12.11.2012 Views

A License Plate Recognition and Speed Detection System - Index of

A License Plate Recognition and Speed Detection System - Index of

A License Plate Recognition and Speed Detection System - Index of

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

9. The algorithm then finds the longest run <strong>of</strong> maxes <strong>and</strong> finds their median. This median is<br />

the optimal intensity threshold. I.E. intensity threshold = (152 + 154) / 2 = 153 (intensity<br />

threshold is an integer).<br />

This completes the description <strong>of</strong> the application developed for this project. By using<br />

existing free (or relatively free) s<strong>of</strong>tware libraries <strong>and</strong> creating our own customized code we<br />

were able to build an executable that could recognize <strong>and</strong> record the license plate numbers <strong>of</strong><br />

passing vehicles.<br />

V.D. Testing<br />

Testing was done throughout the build <strong>of</strong> the system. Tests were run against each<br />

individual component to ensure its functionality. Three separate components; the radar gun, the<br />

serial interface circuit board, <strong>and</strong> the license plate recognition s<strong>of</strong>tware were each tested<br />

separately. A full system test was never accomplished due to a malfunctioning radar gun.<br />

Descriptions <strong>of</strong> the tests performed for each component are provided in the following<br />

paragraphs.<br />

The license plate recognition s<strong>of</strong>tware component underwent much testing through its<br />

development. Several revisions to the source code were made as a result <strong>of</strong> this testing. In the<br />

beginning, tests were made against a stationary license plate. The first version <strong>of</strong> the source code<br />

had several bugs. For one, it was not recognizing license plate characters in certain test cases. It<br />

was found that if the background behind the plate was <strong>of</strong> a certain color or intensity, the<br />

recognition s<strong>of</strong>tware would fail. This was fixed by changing the license plate recognition<br />

algorithms. Instead <strong>of</strong> trying to extract the license plate region from the image, the license plate<br />

60

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!