| 英文摘要 |
The illegal disposers take advantage of the convenient transportation, such as Provincial Highway 61 and National Highway No.1, and transports waste through cross-county and city express roads for disposal. To regulate waste removal operators, the Waste Disposal Act stipulates that waste removal vehicles must install a Global Positioning System (GPS) real-time tracking system. The practical experience has shown that many unscrupulous operators deliberately disable their GPS, and the limitations of the GPS system's functions cannot prove that the operators have turned it off. In addition, some operators also use unregulated vehicles for waste removal because the trajectories of these vehicles cannot be traced, increasing the difficulty of illegal dumping investigations.
To address these challenges, the goal of this project involves: (1) installation and maintenance of the monitoring equipment. (2) construction and improvement of the AI geofencing system. (3) analyzing the removal vehicle abnormal transportation patterns to construct the tracking module and how the system notified the relevant inspectors. (4) Promoting the AI geofencing system to expand. (5) the administrative tasks and duties.
This project involves the monthly maintenance and inspection of 11 existing monitoring points, as well as the installation of 5 new monitoring points. The goal is to ensure that the accuracy of each monitoring point reaches 75% during the day and 60% at night. To improve the accuracy of the alarm correlation, the project implements a dual verification mechanism for license plate recognition, ensuring that the accuracy of each monitoring point reaches 95%.
This system has also introduced several new features this year, including: Real-time video streaming, Video clip of large vehicles, Optimized GIS and the searching function of the license plate recognition list, system’s description and text, Blacklist and alarm functions, Permission-based design tailored to the Environmental Protection Bureau to enhance the applicability of smart fence license plate recognition data, while ensuring that each user operates within the defined access rights.
The system also continues to incorporate vehicle type recognition based on license plate data, focusing on monitoring waste collection vehicles. Using object recognition AI technology, the system identifies what materials the vehicles might transport to prevent the potential for illegal disposal.
To accelerate the expansion of the AI geofencing system coverage, the project assists in collecting license plate recognition data from six agencies. As of the end of November, the system has a total of 135 monitoring points. Additionally, the system includes a data transmission disconnection reporting function to continuously monitor the transmission status of license plate recognition data across all monitoring points.
The project also leverages GPS trajectory data and land-use classification information to establish a correlation logic for detecting high-risk vehicles that stop abnormally in specific land-use zones (such as agricultural, farming, or livestock areas). This has led to the verification of 38 alert cases. Through the integration of various system data and correlation methods, high-risk vehicles can be detected earlier, preventing illegal disposal incidents.
|