Security System

Project Description

We will build a security system which will detect if any person is within a restricted area within a room. The overall system will consist of one base station and two or more camera systems. Each camera system will interface with a camera that is monitoring a room from different angles, use a computer vision library/algorithm to detect and generate bounding boxes for people in the room, and send all this information to the base system. The base system will aggregate all this data, make a decision on whether there are any people in the forbidden area, and additionally display all video feeds and bounding boxes via HDMI.

High-level 'What'

(notes from class 3/11)

• self discovering camera modules
• user determine security zone
• live feed toggleable
  • feed auto toggle
• bounding box for subject
  • max number of subjects?? 10?
• what are we detecting? people? objects?
• HDMI shows text of which camera/camera status, FPS, shapes, RTSP stream

Diagram

Project Organization

This is the main repo for the entire project. Each team has its own branch:

• camera/master
• cv/master
• sysman/master
• hdmi/master

Each team can make additional branches as necessary for each feature, e.g. camera/stream_tx. As we work on the project, we will make pull requests into the main team branch. For full integration, all team main branches will be merged into the main project branch.

Camera Team

Responsibilities

• Connect to camera, retrieve images
• Expose interface to CV team to provide current image via IPC
• Send image data to base station via RTSP/UDP

Members

Ben Palmer (leader)

Emma Brodigan

Muhammad Elarbi

Computer Vision Team

Responsibilities

• Retrieve images from camera driver via IPC
• Process images, detect humans using CV and generate coordinates for boxes around them
• Transmit coordinates of all bounding boxes to base station via TCP

Members

John Privitera (leader)

Jack Horton

Nicole Johnson

System Management Team

Responsibilities

• Overall system info
  • connection between systems: acknowleging and enumerating the camera modules
  • informing each camera module's TX component of any potential instance-specific network config info.
• Aggregate all bounding box data from all camera systems and make decision on if someone is in the restricted area -> send alert to display subsystem

Members

John Craffey (leader)

Sid Chenrayan

Joshua Erickson

Samir Elhelw

HDMI Team

Responsibilities

• Receive video stream via RTSP/UDP
• Receive bounding box data
• Render bounding boxes and camera streams on HDMI

Members

David Boullie (leader)

Tom Doyle

Sam Lambrecht

Joseph Berman

Connor Northway

Matt Downing