Use of Unmanned Aerial Vehicles for Monitoring of Air Pollution

Generated by Stationary Sources

Pavel Jirava and Michal Kuřík

Institute of System Engineering and Informatics, Fac Econ & Adm, University of Pardubice,

Studentská 95, Czech Republic

Keywords: UAV, Stationary Sources of Air Pollution, Ringelmann Scale, Data, Smoke Plume.

Abstract: This paper is focused on the use of unmanned aerial vehicles for the monitoring of air pollution by stationary

sources. A quadrocope with digital image capture was used for experiments. The data thus obtained was

processed and the Ringelmann scale method for determining the darkness of the smoke plume was applied..

The values obtained are relevant and can be used to determine the rate of air pollution by a stationary source.

The advantage of this proposed procedure are the low costs of realization and reduction of the influence of the

human factor on determining the darkness of the smoke and finally no need for direct access to the stationary

source of air pollution.

1 INTRODUCTION

The aim of the paper is to explore selected

possibilities offered by unmanned aerial vehicle in

the field of monitoring of small stationary sources of

air pollution. It is a combination of two areas,

namely the protection of the environment and

unmanned aerial vehicles (UAV). Both areas evolve

over time and it is only a matter of a relatively short

period of time before they begin to mutually

interfere and support each other. UAVs are

multifunctional devices offering a wide range of

uses, including in this area. In the experimental part,

the possibility of using the UAV for data collection

has been verified. For this purpose, several test

flights were carried out, during which videos for

subsequent analysis were made.

Based on legislation in the Czech Republic, the

method of determining the darkness of the smoke is

described. Ringelmann scale is used to evaluate data

(Uekoetter, 2005). This is applied to recorded data

using computer software. This is essentially a digital

determination of the darkness of the smoke plume of

a stationary source of air pollution.

Already today, using the UAV in the field of

emissions monitoring is often considered. This

specialized field is developing very fast. For

example, the European Maritime Safety Agency

deals with "Remote Pilot Aircraft System (RPAS)

services in the maritime environment. RPAS has an

on-board gas analyzer draws samples of air and

monitors of SOx, NOx and CO

levels. However,

this is a relatively expensive technology (EMSA,

2017). Another example is UAV based smoke plume

detection system controlled via the short message

service through the GSM network, proposed system

consist of two parts: flying hardware and embedded

kit for finding smoke plume. This system is

dependent on the operation of GPS networks

(Ramanatha, 2016). The aim of this paper is, among

other things mentioned above, to show a very simple

and inexpensive way to use the UAV for the

detection and evaluation of air pollution. A micro

UAV with standard camera and optics was used. We

used the free software to process the data. The total

cost of experiments (including UAV, insurance,

pilot registration fees, flight fee, free SW) did not

exceed EUR 1000. The economic aspect and the cost

of UAV use is a very important factor in its wider

expansion.

2 UNMANNED AERIAL

VEHICLES

An UAV, sometimes referred to as a drone, is

generally perceived as an unmanned aircraft. It can

be controlled remotely at the pilot's eye or outside

Jirava P. and KuÅ

ZÃ k M.

Use of Unmanned Aerial Vehicles for Monitoring of Air Pollution Generated by Stationary Sources.

DOI: 10.5220/0006510801500155

In Proceedings of the International Conference on Computer-Human Interaction Research and Applications (CHIRA 2017), pages 150-155

ISBN: 978-989-758-267-7

the pilot's reach. In some cases, the distance between

the pilot and the drones may be several thousand

kilometers, as is the case for military UAV. Another

option is the autonomous movement on a previously

planned route. Nowadays computer technology is at

such a level that it allows a completely independent

movement, which is, however, limited by the

legislative conditions of operation. (Karas, 2016)

One of the biggest advantages over piloted

machines is the lower cost of operation and

acquisition of the machine. It cannot be said that this

applies to all UAVs. Some are more expensive than

ordinary planes. But with commercially available

drones, it's a fact. If you need to scan a small area,

deploying these resources is much more economical

than conventional methods (Saadatseresht, 2015).

They are also a suitable solution if you need to

scan or explore only a small area or area where the

classic procedure cannot be used. Their further

advantage is therefore the reach and permeability of

the unfavorable terrain. If allowed constructively,

they can start or land perpendicular, which can be

used in places where such operations are not

necessary for ordinary aircraft. (Rehak, 2012)

Another great advantage is maneuverability and

safety. They can be used in natural disasters and

other situations where rescuers, firefighters, etc.

would otherwise endanger life. (Karas, 2016)

This contribution is also based on some articles

on air quality monitoring. This is mainly research on

the use of quadcopter for monitoring air pollution

(Zhaoyan et al., 2017), and methodology for smoke

detection and monitoring (Gómez-Rodríguez, 2002;

Gómez-Rodríguez, 2003)

The disadvantage of drones can be flight time,

low load capacity. This does not have to be the case

for special army UAVs, which are able to use the

combustion engines to travel to one tank for

thousands of kilometers and can last for several tens

of hours in the air. Commercially available drones

are not capable of so long flights. There are dozens

or hundreds of possibilities and ways to use the

drones. Each field or profession brings another

(d'Oleire-Oltmanns et al., 2012).

The UAV can be seen as a dynamic system that

can be modeled and analyzed using various tools

(Ibl, 2016). No unified international legislation

exists for the operation of UAV. Therefore, it is

necessary to read the laws for the airspace in which

the flight is to be conducted before flight (Cracknell,

2017).

From the historical point of view, there has been

and still remains the greatest benefit of an UAV for

military and military purposes, whether used for

protection, search and rescue of missing persons,

monitoring, communications, exploration or as a

weapon.

In recent years, UAV have become available to

civilians as well. Due to their variability and

features, the most common purpose of the job is to

simplify work and cost effectiveness. They also

provide the ability to represent a person in

potentially dangerous situations, thereby

contributing to increased safety. A wide range of

applications also stems from the use of various

special sensors.

All drones usage can be broken down according

to the following division:

 Aerial photography

 Aerial video

 Aerial monitoring

 Space and terrain mapping

 Special applications in conjunction with

special sensors

 Transport and logistics

 Entertainment (Rodríguez-Canosa et al,

2012.; Towler et al., 2012; Karas, 2016)

3 AIR POLLUTION

In each developed country, air pollution is regulated

by the applicable law. This paper is based on the

Czech Air Protection Act 201/2012 Sb.

The sources of air pollution by origin are divided

into natural and anthropogenic. We mean natural

sources that are not of human origin. This may be,

for example, volcanic activity or dust storms. The

other group is anthropogenic sources related to

human activity, namely industrial and agricultural

production, electricity and heat generation, transport

and waste disposal. We have focused here on

anthropogenic stationary sources.

Depending on location, we can divide our

sources into ground, elevated and elevated. We also

divide the sources of pollution according to their

layout, point, line and bulk order. The point source

can be, for example, a chimney.

A stationary source is, according to the Act on

Air Protection, a stationary technical unit which is

further indivisible and in which the fuels are

oxidised in order to use the released heat.

The list of polluting sources and their

categorization together with the requirements for

individual categories can be found in Appendix 2 to

the Act on Air Protection (The Act No. 201/2012,

2012). We place stationary sources in the following

categories:

 Energetics - Combustion of fuels

 Heat treatment of waste, waste

management and sewage

 Energy - others

 Production and processing of metals and

plastics

 Processing of minerals

 Chemical industry

 Food, woodworking and other industries

 Livestock farming

 Use of organic solvents

 Petrol handling

In order to protect human health and the

environment, limitations on the quantity of airborne

substances discharged by stationary sources are

established. These are emission limits, immission

limits and emission ceilings (Jirava et al , 2010) .

The degree of air pollution by smoke from a

stationary source can be evaluated by means of a

Ringelmann scale. This is a visual method where the

observer visually assesses the darkness of the smoke

plume at the outlet of the chimney. Ringelmann

scale, which consists of six rectangular arrays (see

Figure 1), each representing a given degree of

darkness of smoke.

Figure 1: Ringelmann's scale with RGB values.

Grade 0 corresponds to 0% of black on a white

background with definition. Reflectivity 80%.

Grade 1 corresponds to 20% of black on a white

background.

Grade 2 corresponds to 40% of black on a white

background.

Grade 3 corresponds to 60% of black on a white

background.

Grade 4 corresponds to 80% of the black on a white

background.

Grade 5 corresponds to 100% black on a white

background.

4 EXPERIMENTS

The experiments and calculations performed were as

follows. Firstly, UAV flights were conducted and

data capture was performed using a digital camera.

With the UAV, one stationary source of pollution

was repeatedly scanned under different

meteorological conditions. The obtained data was

preprocessed and digitally processed (Panus and

Simonova, 2005). The outcome was the

determination of smoke darkness according to the

Ringelmann scale.

4.1 Flights and Data Collection

For the experiments, conventional commercially

available UAVs were used (Parrot AR Drone, 2017).

This is a quadrocope controlled by smartphone or

tablet (Benson, 2017). Communication of the control

device (mobile phone) with the machine takes place

via Wi-Fi, which limits the operating range to 50

meters. The UAV has the following features:

720p 30fps HD camera (4)

Photo format: JPEG

Connection: Wi-Fi

With internal frame: 380 g

With external frame: 420 g

Processor: ARM Cortex A8 1 GHz

RAM: DDR2 1 GB at 200 MHz

USB: High-speed USB 2.0 for extensions

Gyroscope: 3 axles, accuracy of 2,000°/second

Accelerometer: 3 axles, accuracy of +/- 50 mg

Magnetometer: 3 axles, accuracy of 6°

The take-off of an UAV was carried out on a

private property with the permission of the owner

and did not leave it for the duration of the flight.

Also, safe horizontal distance from surrounding

buildings and persons not directly involved in the

flight was maintained. As a stationary source of air

pollution a combustion boiler with a manual solid

fuel supply was used, where the chimney mouth was

on the same land from which the UAV was started.

The Figure 2 below shows a graph with the

important flight data. The y axis on the left refers to

the altitude of the flight. The second axis on the left

shows the values related to the flight velocity in

meters per second. The top speed that the UAV

reached was 3.7 m / s, 13.3 km / h. The vertical axis

on the right shows the percentage of battery charge.

At the beginning of the flight, the battery was

charged almost 80%. After five minutes, this value

dropped by half. The time is captured on the x axis.

According to the chart, the flight lasted for just over

five minutes

Figure 2: Flight data.

4.2 Data Processing

During the flight, a digital record of smoke plume

was taken (Aber et al., 2010). Appropriate images

were separated from it, as shown in Figure 3 (with

free SW Avidemux 2.6). Subsequently, only the

smoke plumes were separated (using free graphical

SW GIMP v. 2.8.22). In order to determine the

average values of the RGB components of smoke

plume, it is necessary to place the smoke plume on a

transparent background that does not affect the mean

values of the individual component (realized with

free graphical SW GIMP v. 2.8.22).

Figure 3: Digital image of smoke plume.

In the following figure 4, the smoke plume is

excluded from five different pictures taken from one

video. The interval between the slides is five

seconds. For each of the smoke plume, the RGB

values are determined by the procedure described

above in this paper. In the table 1 are summarized

outputs from one flight. In the first column is

computed RGB value. In the second column is the

closest grayscale from left (based on the Ringelmann

scale). In the third column is the closest grayscale

from right (based on the Ringelmann scale). In the

last column is computed value for pictures a,b,c,d,e.

The last row of the table is the total result.

Figure 4: Preprocessed smoke plumes.

Table 1: Computed values.

RGB

values

Ringel. L

Ringel.

Final

value

0 1

210,8

202,2

205,4

255

209

210

212

0 1

229,0

221,2

212,9

255

209

210

212

1 2

193,2

187,3

197

209

210

212

167

169

172

0 1

239,8

232,6

228,8

255

209

210

212

1 2

201,3

193,1

192,9

209

210

212

167

169

172

Final

Ringelmann

value

4.3 Discussion

The proposed procedure (see Fig. 5) appears to be

applicable according to the obtained results.

Digitization and computer processing minimize the

impact of human factor on final results. These could,

of course, be supported by other methods such as

Gas chromatography, Dynamic Optochemistry,

Gravimetric analysis, or Spectroscopy. These

methods, however, require more complex and

expensive devices (Wolf and Witt, 2000; SenseFly,

2011). Meteorological conditions also affect

measurement. For example, strong wind gusts can

not only affect the smoke plume but also endanger

the UAV.

We believe that the measurements must be

carried out by an authorized person. The question

remains whether it is necessary for this person to be

the pilot of an UAV or only to be present when

carrying out the measurements. An operator could

be another person who has all the permits needed to

make a flight (licence). The officers of the

competent authorities would not have to have the

necessary pilot licenses at all. They would only

perform their own measurements.

Figure 5: Experiments - block diagram.

5 CONCLUSIONS

In this work we dealt with the problem of

determining the darkness of smoke plume from a

stationary source of air pollution using the UAV.

The smoke darkness can be determined using the

Ringelmann scale. This process is usually realized

manually and the measurement is thus largely

affected by human error. The proposed procedure

should eliminate this human error. The experiment

was as follows (figure 5 - left side block diagram).

First we collected data using UAV. The obtained

digital records were pre-processed and modified for

the RGB component decomposition. These were

then compared with the Ringelmann scale using the

software and finally the resulting value was

assigned. In addition, this procedure does not require

direct access to a source of pollution because it uses

the UAV.

The whole process can be done by an embedded

system on board of the UAV (see figure 5 – right

side diagram). The results may be sent via GPS to

the ground stations. But that is only the aim of future

research.

A disadvantage may be the varying legal

regulation of UAV issues in individual countries.

The application of the method would thus have to be

considered from the point of view of the country's

legislation.

ACKNOWLEDGEMENTS

This work was supported by the projects No. SGS

2017_SG19 of the Ministry of Education, Youth and

Sports of CR with title “Models Synthesis and

Analysis for Implementation Support of Smart Cities

and Regions Concept” at the Faculty of Economics

and Administration, University of Pardubice.

REFERENCES

Aber, J., Marzolff, I., Ries, J.B. 2010. Small Format

Aerial Photography: Principles, Techniques and

Geoscience Applications; Elsevier: Amsterdam,

The Netherlands.

Benson, C. 2017. How to make a drone. RobotShop.

[Online] Available from: http://www.robotshop

.com/blog/en/make-uav-lesson-2-platform-14448.

Cracknell, A.P. 2017. UAVs: regulations and law

enforcement. International Journal of Remote

Sensing.Vol. 38. Issue 8-10. Pp. 3054-3067.

d'Oleire-Oltmanns, S., Marzolff, I., Peter, K.D., Ries, J.B.

2012. Unmanned Aerial Vehicle (UAV) for

Monitoring Soil Erosion in Morocco. Remote Sens.

Vol. 4, pp. 3390-3416.

EMSA. 2017. European Maritime Safety Agency.

[Online] Available from: http://www.emsa.europa.eu/

fc-default-view/tagged/82-rpas.html.

Gómez-Rodríguez, F., Pascual-Peña, Arrue, B., Ollero, A.

2002 Smoke detection using image processing.

[Online] Available from: https://www.researchgate.net

/publication/292461320_Smoke_detection_using_ima

ge_processing.

Gómez-Rodríguez, F., Arrue, B., Ollero, A. 2003. Smoke

Monitoring and Measurement Using Image

Processing. Application to Forest Fires. Proceedings

of SPIE - The International Society for Optical

Engineering.5094.

Ibl, M. 2016. A User-friendly Tool for Modeling and

Simulation of Discrete Event Dynamic Systems. Divai

2016: 11th International Scientific Conference on

Distance Learning in Applied Informatics. Pp. 137-

147.

Jirava, P., Krupka, J., Kasparova, M. 2010. Application of

Rough Sets Theory in Air Quality Assessment In:

Rough Set And Knowledge Technology (RSKT)

Lecture Notes in Artificial Intelligence. Beijing, China.

Karas, J., Tichý, T. 2016. Drony. Brno: Computer Press.

Panus, J., Simonova, S., 2005. Pre-computation of

regional data for optimization analysis. In: Eurocon

2005: The International Conference on Computer as a

Tool, Vol 1 and 2, Belgrade, SERBIA.

Parrot AR.Drone 2.0. 2017. [Online]. Available from:

https://www.parrot.com/uk/drones/ parrot-ardrone-20-

elite-edition.

Ramanatha, P., Nelakuditi, U. R., Ravishankar, S.,

Ranganathan, V. 2016. UAV based smoke plume

detection system controlled via the short message

service through the GSM network. In: 2016

International Conference on Inventive Computation

Technologies (ICICT), Coimbatore.

Rehák, M. 2012. Využití bezpilotních prostředků ve

fotogrammetrii. Theses. Prague.CTU.

Rodríguez-Canosa, G. R., Thomas, S., del Cerro, J.,

Barrientos, A., MacDonald, B. A. 2012. Real-Time

Method to Detect and Track Moving Objects

(DATMO) from Unmanned Aerial Vehicles (UAVs)

Using a Single Camera. Remote Sens. Vol. 4, pp.

1090-1111.

Saadatseresht, M., Hashempour, A. H., Hasanlou, M.

2015. UAV Photogrammetry: A Practical Solution for

Challenging Mapping Projects. In: International

Conference on Sensors & Models In Remote Sensing

& Photogrammetry. Iran.

SenseFly: Autonomous flying sensors. 2011. [Online].

Curych, Available from: http://www.sensefly.com/

The Act No. 201/2012 Coll. on Ambient Air Protection

(the Air Protection Act). 2012. [Online]. Available

from: http://www.mzp.cz/www/platnalegislativa.nsf/

d79c09c54250df0dc1256e8900296e32/9f4906381b38f

7f6c1257a94002ec4a0?OpenDocument.

Towler, J., Krawiec, B., Kochersberger, K. 2012.

Radiation Mapping in Post-Disaster Environments.

Uekoetter, F. 2005. The Strange Career of the Ringelmann

Smoke Chart. Environmental Monitoring and

Assessment. Volume 106, Issue 1–3, pp 11–26.

Wolf, P. R., de Witt, B. A. 2000. Elements of

Photogrammetry: With Applications in GIS, 3rd ed;

McGraw-Hill: Boston, MA, USA.

Zhaoyan S., Zhijian H., Shuai L., Qixin W., and Zili S.

2017. A Multi-Quadcopter Cooperative Cyber-

Physical System for Timely Air Pollution

Localization. ACM Trans. Embed. Comput. Syst.

Vol.16, 3, Article 70, 23 pages.