Font Size: a A A

A Study On Identification Of Vehicle Types Based On Spiking Neural Network Model

Posted on:2012-02-11Degree:MasterType:Thesis
Country:ChinaCandidate:H ChenFull Text:PDF
GTID:2218330368483593Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
The intelligent identification technologies have been applied to many intelligent transportation systems. And vehicle recognition technology is very important in identification technology of intelligent transportation systems.The recognition technology can be used to identify and classify the vehicle in a specific time and place, and the results of recognition can be used in transport management, scheduling, statistics in charge systems on roads and bridges. This paper focuses on image processing and analysis of moving vehicles, an approach is proposed to combine the neural network and the current image processing technology for vehicle identification and classification. Artificial neural network is the hottest research direction in intelligent identification domain, and application of artificial neural network in intelligent transportation systems has become an important topic in current research.Through analyzing the prospect of development and application in intelligent transportation systems, this paper proposes a recognition and classification research scheme based on artificial neural network models. In the scheme moving vehicals are captured by using a kind of axonal delay mechanism based on the spiking neural network model and a set of advanced image processing technologies. And then a spiking neural network is used to extract edges of moving vehicles to obtain a firing rate map of corresponding edges. Then, we can extract line moment features from the firing rate map. Finally, the combination of line moment features and other traditional features is used to train the BP neural network, thus an automatic identification system is designed. This system can not only be used to classify and recognize the moving vehicles accurately and correctly, but also can be used to identify other moving objects. Experiment results show that this system achieves very good recognition results. The system can be used in intelligent traffic monitoring and management, and so on, it has important practical significance.
Keywords/Search Tags:Spiking neural networks, line moment, feature extraction, vehicle-type classification and identification
PDF Full Text Request
Related items