| Abstract [eng] |
The main goal of this work is to create an algorithm for classifying vehicles using data gathered from multiple passive sensors. During the research, various possibilities of using passive sensors for the purpose of vehicle classification were investigated. The first part reviews the work of other scientists, the sensors and systems used in their research. It describes the most popular methods of vehicle classification using video cameras and magnetic sensors. Several less popular methods using other sensors are also described. Finally, the results obtained by the authors of the articles using the described methods and systems are compared. The second part describes a data collection system of the project. The principle of its operation is explained, and the sensors and system parameters are described. The third part presents an analysis of 4 types of sensors - magnetic, accelerometers, microphones and piezoelectric sensors. Data collected from these sensors are described and analyzed. Most of the classification models and methods require extracting features from collected data therefore, this section also describes the methods used and the methods that need to be calculated. Finally, conclusions are drawn on the information provided by each type of sensor for classification purposes. The fourth part describes three different types of classification models. Classification results are presented for classification into 5, 6 and 7 classes using different methods obtained from the collected data. The obtained results are compared, and possible situations of their use are described. |
