| 英文摘要 |
For the requirements from Environmental Management Administration （EMA）, this project consists of following works: maintaining and operating the platform of illegal dumping potential zone, performing spatial correlation analyses and risk maps for specific key wastes, generating characteristic model using GPS trajectory data from transportation machinery. Additionally, we produced a risk alert list for illegal dumping and cross-referencing with manifests. Moreover, we conducted remote monitoring and managements of targeted areas using advanced technologies such as drones and surveillance cameras.
In addition to continually optimizing the functionality and operational quality of platform, this project provided 46 sessions of drone monitoring data inquiries and overlay-based zoning queries for risk maps across 19 counties. It also introduced spatial correlation analyses between sites and specific waste sources into the existing risk model, resulting in risk maps for ash and slag waste, with validated prediction accuracies of 96.87% and 78.64%, respectively. Moreover, we proposed an algorithm to trace potential waste sources with those risk maps.
In terms of the applications of those risk maps, pilot research in the high-risk area of Fangyuan Township, Changhua, utilized intelligent checks based on risk maps, identifying eight suspicious anomalies that were referred to the Environmental Bureau for investigation. Additionally, the project ranked risks for 387 freeway interchanges around Taiwan, and provided monitoring equipment deployment plans for 24 interchanges. Analyzed the risks of high-potential illegal disposal area in the southern region, and employed two road network calculation models to prioritize the locations of surveillance equipment.
Furthermore, the project employed transportation vehicle trajectory risk model to approximately 4.09 million GPS tracking data points from waste transportation machinery, identifying 220 abnormal trajectory segments and 213 of them without corresponding manifest records. In collaboration with the EMA and three centers of environmental management, the project expanded the practical application of the trajectory risk model through numerous analyses, such as suspicious vehicle trajectories, trajectory overlay, passage rate, and abnormal stop and start point evaluations.
This year, drones were used to monitor 15 key targets, producing orthophoto mosaics, digital surface models, and 3D volumetric models of sites. Additionally, a site in Changhua was selected for evaluating the integration of multi-spectral aerial systems.
|