Research

Vehicle Tracking using Object Detection via Optical Flow

This work is an extension of my previous work "Traffic Monitoring and Control using Image Processing"

My concentration was on

  • Improvement of threshold method during background elimination.
  • Analysis of object movements to remove static object movements.
  • Implementation of previous research in field data.
  • Intra-analysis of detected object to remove object cluttering
The project was dropped off in middle due to my masters
 


Actual Frame  Optical flow Calculation  

Fault Tolerance in Multi-Core Processors

This is my Master's Research at  Anna University (MIT Campus), Chennai.

The concentration is on development of a Fault Tolerance Scheme for commercially available Multi-Core Desktop Processors with little or no hardware change. The scheme is based on software methods assisted by Hardware and Operating System.

My concentration was on
  • Coverage of both permanent and transient faults with little performance overhead.
  • Mechanism to handle Interrupts, Process Preemption, etc.
  • User control over processes for fault tolerance.

For my research and analysis I used the SimpleScalar toolset.

  Intel Quad-Core

Traffic Monitoring and Control using Image Processing

This is my under graduation research completed on late April, 2012.

Our research was to develop a method to track vehicles in route leading to the computation of speed, direction of travel and possibly vehicle Type.

Our focus was on

  • Using traffic specific details to to speed up computation.
  • Using optical flow between frames to remove background details.
  • Mapping Vehicle's image pixel to actual object.
  • Calculation of vehicle's actual dimension.

Video Frame Optical Flow calculation After Enhancement   After Thresholding Vehicle Detection Final Result