Title The Service and Enhancement Project for the Information Integration of Traffic Events (2/2)
Dept Transportation Technology and Information Division
Year 2020
Month 7
Price 250
Summary        This project extends the previous (2018) project implementation results about traffic incident data platform, continues to expand the collection of all types of incident data, completes the traffic accident data exchange specification standard, and makes progress in developing the functions of theme map production, visual presentation, data statistics and traffic incident information API, to improve the efficiency and application level of traffic incident management. In addition to transfer this platform to the management and use by the Transportation Bureau of Kaohsiung City, also duplicate the experience of traffic incident data integration services and transferred to Tainan City, from the perspective of division of labor from the central and local governments, to construct and complete the "National Traffic Incident Integration Data Exchange Service Platform" and the "Urban Traffic Incident Data Integration and Publication Platform" in Kaohsiung City and Tainan City at the same time.  Besides, in order to improve the auto-detection function for traffic incidents, this project continues the previous project achievements. In addition to part of the equipment originally installed at the intersection of Minzu 1st Road and Dazhong 1st Road in Kaohsiung City, due to the factors of communication power influence, screen overlaps and poor shooting angle, it was transferred to Jianguo Interchange for incident detection in another field, and 12 artificial intelligence image detection devices were installed at the intersection of Zhonghua Road and Zhongshan South Road in Tainan City to carry out detection and identification of traffic incident through the improved Single Shot MultiBox Detector (SSD) deep learning model, and has defined reasonable identification logic for different types of incidents and implemented the practical operation. In addition, the accuracy of the traffic flow parameters under different scenarios has also been improved, and the QKU dynamic graphs of flow rate (Q, Vehicle/hr), density (K, Vehicle/Km) and rate (U, Km/hr) can be automatically produced by the detected traffic flow parameters. In order to make the overall traffic incident management platform and automatic incident detection more complete, in addition to the setting of follow-up rights and obligations for all units to use, this Project also summarizes the ideas and opinions of other units through a number of education and training and marketing publicity meetings as the reference for follow-up promotion.
Post date 2020/08/18
View count: 116