Visual surveillance is widely used in different fields such as security, transportation,
medicine, and logistics. However, these videos and images require appropriate storage
and computing platforms. Cloud based surveillance systems are suitable for big data
processing and storage.
This thesis will discuss how architecture and function modules of a cloud based system
can be created, a cloud-based intelligent surveillance system (CISS) can be published
on a private cloud using a Virtual Machine (VM). Our users are able to connect a camera
at a desired location to CISS so that the system can capture videos and images, and then
push notifications to a web page where the users can watch the videos via the cloud.
This thesis presents an in-depth demonstration of how the portable and efficient cloud
based surveillance system (CISS) works. It makes full use of the cloud computing
technology, and applies computer vision technology to visual surveillance. The system
is able to quickly and accurately detect events from streaming videos, and apply the
feature of push notifications to achieve event based alarming in real time. The new
technology makes CISS more advanced than the conventional visual surveillance
system.
The overall objective of this thesis is to implement four key functionalities, (a)
streaming video input, (b) intelligent visual surveillance (IVS), (c) video transcoding
and storage in real time, (d) message push and media streaming output.
