots are able to extract objects from bins without the object hitting the sides of the bin. Automated inventory robots navigate store aisles and avoid running into aisles, carts, or customers.

Object Detection

Object Detection
Object Detection

Distinguishes foreground objects from background and provides the ability to measure and understand shapes. Factory robots use high resolution imagery to distinguish between a variety of objects. Pick and place robots distinguish between a variety of objects in a bin and adjust picking pressures based on the object’s identity. Greeting robots in the retail space use object detection capability to identify and track objects and interact with consumers.

Volumetric Measurement

Volumetric measurement
Volumetric Measurement

Measures objects in all three dimensions simultaneously and accurately with low measurement error. Factory robots use volumetric data to improve efficiency in packaging, placement, and storage applications. Robots in shipping facilities can determine the volume of boxes and of empty storage slots and manage inventory more efficiently. Retail robots measure empty shelf space for inventory restocking.

SLAM Tracking

SLAM Tracking
SLAM Tracking

Low latency, low power, high quality localization and mapping allows robots to quickly calibrate location. Industrial delivery robots use tracking capability for path planning that enables them to move more efficiently through space. Warehouse delivery robots navigate delivery routes and keep track of their location. Inventory robots utilize mapping technology to navigate store aisles and keep track of their location.

Customer Stories

RightHand Robotics

Automated warehouse order fulfillment

RightPick2 robots use Intel RealSense cameras, to quickly identify items to be able to automatically handle them at high speeds with high reliability.

Learn more
Aethon

Industry-changing mobile delivery solutions

TUG Robots use Intel® RealSense™ technology to help guide autonomous delivery robots in the manufacturing, healthcare, and hospitality industries.

Learn more
Simbe Robotics

The new generation of in-store retail robotics

Simbe's autonomous inventory robot Tally captures inventory and e-commerce grade insights to allow retailers to optimize their in-store operations.

Learn more

A Robust Lineup.

The Intel RealSense family of cameras is the broadest product line of computer vision cameras on the market, with a range of indoor and outdoor options that meet all the needs of today’s robotics engineers.

L515
Intel RealSense LiDAR Camera L515
Depth FOV 70° × 55°
Range .25 m to 9 m
Rolling Shutter RGB
D455
Intel RealSense Depth Camera D455
Depth FOV 86° × 57°
Range .4 m to 20 m
Global Shutter RGB and Depth
D435i
Intel RealSense Depth Camera D435i
Depth FOV 85° x 58°
Range .11 m to 10 m
Global Shutter
D435
Intel RealSense Depth Camera D435
Depth FOV 85° x 58°
Range .11 m to 10 m
Global Shutter
D415
Intel RealSense Depth Camera D415
Depth FOV 65° x 40°
Range .16 m to 10 m
Rolling Shutter

Developers

Start developing with Intel RealSense SDK 2.0.

Our open‑source SDK 2.0 offers a variety of wrappers supporting popular programming languages and platforms.

SDK 2.0

Download SDK

Get the latest Intel RealSense SDK 2.0 to kick-off your project.

Documentation

Documentation

Find answers to your questions with our how-to guides and articles.

code-icon

Code Samples

Get started fast with depth development.

Less words. More Action.

Talk to a representative

Please complete the form below so we can offer you the best solution.

By submitting this form, you are confirming you are an adult 18 years or older and you agree to share your personal information with Intel to use for this business request. You also agree to subscribe to stay connected to the latest Intel technologies and industry trends by email and telephone. You may unsubscribe at any time. Intel’s web sites and communications are subject to our Privacy Notice and Terms of Use.