Skip to content

duckietown/lib-dt-apriltags

Repository files navigation

lib-dt-apriltags: Python bindings for the Apriltags library

These are Python bindings for the Apriltags 3 library developed by AprilRobotics. Inspired by the Apriltags2 bindings by Matt Zucker.

The original library is published with a BSD 2-Clause license.

Installation

The easy way

You can install using pip (or pip3 for Python 3):

pip install dt-apriltags

And if you want a particular release, add it like this:

pip install [email protected]

Build it yourself

Clone this repository and navigate in it. Then initialize the Apriltags submodule:

$ git submodule init
$ git submodule update

Build the Apriltags C library and embed the newly-built library into the pip wheel.

$ make build

The new wheel will be available in the directory dist/. You can now install the wheel

pip install dt_apriltags-VERSION-pyPYMAJOR-none-ARCH.whl

NOTE: based on the current VERSION of this library and the version of Python used PYMAJOR, together with the architecture of your CPU ARCH, the filename above varies.

Build for Python 3

This library supports building wheels for Python 2 and 3. Python 2 will be used by default. Use the following command to build for Python 3.

make build PYTHON_VERSION=3

Build for different architecture

This library supports building wheels for the CPU architectures amd64, arm32v7, and arm64v8. Default architecture is amd64. When building wheels for ARM architectures, QEMU will be used to emulate the target CPU. Use the following command to build for arm32v7 architecture.

make build ARCH=arm32v7

Release wheels

All the wheels built inside dist/ can be released (pushed to Pypi.org) by running the command

make upload

Release all

Use the following command to build and release wheels for Python 2 and 3 and CPU architecture amd64 and arm32v7.

make release-all

Usage

Some examples of usage can be seen in the test.py file. The Detector class is a wrapper around the Apriltags functionality. You can initialize it as following:

at_detector = Detector(searchpath=['apriltags'],
                       families='tag36h11',
                       nthreads=1,
                       quad_decimate=1.0,
                       quad_sigma=0.0,
                       refine_edges=1,
                       decode_sharpening=0.25,
                       debug=0)

The options are:

Option Default Explanation
families 'tag36h11' Tag families, separated with a space
nthreads 1 Number of threads
max_hamming 2 The max number of bits that are allowed to be flipped to generate a successful tag detection. Can help decrease false negatives when noise causes some data bits to be read incorrectly, but can also increase false positives.
quad_decimate 2.0 Detection of quads can be done on a lower-resolution image, improving speed at a cost of pose accuracy and a slight decrease in detection rate. Decoding the binary payload is still done at full resolution. Set this to 1.0 to use the full resolution.
quad_sigma 0.0 What Gaussian blur should be applied to the segmented image. Parameter is the standard deviation in pixels. Very noisy images benefit from non-zero values (e.g. 0.8)
refine_edges 1 When non-zero, the edges of the each quad are adjusted to "snap to" strong gradients nearby. This is useful when decimation is employed, as it can increase the quality of the initial quad estimate substantially. Generally recommended to be on (1). Very computationally inexpensive. Option is ignored if quad_decimate = 1
decode_sharpening 0.25 How much sharpening should be done to decoded images? This can help decode small tags but may or may not help in odd lighting conditions or low light conditions
searchpath ['apriltags'] Where to look for the Apriltag 3 library, must be a list
debug 0 If 1, will save debug images. Runs very slow

Detection of tags in images is done by running the detect method of the detector:

tags = at_detector.detect(img, estimate_tag_pose=False, camera_params=None, tag_size=None)

If you also want to extract the tag pose, estimate_tag_pose should be set to True and camera_params ([fx, fy, cx, cy]) and tag_size (in meters) should be supplied. The detect method returns a list of Detection objects each having the following attributes (note that the ones with an asterisks are computed only if estimate_tag_pose=True):

Attribute Explanation
tag_family The family of the tag.
tag_id The decoded ID of the tag.
hamming How many error bits were corrected? Note: accepting large numbers of corrected errors leads to greatly increased false positive rates. NOTE: As of this implementation, the detector cannot detect tags with a Hamming distance greater than 3.
decision_margin A measure of the quality of the binary decoding process: the average difference between the intensity of a data bit versus the decision threshold. Higher numbers roughly indicate better decodes. This is a reasonable measure of detection accuracy only for very small tags-- not effective for larger tags (where we could have sampled anywhere within a bit cell and still gotten a good detection.)
homography The 3x3 homography matrix describing the projection from an "ideal" tag (with corners at (-1,1), (1,1), (1,-1), and (-1, -1)) to pixels in the image.
center The center of the detection in image pixel coordinates.
corners The corners of the tag in image pixel coordinates. These always wrap counter-clock wise around the tag.
pose_R* Rotation matrix of the pose estimate.
pose_t* Translation of the pose estimate.
pose_err* Object-space error of the estimation.

Custom layouts

If you want to use a custom layout, you need to create the C source and header files for it and then build the library again. Then use the new libapriltag.so library. You can find more information on the original Apriltags repository.

Developer notes

The wheel is built inside a Docker container. The Dockerfile in the root of this repository is a template for the build environment. The build environment is based on ubuntu:18.04 and the correct version of python is installed on the fly. The make build command will create the build environment if it does not exist before building the wheel.

Once the build environment (Docker image) is ready, a Docker container is launched with the following configuration:

  • the root of this repository mounted to /source;
  • the directory dist/ is mounted as destination directory under /out;

The building script from assets/build.sh will be executed inside the container. The build steps are:

  • copy source code from /source to a temp location (inside the container)
  • build apriltag library from submodule apriltags/ (will produce a .so library file)
  • build python wheel (the .so library is embedded as package_data)
  • copy wheel file to /out (will pop up in dist/ outside the container)