Loading presentation...
Prezi is an interactive zooming presentation

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Make your likes visible on Facebook?

Connect your Facebook account to Prezi and let your likes appear on your timeline.
You can change this under Settings & Account at any time.

No, thanks

ROBUST SEGMENTATION OF MOVING VEHICLES UNDER COMPLEX OUTDOOR

No description
by

Sarath Kumar G

on 26 September 2013

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of ROBUST SEGMENTATION OF MOVING VEHICLES UNDER COMPLEX OUTDOOR

ROBUST SEGMENTATION OF MOVING VEHICLES UNDER COMPLEX OUTDOOR CONDITIONS
The aim of the project is to develop a technique to detect and segment moving vehicles in a video sequence under non-ideal outdoor conditions like sudden change in illumination, snowfall, fog etc.

It deals with motion detection and segmentation of traffic vehicles in outdoor environment, particularly under non ideal weather conditions.


Existing System
Gaussian-based background modeling, background separation, code book, frame differencing are the existing techniques.

They cannot effectively deal with non-ideal weather conditions. Such dynamic changes results in false motion detection.

Frame differencing involves comparing successive frames and finding areas having a pixel difference greater than a threshold as foreground regions.

Existing System (cont..)
It is affected by camera noise , small camera movements and adverse weather conditions.

Background modeling keep the copy of background and subtract current frame from background frame.

The major challenge is to create an accurate background model in non-ideal conditions.
Proposed System
This system detect and segment the moving vehicles by making use of dynamically adaptive threshold by the Full-search Sum Of Absolute Difference (FSSAD) algorithm.

Efficiency is increased by localizing object of interest, suppressing noise, and reduce false motion that overshadows object of interest.

Motion energy is calculated by sequence of frames.
Proposed System (cont.)
It is used to differentiate between moving vehicles and dynamic background.

As FSSAD is computationally expensive modification of is done to reduce the false motion and computational efficiency.

Modules
The modules are:-

1. Motion Detection
2. Computation Of Dynamic Threshold
3. Blob Detection
4. Boundary extraction
5. Suppression Of False Motion


1. Motion Detection
Block-based motion estimation is used for identifying moving regions in video.
Blocks of reference frame is matched with current frame.
If the block being matched is not in the same location , then it has moved.
The reference frame Rf is divided into candidate blocks.
Every candidate block carries out search within specified area in current frame, called search window.
2. Computation Of Dynamic Threshold
To identify the foreground and background,threshold value is needed. As the outdoor sequence has dynamic background,an adaptive background is needed to be processed from each frame. The motion energy of the frame is obtained by summation of absolute differences of pixels.
The energy observed at the positions of moving vehicle is more than twice compared to the background having non-ideal conditions and vehicles have a solid cohesive shape. The Average Motion energy per pixel and frame is calculated. From that Adaptive Threshold value is calculated.

3. Blob Detection
The motion blocks are grouped based on their relationship between neighboring blocks.
The block connectivity notation is used to describe relationship between two or more blocks.
Different approaches have been used to perform connected-component labeling which generally fall under the category of either breadth first search or depth first search (DFS).
Full transcript