detection is done using Open CV on to the Raspberry Pi 3.

Transcription

1 Two Stage Security System for Households using Raspberry PI and Open CV A. Harshitha 1, V Prathyusha 2, K. Ashok Babu 3 1 M.Tech in Digital Systems & Computer Electronics, Sri Indu College of Engineering & Technology, Sheriguda, Ibrahimpatnam, RR.Dist. HYD 2 Associate professor, Dept of ECE Sri Indu College of Engineering & Technology,Sheriguda, Ibrahimpatnam, RR.Dist. HYD 3 Professor & HOD Dept of ECE, Sri Indu College of Engineering & Technology, Sheriguda, Ibrahimpatnam, RR.Dist. HYD Abstract: Smart home security and monitoring have become vital and indispensable in recent times, and with the advent of new concepts like Internet of Things and development of advanced authentication and security technologies, the need for smarter security systems has only been growing. The design and development of an intelligent door lock control system using finger print and face recognition technology, for authentication, remote monitoring of visitors and remote control of smart door lock has been reported in this paper. This system uses Haar-like features for face detection and Local Binary Pattern Histogram (LBPH) for face recognition. The system also includes a webbased remote monitoring, an authentication module, and a bare-bones embedded IoT server, which transmits the live pictures of the visitors via along with an SMS notification. The system has been implemented and tested using the Raspberry Pi board, Python along with OpenCV are used to program the various face recognition and control modules. Keywords: Raspberry Pi, Python, finger print sensor, camera 1. INTRODUCTION The surveillance became a big challenging problem in the present world, sake of security purpose in phone or banks or other public places we are using many different security systems such as password, finger prints and pattern recognitions. The pattern or passwords used can be trapped easily once if the user is known well or if the pattern is seen once or well known. The finger print system doesn t achieve full-fledged result the through put is low because of the miss matches or a layer of distraction due to external sources and many other reasons. To provide a proper surveillance we are going for face recognition, the unique features of each individual are taken into consideration. There are different kinds of methods for face detection and recognition, in this paper face detection is done based on haar features and face recognition is done based on local binary pattern histogram. In this paper the Face recognition and detection is done using Open CV on to the Raspberry Pi 3. Face Detection Many kinds of face detections are used in plenty appliance occurrence management, surveillance eventualities, gaming, human-computer interaction, etc. Viola associated Jones devised an formula, known as haar features classifiers, to chop-chop find any object, as well as human faces, victimization Haar classifier cascades that area unit supported Haar-Like options. Different types of ways area unit out there for detecting the face for identification and recognition. Face detection is using haar like features, so we ll work with face detection. Initially, the formula lots of positive pictures (images of faces) and negative pictures (images while not faces) to coach the classifier. Then we d like to extract options from it. For these, haar features shown in image area unit used. Which are similar to our convolutional kernel. There are line features, edge features and rectangle features. Face Recognition Face recognition applications is categorized into the three categories: verification, identification and watch. Face confirmation part is considered to be a one. The system can compare face image to the face image(s) of a similar registered identity within the information to form call on whether declining or not acceptive the identity claim. In distinction, the face identification task may be a one: N matching drawback. The face image is conferred to the system while not associate degree mark claim and also the system can search through the existing identities within the information of face to compare the conferred face image. Usually, it s considered that the conferred face image belongs to at least one of the themes within the information. Lastly, the watch list task is typically very just like the identification task. but in watch list task, the question subjects square measure usually larger than the themes within the information and thus the question subject might not exist within the information 27

2 Local Binary Pattern Fisher faces and Eigen faces are comprehensive Techniques to face recognition. The pixels are the vector of the data some point in High dimensional vector space. High dimensional vector space gives some ambiguity during face recognition. So, we go for the sub space which are lower dimensions and space where the useful data is stored. Total scatter is maximum In Eigen faces approach if the variance is calculated using external sources it may create a problem. Maximum variance components are not useful for the purpose face recognition, so to store some discriminate data we used a linear discrimination analysis and escalation in the fisher face method. In Fisher face we have to go for N number of data base if we have only one image the co variance will be very high and the through put is reduced. IP NETWOR Finger print In (4,1) LBP case, the reason why the four points selected correspond to vertical and horizontal ones, is that faces contain more horizontal and vertical edges than diagonal ones 2.1 System Architecture RASPBERRY ARM1176JZF-S PI ARM I/O L G U U OPEN GL ES1.1/2.0 VIDEO H264, CORE GPU MPEG2, GRAPHICS ACCELERAT CAMERA CA SDIO 1 0 MONITOR SD CARD Fig: Calculating the original LBP code and a contrast measure Local Binary Patterns Extension In order to treat textures at different scales, the LBP operator was extended to make use of neighborhoods at different sizes. Using circular neighborhoods and bilinear interpolation of the pixel values, any radius and number of samples in the neighborhood can be handled. Therefore, the following notation is defined: (P, R) which means P sampling points on a circle of R radius. The following figure shows some examples of different sampling points and radius: III. SYSTEM DESCRIPTION The raspberry pi based obstacle avoiding robot consists of three main module i.e. ultrasonic, camera module, and raspberry pi. The camera module gets the input image which are obtained are real time operation. The raspberry pi is a platform consisting of all necessary hardware module assembled on it. It receives the images from camera module. Ultrasonic check the distance robot to obstacle occurs then it will send the distance raspberry pi. It gives further instruction to motor driver accordingly. The motor driver actually consists of two sub motors i.e. right and left motor. These motors receives the signal from raspberry pi in case of any appearance of the obstacle in its path the motors work accordingly to signal and moves in left or right direction with the help of left and right motor to avoid the obstacles. 3.1 Imaging Module Fig: Circularly Symmetric Neighbour Sets. The imaging module in the proposed system is realized using a USB web Camera, the main reason behind choosing USB Camera over the Pi camera is the cost effectiveness. The camera features a high-quality CMOS sensor, with an image resolution of 25 MP 28

3 (Interpolated), an adjustable lens for focus adjustment, a frame rate of 30 fps and f2.0 lens. The USB camera also is equipped with night vision for low light photography. The camera interfaces with the Raspberry Pi via the USB 2.0 port and is responsible for capturing images when requested, the pictures are captured by using the command fswebcam. Fig. 4 Raspberry Pi 3 Module In this system the core module plays a highly pivotal role and is responsible for various functions, the core module is responsible for acquiring the images from the camera, processing and storing. It s also responsible for maintaining the facial database which consists of pictures of all the authorized persons for reference. It is in charge of employing the face detection and recognition algorithms and has to decide whether is a person is authorized or not. It s responsible for controlling the door lock module by sending lock/ unlock commands using Python code via GPIO to the motor driver. 3.3Embedded Server & IoT Fig. 3 USB Camera 3.2 Raspberry Pi Core Module The core module of the system is realized using a Raspberry Pi 3 board; it s a $ 35 bare-bones computer designed and developed by the Raspberry Pi Foundation, the Pi 3 features a BCM 2837 System-on- Chip which includes a Quad-Core 64-Bit ARM Cortex A7 CPU clocked at 1.2 GHz paired with 1 GB of RAM. It also has VideoCore IV GPU for graphical processing applications, it also includes four USB ports for peripherals and 40 Pin General Purpose Input Output (GPIO) pins for interfacing the Pi with external electronic circuits, these GPIO pins are used to interface the Pi to the door lock module. The Raspberry Pi is designed to run various Linux based operating systems and has Raspbian as its official operating system and Python as its official programming language. Another crucial function of the core module is to act as an embedded web server, the primary responsibilities of this server include, transmitting the visitor/ visitors images via to the owner for s from the owner and find the security code from the s for authorization. This system employs an embedded server approach for communicating with the user and with the internet/ intranet. Python code is used to program certain aspects of this system such as sending and receiving s and text messages. Standard Python libraries corresponding to the web such as urlib2, cookielib for online service; imaplib, poplib, , smtp, etc. for sending and receiving s are imported and used accordingly. The system is also configured using Apache to act as a server, which is useful to remotely monitor the conditions. The owner can log in to the server using a dedicated static IP assigned to the Raspberry Pi, another important function of this server is to provide a secure back door to lock/ unlock the door by bypassing the face recognition feature in case of a failure or emergency. This is a secret feature and is only accessible by the owner. IV.HARDWARE IMPLEMENTATION This section emphasizes on the actual hardware implementation of the proposed system, the various modules, components, peripherals and the interconnections between them are discussed here. The first stage of the implementation is to prepare the Raspberry Pi 2 module for its first boot; this is done by downloading the latest version of the Raspbian operating system from the official Raspberry Pi 29

4 website. A microsd card is the formatted using SD Formatter; it s then flashed with the Raspbian OS using Win32 Disk Imager. The first boot is then completed on the Raspberry Pi connecting the required peripherals, such as power supply, keyboard, mouse, Ethernet cable, etc. The Raspberry Pi for optimal operation requires a quality power supply; the Pi can be driven by using any Micro USB based mobile phone chargers with a good current rating, and this system is powered by a 5V 2A power bank for uninterrupted operation. Using Python programming language preinstalled on Raspbian the source codeof the system is provided and tested appropriately. The USB Camera is interfaced, the GPIO pins are programmed using commands in Linux and Python in this stage. The camera is interfaced to the Pi via the USB port and the door lock module is interfaced via the GPIO pins on the Pi. V. EXPERIMENTAL RESULTS Result VIII. REFERENCES [1] Gurjashan Singh Pannu, Mohammad Dawud Ansari, Pritha Gupta. Design and Implementation of Autonomous Car using Raspberry Pi. International Journal of Computer Applications ( ) Volume 113 No. 9, March [2] Ayush Wattal, Ashutosh Ojha, Manoj Kumar. Obstacle Detection for Visually Impaired Using Raspberry Pi and Ultrasonic Sensors. National Conference on Product Design (NCPD 2016), July [3] Mr. T. P. Kausalya Nandan,, S. N. Anvesh Kumar, M. Bhargava, P. Chandrakanth, M. Sairani. Controlling Obstacle Avoiding And Live Streaming Robot Using Chronos Watch. International Journal of Engineering Innovation & Research Volume 5, Issue 2, ISSN: [4] Xiaodong Miao, Shunming Li & Huan Shen, On- Board lane detection system for intelligent vehicle based on monocular vision, International Journal on Smart Sensing and Intelligent Systems, vol. 5, no. 4, December 2012, pp [5] A. Bar Hillel, R. Lerner, D. Levi, & G. Raz. Recent progress in road and lane detection: a survey. Machine Vision and Applications, Feb. 2012, pp [6] Narathip Thongpan, & Mahasak Ketcham, The State of the Art in Development a Lane Detection for Embedded Systems Design, Conference on Advanced Computational Technologies & Creative Media (ICACTCM 2014) Aug , 2014 VI. CONCLUSION The face recognition was done using LBPH and raspberry pi platform. To reduce the false-positives drastically and increase the efficiency in this research, we are using haar like features and for recognition of face we are using LBPH (local binary pattern histogram).this reference design can be used for authentication and also finger print sensor in home security, and other public places. Thus for a safety purpose in real time we designed a face recognition system in minimum expenses using raspberry pi, open cv. [7] Narayan Pandharinath Pawar & Minakshee M. Patil, Driver Assistance System based on Raspberry Pi, International Journal of Computer Applications ( ) Volume 95 No.16, June 2014, pp [10] J.M.A. Alvarez, A.M. Lopez & R. Baldrich, IlluminantInvariant Model-Based Road Segmentation. Intelligent Transportation Systems, IEEE Transactions on, 12, 2008, pp About Authors A.HARSHITHA 1, Pursuing M.Tech in Digital Systems & Computer Electronics in Sri Indu College of Engineering & Technology, Sheriguda. 30

5 V PRATHYUSHA 2, Currently working as Associate professor, Dept of ECE, Sri Indu College of Engineering & Technology, Sheriguda. K. ASHOK BABU 2, Currently working as professor & HOD, Dept of ECE, Sri Indu College of Engineering & Technology, Sheriguda. 31