AprilTag Targets | High-Accuracy Calibration Boards
Calibration Target Printing

AprilTag Targets

568 Reviews

Apriltag calibration targets are lightweight, durable, and digitally printed with vibrant UV ink. They are made from rigid aluminum composite with a matte finish and can be printed on both sides. Rounded corners and custom shapes and sizes from 4"x4" to 46"x94" are available. New * Now offering an Anti-reflective Print Option, "Ultra Matte," with White Ink! Read More

  • 0.125
W
in
x
H
in
  • 1/2" All Corners $3.00
  • 1/4" All Corners $3.00
  • 3/8" All Corners $3.00
  • Special $56.44
  • Ultra Matte $96.75

$164.25

Demo of April Tag Localization System from 2012

This video demonstrates a localization system using AprilTags.


AprilTag Targets

Are you looking to pre-calibrate your camera with a calibration test target to be as accurate as possible? Our custom AprilTag targets are printed directly onto a rigid on premium 1/8 inch Dibond, making them the perfect calibration boards to calibrate cameras.

Direct to substrate printing at 1000 dpi with anti-reflective surface "flood white" authentic matte finish allows for easy detection in your computer vision pipeline because of its optical performance and physical robustness.

1/8

Dibond Features

Dibond signs have forever been an industry standard for indoor and outdoor business signs. The material is water, weather, warp resistant and rust proof. It's ultra flat and rigid structure makes Dibond the perfect material for calibration targets.

  • Light weight
  • Ultra flat surface
  • Anti-reflective surface (flood white finish) pitch black on pure white
  • Rigid and sturdy
  • Water, weather & rust proof

Technical Specifications

  • Material Thickness: .125"
  • Panel Type: Aluminium Composite Panel
  • Temperature Range: from -50°C to +80°C
  • Fire Classification: CLASS P/0/1
  • Physical Weight: 0.78125 lbs per ft.²

Helpful Ordering Tips

  1. Order the exact sizes that match your file's artboards; use the keyboard to add decimals if needed.
  2. Upload your own custom files, vector PDF files are preferred for optimal print results.
  3. Full-scale files are required to avoid the files being scaled up or down to fit the image on the board's order size.
  4. During the proofing process, always double-check and compare the file size to each board order size (Foamcoreprint.com is not responsible for artwork scaling if the incorrect board sizes are ordered and proofs are approved incorrectly).
  5. If you have questions about the artwork uploaded to an order; please call, email, or start a live chat before sending the art files to print. Production will begin within 10 mins of approval of artwork. Once production begins, we can not remove approved art files from the production line.
  6. Don 't have a Print-ready file? Use our online tool and generate an Apriltag pattern here.

AprilTags: A Versatile and Robust Visual Fiducial System

AprilTags are usable in a wide variety of applications, for instance, camera calibration. Once organized in a uniform sequence, you can use AprilTag makers like other calibration patterns.

So, what are AprilTags? In what instances can you use AprilTags, and under what conditions?

What can affect the accuracy of your results? How do you solve these issues to ensure pinpoint outcomes?

This guide takes you through the AprilTag frequently asked questions to ease your future applications with these tags.

What are AprilTags?

AprilTags refer to the visual fiducial system, usable in numerous applications like camera calibration, augmented reality, and robotics. Apriltags offer a method of 3D positioning and identification, even under reduced visibility.

These tags function like barcodes. They store little quantities of information (tag ID) while simultaneously facilitating straightforward and precise tag pose estimation in six dimensions (x, y, z, roll, pitch, yaw).

With the origin of the AprilTags project being the University of Michigan, you can learn a lot about this AprilTag detection software from the dedicated research on the subject on their website. The team provides implementations in C and Java in reading AprilTags from your camera, and you can find more tag-reading implementations online.

Besides this software, an MIT (Massachusetts Institute of Technology) student developed a C++ implementation. Their website offers printable AprilTags encompassing multiple tag families, which you can access in PDF format.

ROS Implementation

Hydro and ROS Groovy lack inbuilt AprilTag identification software. However, it is possible to convert the MIT C++ implementation code into a functioning ROS package to produce AprilTags seen in your camera image as TensorFlow messages and transforms.

Experimentation Results

April Tags can achieve accuracy within four centimeters from the original pose if your camera lies within a two-meter radius from your tag. The distance of the robot from the tag affects the precision inversely.

Potential Issues

There are various crucial aspects to consider before applying April Tags.

Pose Estimation Accuracy Depends on the Camera You Use

Using a very affordable camera, say costing under $20, can immensely affect accuracy. Your tag detections with such a unit are unlikely to provide reliable outcomes. Moving such a camera while detecting your Apriltags can further degrade the quality of your output.

With a more premium camera, the quality of your results is likely to be good. You can achieve pose estimation accuracy with proper tuning of parameters.

Estimation Accuracy Depends on Tag Motion

As your robot moves, especially during turning, pose estimation can veer off accurate results by even over 20 degrees. The range can also vary by up to 75 centimeters.

The best approach is halting immediately an April Tag gets detected. This way, the likelihood of achieving a more precise pose estimation is higher.

Tag Distance Also Affects Your Pose Estimate

Typically, you can achieve pose accuracy up to about 2.4 meters. Beyond this level, the results tend to vary substantially. Consequently, a robot shouldn't localize unless within the 2.4-meter threshold range.

Determining Ideal Camera Configuration Stipulations is Crucial

As seen, the inaccuracy of robots can result from several reasons. Another potential cause of localization inaccuracy is the failure to meet the set requirements for the given hardware you are using. An example would be using an unsuitable focal length.

For instance, if your Kinect sensor has a smaller focal length than the required value in your software, your outcome may vary. The robot assumes it is closer to your tag than true, resulting in an imprecise tag size approximation.

With the usage of ideal parameters, precision improves by a few centimeters within close range and increases immensely when the distance increases.

Frame Transformation

When using the April Tag node, an issue is your frame's dependency on the moving body's orientation. The body’s orientation can alter your frame and make it challenging to achieve a uniform coordinate system.

The April Tag transform is the source code, which though achieving frame consistency, can be quite noisy. A RANSAC filter can help you with this issue. Relying on the IMU (Inertial Measurement Unit) orientation rather than the April Tag orientation during transformation can help address the errors.

Substitutes

Besides AprilTags, you can also use AR-tag to localize or estimate poses, though these get primarily used in Augmented Reality. And while they may not be as good as Caltags, April Tags offer more precision than ARtags, especially when lighting conditions are below the required levels.

Camera Calibration Using AprilTag Markers

AprilTags support a wide variety of usages as visual markers in localization, object detection, and as a camera calibration target. Although similar to QR codes, AprilTags encode less information. You can, therefore, decode AprilTags faster, a vital feature in various fields like real-time robotics.

We shall use the readAprilTag function in this example to localize and detect AprilTags in calibration patterns. This function accommodates all standard tag families.

The example also performs end-to-end calibration using Computer Vision Toolbox™ functions. We replace the checkerboard pattern with an evenly-spaced AprilTags grid.

The benefits of using AprilTags as calibration model are:

  • Better feature point determination
  • Consistency
  • Repeatable detections- they are typically easily detectable in real-time

You can also use the method below with other calibration patterns like a circle grid rather than the default checkerboard patterns.

Step 1: Calibration Pattern Generation

Download and draw up your tag images. You can download pre-made tags for all supported families with a web browser from AprilRobotics.

You can use a helperGenerateAprilTagPattern function to generate calibration targets using your tag image for a particular placement of tags. You can obtain the pattern image in calibPattern, which you can then use to print your AprilTag pattern (from MATLAB). The example employs the tag36h11 family. These offer a worthwhile trade-off between false-positive detections and detection performance.

Applying the readAprilTag function to the above pattern produces detections in which the border positions of individual tags fall into the same group. You can use the helperAprilTagToCheckerLocations function to change the above arrangement into a column-major positioning like a checkerboard.

tag36h11 family

Image source (tag36h11 family): screenshot

tag36h11

Image source (tag36h11 family): screenshot

Image Preparation for Calibration

You must print the formulated calibration pattern on a flat surface then use the camera you want to calibrate to take the pattern's images.

As you prepare your images for calibration, here are a few issues to keep in mind:

  • Consider printing your Apriltags on a consistently flat surface not prone to deformation from agents like moisture. Paper is therefore not the best object to use.
  • Any pattern imperfections like uneven surfaces can lower calibration precision, given these procedures assume your pattern is planar.
  • Camera calibration needs a minimum of two images of your AprilTags but increasing the number to about 10 to 20 images produces more exact outcomes.
  • Picture various images of your Apriltags and let the Apriltag pattern occupy most sections of your image and cover your entire field of view. For instance, capture images of your Apriltag pattern on all the edges of your image borders for you to ultimately detect lens distortion.
  • Ensure your AprilTags are wholly visible within the taken photos. Pictures with partly visible AprilTags get rejected.

Step 2: Detect and AprilTag Localization

Use the helperDetectAprilTagCorners function to identify and localize your AprilTags from the captured pictures. Arrange the photos like a checkerboard pattern to 0ffer you the vital AprilTag points for the calibration process.

Step 3: Generate World Coordinates for Your AprilTags

The generated AprilTag print appears in a manner like that of a checkerboard target. Therefore, you can determine the world coordinates for the coinciding image coordinates you got in imagePoints with the help of the generateCheckerboardPoints function.

The tag size replaces the size of the square, and you can obtain the board size from step two. Measure the size of your tag between your tag's exterior black edges.

Step 4: Camera Parameter Estimation

Use the estimateCameraParameters function to approximate your camera's parameters. You can do this using the points that coincided between your photo and the world coordinate system.

Gauge how precise your extrinsic camera parameters ( those indicating the planes of your AprilTags in your captured photos) and the calibration are based on your findings.

Assess the positions of the detected picture points and your reprojected points gotten from the approximated camera parameters.

Supporting Functions

There are three main AprilTag supporting functions during the calibration process:

  • The helperGenerateAprilTagPattern function formulates AprilTag-based calibration models.
  • The helperDetectAprilTagCorners function can detect AprilTag calibration models in images.
  • The helperAprilTagToCheckerLocations function changes AprilTag corners into checkerboard corners.

Wrapping Up

Although the message in the example above shows how to use AprilTag markers in camera calibration, it is not specifically tied to AprilTags. The same method is usable in several applications with other calibration models.

The estimateCameraParameters function vital in obtaining your camera parameters needs:

  • WorldPoints: Coinciding world coordinates of the crucial points of your calibration pattern
  • ImagePoints: Vital points in image coordinates of your calibration pattern sourced from the recorded images.

Provided you can achieve these two necessities; you can follow the above calibration workflow.


Reviews for AprilTag Targets

5
Out of 5.0
5 Star
(16)
4 Star
(0)
3 Star
(0)
2 Star
(0)
1 Star
(0)
Overall
Rating
100%
of customers that buy this product give it 4-5 stars.
100%
of customers would like to buy this product again.
Verified Buyer
6 months ago by (IN, US)


Verified Buyer
9 months ago by (MA, US)


Verified Buyer
a year ago by (CA, US)


Frequently asked Questions

FAQ

What is an AprilTag?

The AprilTag visual fiducial system can be used for various tasks, including augmented reality, robotics, and camera calibration. AprilTag detection software computes the precise 3D position, orientation, and identity of the tags in relation to the camera using ordinary printers as targets.

How many AprilTags are there?

In the AprilTag families, there are two generations to be found: Apriltag 2 and the newest generation, AprilTag 3. AprilTag 2 is a classic version of the Apriltag. It is recommended that you utilize Apriltag 2 (gen2) if you want a better cookie cutter out-of-the-box experience.

How are April Tags determined?

Detecting AprilTags in a picture begins with specifying the parameters, and specifically, the family of AprilTags. The set of six likely AprilTag families that our AprilTag detector can identify. When using AprilTags, a family represents a collection of tags that the AprilTag detector will assume the input image.

How accurate are Apriltags?

The AprilTag achieved a location accuracy of 1.77 cm and an orientation estimation accuracy of 4.15 cm with a precision of 1.77 cm. AprilTag has been utilized as a landmark by particular academics, who have used it to track it in robotics applications.

How do AR tags work?

With AprilTag, video tracking capabilities are available that calculate a camera's location and orientation in real-time relative to physical markers. In augmented reality applications, once the camera's position is known, a virtual camera can be positioned at the exact location, revealing the virtual object at the location of the AR tag.

YOU MAY ALSO LIKE

Notify on stock availability