Font Size: a A A

The Monitoring Of Transport Line Which Based On Gait Recognition

Posted on:2013-02-19Degree:MasterType:Thesis
Country:ChinaCandidate:X F LiuFull Text:PDF
GTID:2218330371959466Subject:Circuits and Systems
Abstract/Summary:PDF Full Text Request
Rail transport has been rapid development in recent years, both passenger traffic and the speed has being much increased.With the continuous development of bio-technology, rail traffic control slowly changing from analog control to intelligent monitoring. Focus of this study is to identify the biological characteristics of gait in the railway transport monitoring applications.Gait-based identification technology the emergence of a new type of biometric technology in recent years.It's aim is based on people walking posture to achieve individual identification.This technology in a very long time scientists in various fields has been a research focus.Gait recognition mainly on sequence analysis of human walking movement, with its emphasis on efficient extraction of gait features and select the appropriate classification. It combines the moving object detection and video sequence processing, pattern recognition and other technologies.This paper presents a method based on Fisher Discriminant feature extraction method and improve the nearest neighbor classifier. Fully amplified gait characteristics. Greatly raised the profile-based feature extraction of gait recognition success rate.This article focuses on human gait feature extraction and recognition have been studied to obtain the following results:1) Research-based gait recognition for surveillance systems used in the basic methods and algorithms, including motion detection, image sequence coding, feature extraction, pattern classification and recognition, the final identification of the object according to select the appropriate gait database.2) To study the gait recognition preprocessing, the process is as follows:First, the time average image was established by using the background model, then use the background difference method to extract contours of the body moving target, and then the body morphological processing area and normalized.3) Proposed a method based on Fisher discriminant gait feature extraction methods. Fisher discriminant method using the weighted matrix of the characteristic gait features become more pronounced. Then the characteristic matrix of the weighted analysis, dimension reduction obtained after the gait characteristics.4) The classification used in this paper, an improved nearest neighbor classifier. Commonly used classifiers are ignored in the same class of similarity between all training samples, as well as the correlation between them. You can determine the object to be identified in samples with each class to consider the relationship between similar classification. This idea is that when the classifier to identify the object to be placed in a class, this class of high-dimensional feature space in terms of shape and internal structure is very regular.Single frame of image based detection method compared to the gait, Fisher Discriminant full account of the inter-frame changes in gait and amplified this change, and then extract the more obvious characteristics of gait. The research is a good solution to the current widespread use did not change significantly based on single-frame feature gait problems.
Keywords/Search Tags:Traffic alert, Biometrics, Fisher Discriminant method, Featureextraction, Gait recognition
PDF Full Text Request
Related items