DSP Lab Project
DSP Lab Project
Mini Project
Motion Detection from a video recorded using a Still Camera
Krish Mangal 0801EE21 R Harsha Vardhan 0801EE29 R Umesh Reddy 0801EE20
EE321: Digital Signal Processing Lab
Aim
To identify a moving object from a video shot from a still camera and to change its color with
respect to the background.
To identify the exact path of a moving object from a video captured.
Basics of Motion Detection
Moving object detection is the basis of moving object identification and tracking. Moving object
detection from video sequences is an important research these days.
It is used in many places like in intelligent video surveillance, motion analysis etc. Moving object
detection is being done using the techniques like Edge Detection, Background Extraction, Frame
Difference, Optical Flow Method etc. with Optical flow method being the most complicated
among them.
Edge detection gives us the space gradient of the video, while Frame difference and Optical flow
methods gives the time gradient of the video.
Algorithm used
We have used the frame difference and edge detection techniques to perform the analysis of the
moving body from a video recorded using a still video camera.
Firstly, we perform edge detection on each of the frames of the video, and then we compare the
consecutive frames for difference in edges.
From this we get the moving pixels in a frame, we now divide each of the frames into blocks of
small size, and if the number of moving pixels in a block is over a threshold, we consider that
block to have motion, else that block is considered to be without motion.
All of the blocks with some motion are shown in color, while the rest are shown in complete EE321
grayscale.
Estimation of Path: we can estimate the path of a moving object by plotting the position of the
moving object’s centroid in each of the frames and with this we may map the moving object’s
path.
2 | P a g e
EE321: Digital Signal Processing Lab
Edge Detection
Applying edge detection to an image leads to a set of connected curves that indicate the boundaries of
objects, the boundaries of surface markings as well as curves that correspond to discontinuities in
surface orientation. Thus, applying an edge detection algorithm to an image may significantly reduce the
amount of data to be processed and may therefore filter out information that may be regarded as less
relevant, while preserving the important structural properties of an image. Edge detection can be done
using many methods such as Sobels, Roberts, Prewitts, Canny, Kirsch etc.
A Comparison of various Edge Detection techniques, and this shows that Canny Edge detection is the best.
EE321
The Flow Chart for our Algorithm
3 | P a g e
EE321: Digital Signal Processing Lab
Uses
This may be used to modify recordings done through a still camera, in live TV recordings etc.
This may be used to identify the path of a cricket ball when its motion is captured during a
match.
Can be used for Toll booth automation, and also as a part of a larger Intelligent Transport
System.
Proposed Extensions
Calculating the velocities of moving objects can be a good extension to this project. This may also be
implemented on a DSP Kit for further use on field.
References
An Improved Moving Object Detection Algorithm Based on Frame Difference and Edge
Detection by ZHAN Chaohui ,DUAN Xiaohui, XU Shuoyu, SONG Zheng, LUO Min, 2007,
Electronics Department, Peking University, Beijing, 100871.
Edge Plane Detection in Spatio‐Temporal Images by Using Edge Vector and Edge Reliability,
Fumiaki YAMANA, Takahiro SUGIYAMA, Keiichi ABE, 2000.
Covariance Analysis of Non‐Translational Motion‐Compensated Frame Differences, 2005,Anant
G, Veeravalli, W. D. Pan, Mahesh Nalasani, and Manoj Kumar Santha Mohan, Dept. of Electrical
and Computer Engineering, University of Alabama in Huntsville, Huntsville, AL 35899.
EE321
4 | P a g e