整合地理資料之多攝影機移動物件追蹤技術研發及應用

The Internet of Things (IoT) has been attracting many attentions in recent years.The goal of the IoT is to connect the sensing and tasking capabilities of everyday objects to achieve automatic and efficient applications. Among the sensing capabilities, remote sensing sensors like cameras produce images and videos that require images processing to fit application requirements, such as optical Inspection. However, the connectivity characteristics from multiple sensors is seldomly utilized. Therefore, this project focuses on the integration, analyses, andapplication of CCTV cameras. To be specific, this project discusses (1)integration of heterogenous CCTV resources, (2) moving object tracking from multiple CCTV camera images, and (3) applications of moving object tracking.In terms of the integration of CCTV resources, as various organizations operate their own CCTV networks, the integration of camera resources is difficult. To address this issue, apply international open standard to unify the data model and protocol is a necessary step. Regarding the moving object tracking from multiplecameras, this type of Multi-Target Multi-Camera (MTMC) tracking is usually achieved by identifying and tracking objects on one single image, and applying re-identification methods to compare the similarities between objects. Existing researches usually compare objects’ appearance, such as facial features, color of clothes, gender, etc. Although some researches utilize spatial and temporal