Cantag

Cantag is a marker-based machine vision system designed to allow the user to select, and compare, different tag designs and tracking algorithms. It allows investigation into fundamental properties and limitations of particular 2-dimensional marker tag designs. Cantag is open source software written in C++.

Many possible tag designs and symbolic coding schemes are available. Cantag has been used to examine the trade-offs between payload size, decoding error and transformation accuracy.

Cantag supports multiple types of image source for use with the testing and deployment of the final application. The use of generated images from the OpenGL test harness permits comparison between different tag designs and processing algorithms and iterative refinement of the system. Real-time processing may be applied to image data from video cameras, or batch processing from images and movies stored on disk.

The results from the image processing pipeline have many possible applications. Cantag allows the system designer to use the same algorithm implementation used in testing the tag design to eploy the application. Applications include: barcode reading for object identification, visual overlay onto the image and camera location estimation.

Introduction to the Tracking Process: A description of the basic tag tracking process.

Tag Terminology: Terminology for discussing the tag designs currently supported by Cantag.

The Vision Pipeline: Cantag consists of a large number of image processing algorithms. These operate on data stored in entities which form the basic data abstraction within Cantag.

Libraries and Dependencies: The libraries Cantag can make use of and where to find them.

Algorithms: A description of some of the algorithms within Cantag.

Dependable Coding of Fiducial Tags: The rotational symmetry common to many tag designs requires particular consideration in order to understand the performance of the coding schemes when errors occur. Rotational Invariance is abstract representation of tags carrying symbolic data which allows existing information coding techniques to achieve robust codes.

Information Theoretic Limits of Tag Tracking?: The amount of information present in a camera image places an upper bound on the ability to read the payload of a tag. The minimum sample distance metric allows us to quantify these limits and analyse systematic errors due to the pose and position of the tracked tag.

Simulating System Performance?: Cantag allows system designers to investigate the effects of different tracking algorithms. This ability is combined with an OpenGL test harness to provide a fair comparison of the performance of different algorithms and designs.

Real-world performance?: The real-world performance of the Cantag system shows strong similarities to the trends predicted by the theoretical and simulation based analyses.

Documentation and Downloads for Cantag are currently unavailable. We hope to restore them shortly

Documentation

The  Doxygen documentation for Cantag is available online:  View Documentation

Download

As yet, there are no official releases of Cantag. A nightly snapshot from our subversion repository is available from: Download snapshot