Synopsis Proj

You might also like

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 4

CAMERA MOTION TRACKING AND AUTOMATIC BACKGROUND AND FOREGROUND SEGMENTATION OF MOVING CAMERA DATA

AREA :
Software project based on Image and Video Processing Using MATLAB

OBJECTIVE:
The objective of our project is to develop an algorithm which can detect the vehicles and pedestrians and track them on a road using a moving camera fixed on vehicles . The simulation of this algorithm is done using the MATLAB. The project can be further extended to the detection of Lane and other moving and stationary object like Traffic lights etc .

PRINCIPLE:
Target detection is the process of compensation of camera motion and localizing those areas in the image where a potential target is likely to be present. Conventionally the camera motion detection technique is based on the shift in the background of image sequences and detection technique of foreground is based on the contrast between target and its immediate background. Once the global motion is estimated and target is localized, it is extracted from the background as accurately as possible in the segmentation step. Moreover contour of moving object can be extracted by simple frame difference of previous image sequence of video and motion compensated present frame sequence of video.

THEORY:
The steps into which the algorithm for automatic target contours detection of moving objects for moving camera are the following:

The first step involves the motion estimation of camera. The second step involves motion compensation of the camera with respect to previous frame. On subtraction of motion compensated present image frame with previous frame, contour of moving object is extracted. For automatic detection of intrusion in the moving camera, the first step involves motion estimation of moving camera. Second step is background learning. Third step involves the background subtraction for each incoming frame to get the foreground blob.

The most common paradigm for performing background subtraction is to build an explicit model of the background. Foreground objects are then detected by calculating the difference between the current frame and this background model. A binary foreground mask can be constructed by classifying a pixel with an absolute difference above a threshold as being from a foreground object.

However the motion estimation algorithm and background subtraction should be robust to ignore the movement of small background elements and capable of incorporating new objects into the background model. The computations have to be minimum to allow real time implementation. Same is the scenario with the global motion estimation algorithm

In the last few decades a large amount of work has been done in designing and developing algorithms for moving objects detection in complex scenes for a static

camera but for a proper video surveillance system a proper robust algorithm for a moving camera is to be developed.

ALGORITHM TO BE DEVELOPED:
Global Motion Detection using EDGE based method using Gaussian Filtering and Enhanced SHIFT method . The algorithm can be named as EBM- NPFA (Edge Based Method-N Pass Filtering Algorithm).

BLOCK DIAGRAM

APPLICATIONS:
Moving objects detection is widely employed in many applications such as video compression, video surveillance and reconstruction of useful 3D motion information. It can even be used in Spacecrafts etc .

GROUP MEMBERS:
Ankur Nigam Asma Irshad 0812231015 0812231027

Avinash Chauhan 0812231028

You might also like