Scribd
Upload
7 views

Image Processing - Copy (3)

Uploaded by

Matt
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
7 views

Image Processing - Copy (3)

Uploaded by

Matt
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

EE368: Digital Image Processing Project Report Ian Downes

[email protected] Stanford University Abstract— An algorithm to detect


and decode visual code markers in medium resolution images is presented.
The algorithm uses adaptive methods to segment the image to identify
objects. The objects are then used to form candidate markers which are
examined for several criteria. Potential markers are then sampled and guide
information present in the marker is used to verify the data. The algorithm is
invariant to scale and rotation and is robust to motion blur and varying
illumination. The algorithm is implemented in C and takes approximately 100
ms to execute per image using a Pentium 4 3.40 GHz computer. I.
INTRODUCTION The presence of cameras in cell phones is becoming
extremely commonplace as the price of inclusion plummets. As more and
more people are equipped with these cameras it becomes feasible to
develope a range of applications that utilize the camera for purposes other
than simply taking a snapshot. One such application is to use the camera to
sample the data contained in a visual code marker and to use this as a code
to reference information. A typical use case might be to include a visual code
marker next to the advertisment for a movie. By taking an image of the
marker the phone can decode the data and then query a backend database
for the local screening times of the movie. To achieve this type of application
the camera must be able to reliably identify and decode visual markers. This
report details the development and implementation of an algorithm that can
be used for this purpose. The report first analyses the problem and
establishes the requirements for the algorithm. It then examines the major
steps of the algorithm and how it meets the requirements. The results of
testing the algorithm are discussed before the report concludes with a
summary. II. ALGORITHM DESIGN Before the design of the algorithm was
started the problem was analysed to develop some ideas about how the
visual code markers could be effectively detected and the issues that would
need to be solved. Once a basic high level method was established the lower
level stages of the proposed algorithm were designed and tested. The follow
sections detail how the problem was analysed and the resulting algorithm. A.
Problem Analysis Figure 1 shows the format of the visual code marker. The
marker is composed of an 11×11 grid of binary elements. The binary
elements are represented by a black square for a ‘1’ and a white square for
‘0’. The origin of the array is defined to be the top left corner (that is, the
corner opposite the intersection of the guide bars) and the data is ordered in
column major order. The central region of the marker contains 83 bits of
data. The remaining 38 bits are spaced at the corners and are used to form
guide elements. The guide elements consist of three corner square elements
and two guide bars. The guide bars are perpendicular to each other and are
of different lengths. In addition to the 11×11 array the marker also has a
white border extending (at least) one element width around the array. This
border serves the same purpose as the border around barcodes and ensures
that the marker is separate from any background it is placed on. The image
source is a digital image from a camera equipped cell phone. The image is of
VGA resolution (640 × 480) and provided as a compressed 8-bit RGB JPEG
file. The compression ratio is approximately 18:1. This resolution is
comparatively low but it is of course definately adequate to sample and
extract the data from the markers under the right conditions. Of more
importance are the conditions under which the image is obtained.
Specifically, the camera is most likely to be handheld, leading to a degree of
motion blur, and to be positioned somewhat haphazardly, leading to
arbitrary transformations in angle, translation and perspective. Furthermore,
the lighting conditions are not controlled and it is possible that the
illumination will not be constant over the image and definately not between
images. The high contrast between the binary elements (black vs. white) will
help to differentiate the binary values under a wide range of lighting
conditions. The marker is likely to be a small fraction of the total image data.
It is expected that the marker will be approximately between 100 × 100 and
150 × 150 pixels. As part of the project brief the image may contain one to
three markers in an image. Although it may be possible to find the location
of markers by perhaps looking for mixed dark/light regions it is essential that
the guide elements be used to orient the marker. Correct orientation is
required so that the data can be read in a meaningful way. The provision of
the perpendicular guide bars allows the detection of any rotation of the
marker. In addition, if necessary, the observed change in the 90o angle could
be used to estimate the perspective of the camera view, opening up the
possibility of rectifying the image. The corner guide elements can be used to
accurately determine the extent of the marker. This will be particularly
important when the viewing perspective distorts the shape of the marker.
This analysis has shown that the algorithm must be able to

You might also like