|
|
Comparison of technologies for vehicle passage detection
Mareková, Martina ; Frolka, Jakub (referee) ; Krajsa, Ondřej (advisor)
System for vehicle detection and speed measurement system is an important part in traffic control as well as in providing data for intelligent and automated traffic signs. These technologies are divided into two types: intrusive and non-intrusive methods. Traditional traffic control is based on inductive loops. This method is intrusive and therefore requires intervention under the road and higher maintenance requirements. We can avoid this problem with non-intrusive methods, including video analysis, microwave radars, geomagnetic radars, weighing sensors and laser sensors, but these are more expensive alternatives. By implementing radars we performed measurements on road sections, on the basis of which the output data from microwave, weigh-in-motion sensors and induction loops were displayed. We receive processed and unprocessed output data of measurements from each sensors. For processing of those data is created software solution for graph rendering and direct measurement of the sensor accuracy in comparison with competing technology. The results of the measurements showed that the efficiency of the individually implemented sensors is sufficient for use in traffic control. As well we can consider the implementation of microwave radars and cameras for video analysis as a replacement for the conventional method of induction loops.
|
|
|
Vehicle classification using inductive loops sensors
Halachkin, Aliaksei ; Klečka, Jan (referee) ; Honec, Peter (advisor)
This project is dedicated to the problem of vehicle classification using inductive loop sensors. We created the dataset that contains more than 11000 labeled inductive loop signatures collected at different times and from different parts of the world. Multiple classification methods and their optimizations were employed to the vehicle classification. Final model that combines K-nearest neighbors and logistic regression achieves 94\% accuracy on classification scheme with 9 classes. The vehicle classifier was implemented in C++.
|
|
|
Comparison of technologies for vehicle passage detection
Mareková, Martina ; Frolka, Jakub (referee) ; Krajsa, Ondřej (advisor)
System for vehicle detection and speed measurement system is an important part in traffic control as well as in providing data for intelligent and automated traffic signs. These technologies are divided into two types: intrusive and non-intrusive methods. Traditional traffic control is based on inductive loops. This method is intrusive and therefore requires intervention under the road and higher maintenance requirements. We can avoid this problem with non-intrusive methods, including video analysis, microwave radars, geomagnetic radars, weighing sensors and laser sensors, but these are more expensive alternatives. By implementing radars we performed measurements on road sections, on the basis of which the output data from microwave, weigh-in-motion sensors and induction loops were displayed. We receive processed and unprocessed output data of measurements from each sensors. For processing of those data is created software solution for graph rendering and direct measurement of the sensor accuracy in comparison with competing technology. The results of the measurements showed that the efficiency of the individually implemented sensors is sufficient for use in traffic control. As well we can consider the implementation of microwave radars and cameras for video analysis as a replacement for the conventional method of induction loops.
|
|
| |
|
|
Vehicle classification using inductive loops sensors
Halachkin, Aliaksei ; Klečka, Jan (referee) ; Honec, Peter (advisor)
This project is dedicated to the problem of vehicle classification using inductive loop sensors. We created the dataset that contains more than 11000 labeled inductive loop signatures collected at different times and from different parts of the world. Multiple classification methods and their optimizations were employed to the vehicle classification. Final model that combines K-nearest neighbors and logistic regression achieves 94\% accuracy on classification scheme with 9 classes. The vehicle classifier was implemented in C++.
|
guest ::
