Summary: | 碩士 === 大葉大學 === 資訊管理學系碩士班 === 102 === Modern technological advancement, the type of commercial signs are
more diverse, in general terms, to obtain external information visually
is the fastest way to have the most direct and we can sign on from the
content business, get the necessary information, we want to find the
place. While the car is moving at night as seen on both sides of the
street there are many different types of signs, the Central Standing
Committee nighttime image recognition studies are subject to different
sources of interference identification more difficult to improve, so the
video image is often affected by the light source. This study aims to
capture the flashing signs from nighttime video screen and identify its
internal text, in order to overcome the difficulties in tracking nighttime
detection, the use of light for the signature characteristics, color and
size variation determination flashing signs, flashing signs while doing
Type determination, there are three main types of flashing, flashing
different ways determine the difference, according to the last flicker
type do text recognition. Mainly for the night environment, to detect
signs flashed through experimental simulation experiment, experiment
results and error analysis, to verify the feasibility of the proposed
method in this study.
|