• No results found

Disaster Autonomous Robot - Finding People through Face Detection Using ARM-11

N/A
N/A
Protected

Academic year: 2020

Share "Disaster Autonomous Robot - Finding People through Face Detection Using ARM-11"

Copied!
6
0
0

Loading.... (view fulltext now)

Full text

(1)

ISSN(Online): 2320-9801

ISSN (Print) : 2320-9798

I

nternational

J

ournal of

I

nnovative

R

esearch in

C

omputer

and

C

ommunication

E

ngineering

(An ISO 3297: 2007 Certified Organization)

Vol. 4, Issue 3, March 2016

Disaster Autonomous Robot - Finding People

through Face Detection Using

ARM-11

Hemali M Gauswami1, Harsha A Gupta2

PG Student, Dept. of Electronics & Communication, SOCET, Ahmedabad, Gujarat, India1

Asst. Professor, Dept. of Electronics & Communication, SOCET, Ahmedabad, Gujarat, India2

ABSTRACT: This paper presents the implementation of an autonomous mobile robot which is capable avoiding obstacles and aims to find human faces and human body. In a disaster such as a landslide or earthquake, flood, Tsunami, windstorms, Tornado, where it is possible that the initial condition of danger persists as would be the case of an earthquake, in which within a building beams and parts of the structure remain committed, as well as the people who risk their lives finding victims. Aim of this robot is to improve the auto navigation to search & detect human body and multiple face detection. In response, the designed robot provides a solution by using artificial intelligence algorithms and image processing running on embedded systems. Using the face detection algorithm Haar cascade classifier it will detect the image of whole face, eyes, ears, and mouth. By the use of python shell, Robot will detect the face through the Haar cascade face detection Algorithm. Physically Robot will be as small as possible with the characteristics of all terrain vehicles.

KEYWORDS: Face Detection, Haar Cascade Classifier Algorithm, Semi Autonomous Control, Machine Vision, Image processing.

I. INTRODUCTION

There is many different kind of catastrophe in natural and man-made disaster: earthquake, flooding, hurricane and they cause different disaster area like collapsed building, landslide or crater. During these emergency situations and specifically in urban disaster many different people are deployed like policeman, fire fighters and medical assistance. They need to cooperate to save lives, protect structural infrastructure, and evacuate victims to safety.

Among the new electronic devices developed, necessary to implement in the electronics of a mobile robot raspberry pi is selected in [1]. Generally this minicomputer is being used to implement the hardware of control required in many applications. These autonomous mobile robots are helpful because they are small enough to enter confined spaces. Detecting and tracking multiple humans on mobile robot platforms still remains a challenging task [2].

The aim of this project is to detect a victim in a disaster environment. This is a very difficult task especially in the unstructured environment of a collapsed building. The physical parameters of a victim that can be detected using different kinds of sensors are: Voice, Temperature, Clothing texture, Motion, Skin color, Shape.

Through the face detection procedure the face structure of the victim can be found out. This detection system is based on the python programming. This runs on an embedded raspberry pi system, which has a camera to capture images. Face detection is to detect and extract face, which is a fundamental issue in many computer vision applications [3].

II. RELATED WORK

(2)

ISSN(Online): 2320-9801

ISSN (Print) : 2320-9798

I

nternational

J

ournal of

I

nnovative

R

esearch in

C

omputer

and

C

ommunication

E

ngineering

(An ISO 3297: 2007 Certified Organization)

Vol. 4, Issue 3, March 2016

AM-CC face algorithms are superior to the other two methods whether in detection rate or the detection average time, and it is very robust to the changes of illumination, face pose, facial expression etc, it can be applied to the real time face detection system. In [6], the author demonstrates that the contributions of geometric constraints, navigation information and depth-based skin segmentation to face detection are remarkable. Based on their combination, mobile robots are able to localize human faces in real time, and interact with humans in a more reliable way. In our future work, we focus on a promising direction of face tracking using navigation information. When human moving directions are repetitive, a robot can compute prior probabilities to predict the next face positions in 3D space.

III.FUNDAMENTALS

The basic thought of face detection is to get the face features by the statistical method to construct face model and compare to the matching level of detecting region and face model. using the Haar feature based cascade classifier to detect things like faces, eyes, noses, and so on. The key advantage of a Haar-like feature over most other feature is its calculation speed. For the face detection, algorithms needs a lot of positive images (images of faces) and negative images (images without faces) to train the classifier.

Object Detection using Haar feature-based cascade classifiers is an effective object detection method proposed by Paul Viola and Michael Jones. Here we will work with face detection. Initially, the algorithm needs a lot of positive images (images of faces) and negative images (images without faces) to train the classifier. Then we need to extract features from it. For this, haar features shown in below image are used. Each feature is a single value obtained by subtracting sum of pixels under white rectangle from sum of pixels under black rectangle.

Fig.1. Haar like Features

(3)

ISSN(Online): 2320-9801

ISSN (Print) : 2320-9798

I

nternational

J

ournal of

I

nnovative

R

esearch in

C

omputer

and

C

ommunication

E

ngineering

(An ISO 3297: 2007 Certified Organization)

Vol. 4, Issue 3, March 2016

Fig.2. Image

IV. FACEDETECTION

Face detection is a very active research topic in the field of computer vision and pattern recognition. It is widely applied in the identity authentication, man-machine interface, visual communication, virtual reality, management of public security files, content based retrieval and many other aspects [7]. Because of the complexities of face image, explicit description of the Facial Feature has certain difficulties; therefore other methods which based on the statistical model have received much more attention.

This method considered human face region as one pattern, namely template, constructed classifier by training a lot of “Face” and “non-face” samples, and then determined whether the images exist human face by analyzing the pattern of the detection region [8]. The objective of the Face detection is to determine whether these exists any human faces in Image judge or not. If exists, then return the face’s spatial location.

In [1] the position detection system is based on a web camera located on the platform; through this camera the scene information of the reference path is obtained. This requires the identification of the corresponding pixels to such reference, which in this case consists of a small red circle. The Detection system is based on python programming, which runs on an embedded raspberry pi system, which has camera to capture images, as seen in Fig. 3. The raspberry pi has an embedded Linux operating system called raspbian, and it has installed the machine vision library developed by intel and called open CV.

Fig.3. Raspberry Pi with camera

PROCEDURE FOR FACE DETECTION:

(4)

ISSN(Online): 2320-9801

ISSN (Print) : 2320-9798

I

nternational

J

ournal of

I

nnovative

R

esearch in

C

omputer

and

C

ommunication

E

ngineering

(An ISO 3297: 2007 Certified Organization)

Vol. 4, Issue 3, March 2016

Step1 Choose the observe face images;

Step2 Determine the size of face area and select face area; Step3 Normalize selected face areas into the same size;

Step4 Compute the average value of every corresponding pixel of face area;

Before constructing template, some images including observe face are chosen. At first, the size of face area is determined. Then, the face area selected manually in these face image. The number of the selected face area is n. The matrix vectors of face areas are distributed independently, so are the pixel values in same position of these face areas. There are single face detection and multiple face detection and it is quite difficult to detect single and multiple face in danger area.

V. RESULTANALYSIS

Fig.4. Flow chart of Hardware

Face Detection Procedure by the use of IDLE (Integrated Development Environment) python 2.7.10 shell version through windows camera application. First open python 2.7.10 shell show in Fig 5.

(5)

ISSN(Online): 2320-9801

ISSN (Print) : 2320-9798

I

nternational

J

ournal of

I

nnovative

R

esearch in

C

omputer

and

C

ommunication

E

ngineering

(An ISO 3297: 2007 Certified Organization)

Vol. 4, Issue 3, March 2016

In this shell open the Haar Cascade Classifier Face Detection Algorithm. And then run this algorithm. By the using of CyberLink YouCam Detect single Face With Eye Detection shown in Fig.7. First try to Detect single Face in Fig 5. Single Face Detection With Eye Detection. Two Face Detect With Eye Detection in Fig 8.

Fig.7. Single Face Detection With Eye Detection Fig.8. Two Face Detect

Then Face Dtection Procedure is also run by the ARM-11 Board Using of 5MP Camera.Setting up the camera module and then it connect with the ARM-11 board. When unboxing the camera module, it is important to avoid exposing the camera to any static electricity that may build up on surfaces and clothes. The camera module is quite sensitive to this, and this has been a common cause of failure of the camera module.

There are two long thin connectors in raspberry Pi; one near the Ethernet and HDMI ports, and one near the power and GPIO ports (the exact positioning depends on the model of the Raspberry Pi but this is a good guide for all boards). The connecter for the camera is the one near the Ethernet and HDMI ports, This can be avoided by touching an exposed earth connection.

Camera connect with ARM-11

Now that the camera is connected, we will enable it in Raspbian using the raspi-config utility.

1. Open a terminal and start the utility using the following Command:

sudo raspi-config

2. Using the arrow keys, navigate to the option 5 Enable Camera shown in Fig. 9and select it by pressing Enter:

Fig.9. Enable Camera

(6)

ISSN(Online): 2320-9801

ISSN (Print) : 2320-9798

I

nternational

J

ournal of

I

nnovative

R

esearch in

C

omputer

and

C

ommunication

E

ngineering

(An ISO 3297: 2007 Certified Organization)

Vol. 4, Issue 3, March 2016

4. Press command on terminal of Linux:

raspistill -t 5000 -o image.png

5. Using raspistill command will produce a capture every 2 seconds.

6. We can also use the following command to save a still image as a JPEG file. Here, the –t parameter controls the amount of time in milliseconds between the camera preview starting and the image being captured.

7. For example we enter a command on terminal of Linux and get the result in Fig.11:

[Raspistill –t 2000 –o image.jpg]

Fig.11. Result of Face Detection

There are 3 commands of raspberry pi camera module, raspistill and raspistillyuv both are very similar and for capturing images. Raspivid is for capturing video.

raspistill

raspivid

raspistillyuv

VI.CONCLUSION AND FUTURE WORK

Hence, concluded that the using of ARM-11, implemented the solution of obtain human detection by using of the command of Linux and that allows the autonomously navigation of a mobile robot until it finds its objective. Detected whole face, eye, nose Using of Face Detection Algorithm Haar Cascade Classifier in open CV open source library by the using of Python Shell. Face Detection using of LINUX command also.

In future work, To implement multiple face detection and human body detection and implement Face detection algorithm on ARM-11 using of open CV which is a open source platform library. Also implement face recognition procedure on ARM-11.

REFERENCES

1. Robinson Jimenez moreno “Robotic explorer to search people through Face Detection” Universidad autonoma de Colombia,2014, IEEE 2. Duc my vo, lixing jiang and andreas zell “Real time person detection and tracking by mobile robots using RGB-D images” international

conference on robotics and biomimetics december 5-10,2014, IEEE page no.689-694

3. xianghua fan, fuyou zhang, haixia wang, xiao lu”the system of face detection based on open CV” key laboratory for robot and intelligent technology of shandong province, shandong university of science and technology, IEEE page no. 648-651

4. Wei chen, tongfeng sun, xiaodong yang,Li wang “Full face-template and half face-template, template matching”,2009, IEEE00 5. Liying Lang,Weiwei Gu “Adaboost algorithm for real time face detection”,2009,IEEE

6. Vo Duc My,Andreas Masselli,Andreas Zell “Real time Face Detection using Of Viola-Jones algorithm” 7. “Linux BIBLE”, Wiley, Ninth edition

Figure

Fig.9. Enable Camera

References

Related documents