supervisor: mr. j. connan. the main purpose of system is to track an object across multiple frames...
TRANSCRIPT
Security Tracker By
Jered Olckers
SUPERVISOR: MR. J. CONNAN
Introduction
The main purpose of system is to track an object across multiple frames using fixed input source.
Input real-time video source
Extract frames Back-Ground subtraction
Detect movement Draw rectangle around object
Current Operation State
Capture code - Changed from VB to VC++ - Merge VC++ with C++
RGB-to-HSV conversion - Skin color detection
Level of accuracy close to 96.5%
Changes
Accordingly scale trace Area Get Accumulative Pixel Average(APA)
value inside the square Store and Compare values
Display object matched Results
Changes(continue)
Start
Movement Detection Implementation
Get threshold values Draw square
A [i--]<10 && A[i]>50
Design Old Design Limitations
- Physically need to open video for extraction - Cannot detect video input source
New Design-Operates in real time-Result of Background
subtraction
Tools & Languages Used
Tools Description
Microsoft Visual Studio 2008 IDE
Microsoft SDK Libraries
Code Blocks IDE
M-encoder Convert image format
Gimp2 Diagrams
Tools
Languages Implementation
C++ Back-End Calculation
VC++ Used to create interface and capture/display frames
Batch Used to connect connect programs (inter-link)
Tools & Languages Used
Languages
Project Timeline
Term3 Term4
BG-Subtraction detect movement Implement skin detection
Extend functionality on interface Track object across camerasScale tracked square
Testing for correctness. Testing and evaluating software
Display Tracked object Create User Manual.
Skin Detection in Luminance Images using Threshold Technique: Hani. K. Almohair, Abd Rahman Ramli, Elsadig A. M, Shaiful J. Hashim
A Robust Vision-based Moving Target Detection and Tracking System.
Frame-skipping tracking for single object with global motion detection!: Ming Anlong, Ma Huadong
References