You are on page 1of 13

OBJECT mOTION

DETECTIOn
Table of Contents:
• Abstract
• Explanation
• Functional and Non-Functional Requirements
• Software and Hardware Requirements
• Technology used
• Design
Abstract:
• In recent years, motion detection software has been generally accepted as
one of the most critical security features. It is used to improve existing
monitoring systems such as motion sensor lighting on indoor and outdoor
surveillance cameras. One of the most popular devices that employ motion
detection technology is security cameras.
• This project detects the movement of objects in front of the camera
Explanation of Project:
• We'll start with one video file as input and then use OpenCV to convert it to
"n number of frames." and save the frames in the appropriate format with a
proper name. Here we do 30 frames of video.
• Image resizing is used in frame separation.In that video, the first frame is
considered the background image (or previous image), while the rest of the
frames are considered the current image.
• We get difference image by subtracting background image from current
image. Prior to background subtraction, each current image must be
converted from RGB to Greyscale.
•We use the following thresholding for this difference image: Image difference
>35=1 otherwise 0. Difference image >100=1 for dynamic thresholding.Otherwise,
the value is 0.

•After background subtraction, we'll use loop Fn to update the current picture.

•Finally, we'll use the median filter to remove noise.

•We trace an object using Morphological functions. If the background image and
the current image are identical, it means that there is no moving object detected.
To track multiple objects, we use different outboxes of different colours.

•By comparing the Ground Truth image to our output image, we estimate
parameters such as MSE, Entropy, Precision, PSNR, and so on.
Functional Requirements:
 Enables the viewers to detect the distance between the people
• The system takes the real time video through the camera installed and enters them
using the deep-learning technique to identify the distance between the people
 Viewing the distance between the people by indicating the amount of
risk
• The deep-learning technique helps in binding a color box around the people and
indicate different color according to the distance they are maintaining and type of
risk they are facing. The type of risk is identified in accordance with the color of
box.
Non-functional requirements:
o Compatibility
• System should be compatible and integrate-able with the Moodle as it can be further
added to the new feature of Moodle
o Ease to use
• A person will interact with the system to generate the distance report through a user-
friendly graphical user interface. Furthermore, the generated reports will contain
visual representation for the results.
Software and hardware requirements:

• Software Requirements • Hardware Requirements


• Python interpreter • PC (4GB RAM)
• Camera (min 2MP)
METHODOLOGY
This system follows following stages to track moving object.
Input video
•Actually, the input video is taken from a real-time camera as well.
Frame Separation
•In this project, we will use python frameworks to convert input video to frames. The term video
refers to the combination of frames.Every video has a number of frames; by using that value, we
can calculate the number of frames.
Current Image & Background Image
•After transforming video to frames, the first frame (image) is referred to as the background
image, and the subsequent frames (images) are referred to as the current image, and these images
are not similar at the time of moving object detection video. Moreover, when it comes to non-
moving object detection, these current images are identical.
METHODOLOGY
Background Subtraction
•Background subtraction implies that we simply subtract the current image from the background
image, and the current image is changed each time, while the background image remains
unchanged.
•We can easily find the moving object by employing these techniques.For example, Background
image pixel – current image pixel.
•If the output is 0, it indicates that no moving objects were observed.
•If the output is 1, it indicates that a moving object has been observed.
Preprocessing & shape analysis
•Using this method, we can determine the shape of a moving object (if the background image and
current image differ). Otherwise, no moving object is observed (if there is no difference between
the background image and the present image).
Technology used:
o Language
• Python
o Techniques
• Background Subtraction
• Blob Classification
Design:
Thank You

You might also like