REMOTE HOME MONITORING SYSTEM USING IOT DEVICES.

Reenu Varghese 1 , Saju A 2 and Kuruvilla John 2 . 1. PG Scholar, Department of Electronics and communication Engineering, Believers Church Caarmel Engineering College, Ranni-Perunad, Pathanamthitta,689711. 2. Associate Prof., Department of Electronics and communication Engineering, Believers Church Caarmel Engineering College, Ranni-Perunad, Pathanamthitta,689711. ...................................................................................................................... Manuscript Info Abstract ......................... ........................................................................ Manuscript History

The main aim of this work is to the IoT concept using cloud server [2] to transfer the images. The motive behind the work is to develop a motion sensing (image processing) algorithm for IoT device considering the resource limitation and storage space limitations of such device, Further, the whole cost of the system is designed to be low. The figure  1 shows internet of things, it inter-connect with whole world. In 2020 the most of the world things are interconnect with IoT device.

Motion Detection: -
The existing systems use motion detection kits that usemotion sensors and Ethernet modules to send or receive data capturedby the sensors. This requires an external server to processthe data. The external server processes the data live streamed by the sensors. There is no local storage of the data in such systemsand if the cloud server is insecure, sensitive data can be leaked. Further, present secure systems use heavy data encryption algorithms to securely upload the data to the server. This approach is impractical in IoT in common IoT devices due to the hardware limitations.
Present systems also require large cloud storage to store such data and can use up the home internet bandwidth. The cost of such implementations are high. So, ordinary home users need an alternate method to reduce the cost of implementation using IoT based security hardware -without loss in efficiency and use ability. Further the requirement of existing systems to be always connected to the cloud server can handicap the functionality of the security of the intruder alert/notification function of the system since any problems in connectivity to the remote server or problems in the server can hinder communication between the sensors and the user.  In this work we can use the Gaussian Mixture Model algorithm [3] [4]. This algorithm is generally used as a preprocessing stage in large image process project. The algorithm uses low resource. In addition to background subtraction, this also has ability to neglect small changes in the background such as changes in lighting condition etc which should not be perceived as motion. The segmentation approach also allows shadow [4] from the images to be removed.  Background Subtraction: -For motion detection, the background needs to be separated from the foreground. Background subtraction or Foreground Detection is used to extract an images foreground for further processing and object localisation.The pixels which remain static between frames are considered as the background and the pixels which are displaced from frame to frame are considered as the foreground. The absolute value of the pixel intensity differences difference between frames is obtained by a simple subtraction.

Image segmentation for motion detection: -
The goal of segmentation is to simplify and/or change the representation of an image. Our approach to image segmentation is to represent the background of the image as black (0) and the pixels where the motion is taking place in the image, called the foreground it represented in white (1). Finally, only regions where data is clustered are considered and unwanted data is removed.

Image segmentation for motion detection Minimum area threshold: -
If the contour area of the foreground segment is larger than our supplied minimum area called the threshold, motion can be considered to be detected. This method of comparing the foreground segment with that of the threshold area also helps to neglect any relatively small motion, like a paper moving in the breeze.
In this work, python script is used for motion detection. We can place camera in particular position and continuously monitor the area. If there is considerable difference between two consecutive [5] recent image the notification of motion detection is send to email and user phone as SMS. Through the website we can see the room from where ever we are and we can control raspberry pi 1900 Fig 6: -Controlling raspberry pi remotely

Conclusion: -
The motion detection less space to store the data. Data privacy and safe communication is also considered to be achieved in the new communication system. The motive behind the project is to develop a motion sensing (image processing) algorithm for IoT device considering the resource limitation and storage space limitations of such device, Further, the whole cost of the system is designed to be low. Using this project, we can identify intruder motion. Which help to secure the area.