unusual(moving) object recognition in video surveillance

The following goals and objectives state in general and specific terms what this project

should accomplish:

Goal #1 – Develop a GUI to a video surveillance application with simulated video feeds from IP cameras.

Objective 1.1 – Create a Document/View style user interface with a System Resources View, a Live-Video View, and an Archived Video View.

Objective 1.2 – Develop a way to use pre-recorded video (.avi files) to

simulate live video feeds from IP cameras and playback of archived video streams.

Objective 1.3 – Develop the System Resources View as a tree view with

different leafs/nodes being added to control system resources such as the server settings, the host computer settings, or the IP camera settings.

Objective 1.4 – Develop a Live-Video View that allows live viewing of the cameras video feed.

Objective 1.5 – Develop an Archived Video View that allows playback of an archived video stream with controls similar to a DVR

Goal #2 – Recognize the patterns of different objects (like human or animal or vehicle etc) and classify into different classes.

Objective 2.1 – all moving objects are detected using a temporal differencing algorithm which are called motion regions

Objective 2.2 –Each motion region is classified at each time frame using an image-based classification metric

Objective 2.3 – Classifications for each individual motion region are recorded over a period of time, and a simple Maximum Likelihood Estimation (MLE) criterion is used to correctly classify each target.

.NET Pengaturcaraan C#

ID Projek: #1593096

Tentang projek