Abstract— Vision based fire surveillance is a useful technique.

With the increased number of surveillance cameras, a vision

based fire detection capability can be incorporated in existing

surveillance systems at relatively low cost. Vision based fire

detection offers advantages over the traditional methods. In this

paper we provide an ideal solution for surveillance that involves

the assortment of necessary sensors and devices using pertinent

tools.In our project, we are aimed at developing a low- cost, real-

time video surveillance kit which will be proficient of providing

real-time video streaming of the area which we want to observe.

During rescue mission it can lessen the human effort in

menacingcircumstances. A Bluetooth module is used for the

communication between the controller and the android

application. We will program the microcontroller in such a way

that say whenever the smoke sensor perceives smoke in air and if

the temperature is beyond a certain threshold level, temperature

status, smoke level etc. will be immediately sent and stored into

mobile via Bluetooth. Thereby we can detect the safest path thus

avoiding loss of life.

Index Terms—Live Video streaming, (HC-05)Bluetooth

module, Arduino,Android Mobile application.

I. INTRODUCTION

According to the studies conducted, it is observed that inthe

event of any natural calamitythe first 48 hours are verycritical

in saving human life.Fire rescue is one of the most important

public safety activities.Development of automated and smart

video surveillance systems has been the core trend. This

project presents a simple and effectual method for detecting

the fire crisis automatically in the monitoring area via live

video transmission.By observing and utilizing features of fire

event, a quick and accurate detection process is developed for

early fire warning purpose thus to reduce the mortality rate

and loss of property caused by fire accidents. The proposed

algorithm not only achieves real-time requirements and has

better performance (more robust and correct) than the existing

surveillance systems, but also cost effective. Our fire detection

system is a reliable system for surveillance and recognition of

spontaneous fires inside commercial buildings or homes.Real

time applications, such as habitat surveillance, environmental

and structural observance,start to work in practical since they

have attracted a lot of attention in recent years. The proposed

system works in case if predefined temperature limits are

exceeded and the smoke is detected in the smoke

sensor.Alarms and the system status are displayed on the

monitor of the android device. By observing that the demand

of using the digital surveillance system gets increasing, an

intelligent, integrated, and automatic system should be

developed to overcome the drawback in the existing systems.

People and building are getting more crowded in urban area,

thus it makes fire accident more harmful than sparse area. The

main objective of this work is to develop a fully automatic

warning and surveillance system for potential danger in the

monitoring area via video transmission.In this paper, the

sections to be discussed are arranged sequentially. Some

similar and existing technology discussed in some paper are

reviewed in Section II. The proposed system architecture &

features are discussed in Section III. The system

implementation against the features are discussed in Section

IV. Finally the future scope& conclusion is remarked in

Section V.

II. RELATED WORKS

There are large number of papers related to fire detection. One

such paper is about fire detection in computer vision literature.

A probabilistic approach is used for vision based fire

detection. It is one of the method for identifying fire in videos.

It has a background control for CCTV surveillance.

Additionally an automatic video classifier is used. The

features for detection include color, area, size, boundary

roughness and skewness that describe the fired region. It also

proposes the use of variance because of the randomness in fire

surfaces. Bayes classifier is used to achieve this practically.

All these features are considered as a major discriminants

because of the flickering and random characteristics of blaze.

A similar paper for detecting fire uses flame detectors. With

the advance in IT, optic temperature sensor is widely used to

detect fire. The cost is too high, the basic working is to judge

the color and motion of fire using video processing technique.

They cannot detect the temperature, gas and other sources of

fire. The video processing technique is based on thermal

image. For thermal image, they use omnidirectional thermal

sensors which uses eight thermal infrared sensors for 360

degree motion detection and tracking. The thermal camera

gives only the raw thermal data. The fire detection device

comprises of 3 important things-the first one is thermal

camera, the second one is conic mirror for omnidirectional

view and the last one is pan-tilt module for wide and narrow

field view. The mirror system consists of three mirrors. The

first one is conic mirror which is easy to build. The next is

spherical mirror which is also easy to build. The last one is

hyperboloid mirror which is difficult to machine. The pan-tilt

module provides two modes (i) Wide range search mode by

SURVEILLANCE AND APPREHENSION OF BLAZE WITH MOBILE

APPLICATION INTELLIGENCE

Mrs. K.Lavanya, Assistant Professor, A.Aishwarya, D.Kavya, M.Shwetha, S.Srinidhi

Velammal Engineering College.

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 14, Number 6, 2019 (Special Issue) © Research India Publications. http://www.ripublication.com

Page 282 of 285

positioning the thermal camera upward to conic camera. (ii)

Narrow range accurate search mode. Finally method detects

sources of fire detection only within 1.5m distance range. In

the prevailing methods of fire detecting systems an approach

is based on video stabilization which is completely reliable on

SURF which involves two main steps: interest point detection

and interest point description which are implemented by

systolic two layer array Architecture designed on an FPGA

platform. The other approach involves rescue missions of

victims finding, positioning and identification. While the state

of the victim is not completely achievable autonomously and it

needs a continuous monitoring end that is challenging which

ultimately consumes time that is unfit for a fire accident

situation. Also the use of audio sensor is not advisable as it

may encounter false-positive victim detection. With reference

to the related work, it builds an end to end connectivity over

wireless networks between two heterogeneous systems for live

video streaming. In this system, the wireless networks are

varying in time and have inferior performance when compared

to wired network. Moreover, it requires short transition time

during hand off and thus the video quality gets deteriorated.

III. DESIGN AND IMPLEMENTATION

A. ATMEGA 2560 ATMEGA 2560 has the following pins: 54input and output

pins,16 analog inputs,4 UART’s which are hardware serial

ports, ATMEGA 2560 is a arduino mega 2560 microcontroller

based board. 16 MHz crystal oscillator, ICSP (Incircuit serial

programming) header, power jack, USB connection, a reset

button. The arduino software (IDE) is used to program the

ATMEGA microcontroller. The Integrated Development

Environment(IDE) runs both online and offline on the boards

and allows to write and upload programs on board. The

microcontroller is coded using C language

B. Firesensor The fire sensor that we use is silicon PIN photodiode which is

BPV10. It is very fast and sensitive to blaze with a plastic

package. The device is sensitive to visible and infrared

radiation. It has many features like fast response time, high

sensitivity to flame, high bandwidth of 250 MHz at 12 V.

Application : Used as wide band detector for industrial

electronics, interrupters and control circuits.

C. Smoke sensor MQ-2 is a semiconductor sensor for detecting combustible

gases.MQ-2 gas sensor has a sensitive material of 𝑆𝑛𝑂2

which has lower conductivity in clean air. When the gas

concentration increases in the fired location, the

sensors’sconductivity also raises and this increase in

conductivity corresponds to the gas concentration output

signal. It has high sensitivity to propane, hydrogen, LPG,

methane and other combustible gas. The features are long life

with low cost and has a simple drive circuit.

Application: Used as portable gas detector, Industrial

combustible gas detector, domestic gas leakage detector.

D. Temperature sensor The temperature sensor is of two types: contact and non-

contact. The contact sensors measure the thermal radiation

from the object and then measure the temperature. The non-

contact sensors measure the temperature directly from a

distance in a hazardous environment. The temperature that is

read in the environment is the voltage that is read across the

diode.

E. Relay Relays are small electromagnetic switch that are operated by

small electric currents. It acts as both switches and amplifiers.

The relays are made of a coil which acts as a temporary

magnet when electricity flows through it. Generally, the

various sensors that we use are sensitive electronic equipment

that often produces only small current. This current is not

enough to drive big apparatus. So the relay helps small

currents to activate the larger objects by converting small

currents into larger currents thereby acting as amplifiers here.

They are generally single pin or dual pin, 3-pin, 4-pin, 5-pin,

6-pin, etc.

Applications: Used in telephone switching and as amplifiers

till the arrival of transistors.

F. Bluetooth HC-05 module is a bluetooth serial port protocol (SPP)

module which is used to communicate with two arduino

boards. The arduinos communicate via serial communication.

HC-06 can be only set as slave while HC-05 can be set as both

master and slave between two arduino boards. This module is

fully qualified bluetoothV2.0+EDR(Enhanced data rate) with

2.4 GHz radio baseband and transceiver. It uses CSR

bluecore-04 with CMOS technology and adaptive frequency

hopping(AFH). The slave mode is the default factory setting.

In slave mode it can accept only connections and cannot

initiate connections while in master mode it can initiate

connections. HC-05 has 6 pins.

Enable : When it is connected to 3.3V the device is on and

communication takes place.

𝑉𝑐𝑐 : 3.3V-5V.

Gnd: Ground pin.

TXD and RXD: UART interface for communication.

Status: Acts as status indicator.

Button: Switch the module into AT command mode.

G. Power supply The ac voltage which is 220 V rms is connected to

transformer which steps down the voltage to desired DC

output level. The diode rectifier provides full wave rectified

voltage that is initially filtered by a simple capacitor to

produce a DC voltage. The resulting DC voltage has some

ripples and the ripples are removed by the regulator circuit

andthe DC value remains the same even if the input DC

voltage varies or the load connected to the output DC voltage

changes. The voltage regulation is obtained by one of the

popular voltage regulator IC circuits.

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 14, Number 6, 2019 (Special Issue) © Research India Publications. http://www.ripublication.com

Page 283 of 285

IV. PROPOSED WORK

In the proposed system, Arduino Mega 2560 is used as the

Control unit through which various sensors are controlled.

The goal is to develop a real time video surveillance.

Arduino kit which will be fixed at various locations in

commercial buildings. The kit developed is interfaced with

temperature sensor, flame sensor, smoke sensor, camera,

Bluetooth module and an LCD display. These are usually

cheap and stationary solution for fire detection. The sensors

are interfaced with the analog input pins of the control unit.

The data collected from the sensors are viewed through the

LCD display. The display information includes temperature

rise, presence of smoke and flame of the surrounding

environment. If the temperature rises above the threshold

value of 45 degree Celsius or the flame is detected through

the IR sensor, then an alert is given via the DF Mini player

to the people in the fired building. The same information is

obtained in the developed Android Application to the

security personnel who will be given access. The smart

microcontroller unit named as Arduino Mega 2560 can be

programmed with the Arduino IDE softwareto upload new

code to it without the use of an external hardware

programmer. The IP web camera is used to enhance the

situation awareness of the damaged environment through

live video streaming. The harmful and poisonous gas is

detected from the gas sensor so to protect the victims from

death. A HC-05 Bluetooth module is used for the

communication between the Arduino controller and the

android application.Communication is using the original

UART protocol (reference, C header files). Additionally, the

DF Player Mini SKU gives necessary commands about the

safest path to the victims so that the victims can escape via

non-fired regions. This protects them from various fire

disasters as well as from other destructions. Finally, the

whole system is designed to obtain the required quality

video streaming and the necessary parameters are viewed.

ALGORITHM

1. Start.

2. LCD displays “FIRE SAFETY SYSTEM” and

temperature of each floor.

3. Micro-controller reads the input from temperature

sensor, flame sensor and gas sensor.

4. If poisonous gas leaks or detects smoke or detects fire,

i. HC-05 Bluetooth module turns on.

I. Mobile application indicates the fire

prone floor with the corresponding

exit number.

ii. Speaker turns on.

iii. LCD displays the fired floor.

Else microcontroller continuously reads input from the

sensors.

5. Stop.

Fig., Block diagram

V. RESULTS AND DISCUSSION

The developed project was tested with inputs to different

sensors and was found to work accordingly. The response for

the input to the sensors were noticed on the LCD display,

mobile application and DLF Mini player. The LCD Display

displays the floor where fire has occcured and it is also

specified in the mobile application using a flame image. The

DF Mini player plays the voice recorded in the SD card

according to the current situation describing the fire prone and

the safe floors, by collecting information from the sensors.

The above figure shows the output for fire detected on the first

floor by the fire sensor indicating the first exit. The DF Mini

player plays the audio indicating that the first floor has caught

TRANSFORMER RECTIFIER FILTER

IC REGULATOR LOAD

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 14, Number 6, 2019 (Special Issue) © Research India Publications. http://www.ripublication.com

Page 284 of 285

fire whereas second and third floors are safe.The second and

the third exits of each floor corresponds to the gas and

temperature sensors respectively. When any of thepoisonous

gases such as LPG, Butane, Propane, Methane, Alcohol,

Hydrogen and Combustible steam is detected, then it indicates

fire in the second exit of the corresponding floor in the mobile

application along with the DF Mini player alert. When the

room temperature exceeds the preset threshold temperature, it

indicates fire and flame symbol is indicated in the third exit of

the corresponding floor along with the DF Mini player alert.

VI CONCLUSION

Thus this project brings an intelligence to view and

communicate a situation during a blast using microcontrollers

and mobile application. Thus the camera linked with the

device provides a live video surfing of the blast environment

that ultimately guides the control unit.In addition to this, the

system is supported by an automated alarm triggering. The

success of any innovation relies on the attribute against real

life problems and its viability. Fire is one of the dangerous

events which can result in great losses if it is not controlled on

time. This necessitates the importance of developing early fire

detection systems. Therefore, in this research article, we

propose a cost-effective fire detection.Additional

improvement in the design and functionality of the kit could

make this project more successful.

VII FUTURE WORK

It is vital to note that noise generated by the sensors are an

important factor and cannot be ignored and has certain impact

on the control systems. At future works many filters, both

analog and digital can be used to filter sensor noise to provide

a better control. Complexity of the system components,

frustrated environmental factor and unpredictable human

factorsmay affect the performance of the system seriously.

This is verified with the help of fire proof Wi-Fi camera which

has 360 degree vision, full HD picture quality, full color in

low light, two way audio talk back features. The camera is

turned on by the controller as soon as fire is detected. The

application can be further evolved such that it automatically

notifies the nearby fire station and the entitled person

regarding the event of fire accident and update them the

ongoing situation.The obtained particulars regarding the zone

of fire can be employed for improvisation of the mobile

application, to acquire the safest path inside the building. This

could be utilized by the firemen in an effective manner for

quick and safe entry into the building been in flames. This

could potentially reduce the loss of lives and property.

REFERENCES

[1] Khan Muhammadi,Jamil Ahmadi, IrfanMehmood, SeungminRho3 and

Sung WookBaiki, “Convolutional Neural Networks based Fire Detection in

Surveillance Videos.,”IEEE Access,vol.6.06 March 2018, DOI:

10.1109/ACCESS.2018.28128355.

[2] Paulo ViniciusKoerich Borges and EbroulIzquierdo, “A Probabilistic

approach for vision based fire detection in videos,” IEEE Transactions on Circuits and Systems for video technology., vol. 13, no. 1, pp. 11-23, Jan.

1995, DOI: 10.1109/TCSVT.2010.2045813.

[3] TahiyahNouShene, K. Sridharan and N. Sudha, “Real-Time SURF-Based

Video Stabilization System for an FPGA-Driven Mobile Robot,” IEEE

Transactions on Industrial Electronics., vol. 11, no. 2, pp. 115–117, Feb.

1986, DOI: 10.1109/TIE.2016.2551684.

[4] Sangchun Han, Hyunchuljoo, Dongju Lee and Hwangjun Song,

“An End-to-End Virtual Path Construction System for Stable Live

Video Streaming Over Heterogeneous Wireless Networks.,” IEEE Journal On Selected Areas In Communications, vol. 29, no. 5, May

2011, DOI: 10.1109/JSAC.2011.110513.

[5] Pasquale Foggia, AlessiaSaggese and Mario Vento, “ Real-time

Fire Detection for Video Surveillance Applications using a

Combination of Experts based on Color, Shape and Motion.,” IEEE Transactions on Circuits and Systems for Video Technology,vol.

25 , no. 9 , Sept. 2015, DOI:10.1109/TCSVT.2015.2392531

[6] ByoungChulKo, Kwang-Ho Cheong, Jae-Yeal Nam, “Fire

detection based on vision sensor and support vector machines.,” Fire Safety J., vol. 44, no. 3, pp. 322–329,Apr. 2009.

[7] C. L. Lai, J. C. Yang, and Y. H. Chen, “A real time video

processing based surveillance system for early fire and flood

detection.,” in Proc. IEEE Instrum . Meas. Technol. Conf., Warsaw,

Poland, May 2007, pp. 1–6. [8] P. Barmpoutis, K. Dimitropoulos and N. Grammalidis, “Real time

video fire detection using spatio-temporal consistency energy., ” IEEE International Conference on Advanced Video and Signal-Based Surveillance, Krakow, Poland, 2013.

[9] Jung-Hoon Hwang, Sewoong Jun, Seung-Hun KimDonghoon Cha

and Kaehoon Jeon,”Novel fire detection device for robotic fire

fighting,” ICCAS 2010,30 Oct 2010,

DOI: 10.1109/ICCAS.2010.5669964.

[10] Y. H. Habiboglu, O. Günay, and A. E. Çetin, "Covariance matrix-

based fire and flame detection method in video.," Machine Vision and Applications, vol. 23, pp. 1103-1113, 2012.

[11] M. Mueller, P. Karasev, I. Kolesov, and A. Tannenbaum, "Optical

flow estimation for flame detection in videos.," IEEE Transactions on Image Processing, vol. 22, pp. 2786-2797, 2013.

[12] R. Di Lascio, A. Greco, A. Saggese, and M. Vento, "Improving

fire detection reliability by a combination of videoanalytics.," in International Conference Image Analysis and Recognition, 2014, pp.

477-484.

[13] A. Ullah, J. Ahmad, K. Muhammad, M. Sajjad, and S. W. Baik,

"Action Recognition in Video Sequences using Deep Bi-directional

LSTM with CNN Features.," IEEE Access, vol. PP, pp. 1-1, 2017.

[14] A. Ullah, J. Ahmad, Khan Muhammad, Irfan Mehmood, Mi

Young Lee, Jun Ryeol Park, Sung WookBaik, "Action Recognition in

Movie Scenes using Deep Features of Keyframes.," Journal of Korean Institute of Next Generation Computing, vol. 13, pp. 7-14, 2017.

[15] J. Yang, B. Jiang, B. Li, K. Tian, and Z. Lv, "A fast image

retrieval method designed for network big data.," IEEE Transactions on Industrial Informatics, 2017.

[16] D. Y. Chino, L. P. Avalhais, J. F. Rodrigues, and A. J. Traina,

"BoWFire: detection of fire in still images by integrating pixel color

and texture analysis.," 28th SIBGRAPI Conference on Graphics, Patterns and Images, 2015, pp. 95-102, 2015

[17] S. Verstockt, T. Beji, P. De Potter, S. Van Hoecke, B. Sette, B.

Merci, et al., "Video driven fire spread forecasting (f) using multi-

modal LWIR and visual flame and smoke data.," Pattern Recognition Letters, vol. 34, pp. 62-69, 2013.

[18] B. C. Ko, S. J. Ham, and J. Y. Nam, "Modeling and formalization

of fuzzy finite automata for detection of irregular fire flames.," IEEE Transactions on Circuits and Systems for Video Technology, vol. 21,

pp. 1903-1912, 2011.

[19] S. Rudz, K. Chetehouna, A. Hafiane, H. Laurent, and O. Séro-

Guillaume, "Investigation of a novel image segmentation method

dedicated to forest fire applications.,”Measurement Science and Technology, vol. 24, p. 075403,June 2013, DOI:

10.1088/0957233/24/7/075403.

International Journal of Applied Engineering Research ISSN 0973-4562 Volume 14, Number 6, 2019 (Special Issue) © Research India Publications. http://www.ripublication.com

Page 285 of 285