Dissertation > Excellent graduate degree dissertation topics show

Vehicle Logo Location Algorithm Research Based on Video Image

Author: LiHongLin
Tutor: WangYunQiong；FengQiaoSheng
School: Yunnan Normal University
Course: Computer Software and Theory
Keywords: Edge colors mean on License Plate Location Front position Difference Symmetry Detection Car standard positioning
CLC: TP391.41
Type: Master's thesis
Year: 2007
Downloads: 36
Quote: 0
Read: Download Dissertation

Abstract

Car logo recognition system is an important part of the Intelligent Transportation Systems, is also a hot topic of cross-disciplinary research in computer vision, image processing and pattern recognition, so the the car logo recognition related technologies in the field of research is of universal concern. The logo is iconic image of the vehicle includes not only the model information, but also contains a manufacturer's information and, more importantly, it is difficult to replace. If the car standard information combined with the license plate information, will greatly improve the reliability of the vehicle identification. Car logo recognition technology can be widely used in traffic, automatic monitoring of illegal vehicles, automatic fare collection, the automatic safety management of airports, ports, car parks, residential and other vehicles, combat vehicle crime, so car logo recognition on intelligent transportation vehicles The recognition technology improvement and development of important theoretical significance and application value. Car logo recognition system, the car standard positioning is a crucial part of its positioning speed and accuracy of a direct impact on the performance of the car logo recognition system. This paper combines the Chinese license plate, the vehicle the subject characteristics, and conducted in-depth research on how accurate extraction of the license plate and car standard, the main research work are as follows: (1) an improved HSV color model. Because of the gradation level of the pixels corresponding to the V with the image, so this article V is to take after the luminance value of the grayscale, experiments show that the new and improved model can effectively extract the mean value of the pixel of color plates. (2) a new method based on the edge colors mean on the license plate segmentation. Fixed with this method makes full use of the background color of the license plate character color of the license plate, struck in the RGB color space the adjacent pixel RGB component of the mean, and then converted to the HSV color space, and if there is in line with the colors that the license plate the candidate region. (3) license plate location method based on projection and geometric characteristics of the combination. The horizontal projection of the image, and then calculates the area of ??each connected region; Finally, according to the license plate of the width, height, aspect ratio and other geometric feature constraint, further removal does not comply with the area of ??the license plate feature, the plate positioning. (4), according to the position of geometry, differential, projection and area combined crude car standard positioning. The location of use of the license plate and the plate size, as well as the relative position between the license plate with the front and size, to determine the front position; combined with the texture of cars marked region rich characteristics, the use of the differential, the projection and the communicating area method car standard coarse positioning. (5) The proposed car standard combination of interval-based differential symmetry detection fine positioning. Candidate region of the car standard car standard area rich texture interval pixel difference, there are several regional connectivity some image differential symmetry detection method to find the area of ??the center axis of symmetry. Only after the center axis of symmetry of the connected region was considered to be the area where the car standard, the car standard precision positioning. The experimental results show that: the license plate positioning with high accuracy positioning, a little faster. Car standard positioning method works better when the license plate just above the car standard is not able to find the correct location of the vehicle the subject, to overcome the inadequacies of the traditional use of license plate positioning vehicle the subject, and to meet the basic needs of practical application.

Related Dissertations

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net Mobile