影像式車輛偵測器擴充模組研發－事件偵測功能模組研發與試作

Abstract

本研究計畫首先針對國內外影像式事件偵測系統與研究進行文獻蒐集與回顧，同時針對國內交通管理單位進行訪談，確實掌握產品技術發展與國內使用單位需求，使得本計畫研究成果能夠更符合所需。在研究部分，本計畫將針對影像式車輛偵測器進行事件偵測功能之擴充模組研發與測試，該模組針對車輛壅塞、掉落物、異常車速以及事故車進行自動偵測，同時將事件過程之影像進行錄影保留。除了事件偵測核心演算法，本計畫亦將完成一個後端管理介面，以協助系統操作人員進行事件發生相關應變措施。在實機測試部分將陸續完成自行架設攝影機地點ㄧ星期資料分析以及進入交通管理單位進行為期一個月的試作。預期於本計畫能夠完成事件偵測擴充模組雛型研發。

The main purpose of this project focuses on the expansion module development of the incident detection for the developed vision-based vehicle detector. In the beginning of our development, some control centers will be visited since their demands are very important for us to design the module exactly. The proposed incident detection module in this project will detect several incidents including traffic jam, dropped objects, unusual speed and the accidents. Simultaneously, the videos of the incident will also be recorded in the vision-based vehicle detector for the evidence and the identification. For the operators, a friendly user interface is necessary, which will also be involved in this project. Finally, the proposed module will be installed in the traffic control center for real test to verify its performance.