System for Monitoring the Technical State of Heating Networks Based on UAVs

  • Artur ZaporozhetsEmail author
  • Svitlana Kovtun
  • Oleh Dekusha
Conference paper
Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 1080)


The article presents the causes of defects in pipelines of the centralized heat supply. The possibilities of thermal aerial photography for detecting different types of defects on pipelines in a functioning state are explored. The characteristics and capabilities of the proposed set of devices for monitoring thermal losses in pipelines based on quadrocopters are considered. A method for monitoring the technical condition of pipelines using UAVs is presented. A method for processing thermal images for highlighting anomalous areas is presented. The created hardware-software complex for monitoring the state of trunk pipelines of heat networks based on the UAV is considered. Experiments on the use of UAVs for monitoring heating networks have been conducted. The obtained experimental results, confirming the possibility of differences in the technical condition of pipelines.


Heating network Main pipelines Thermal image Aerial photography Monitoring Image processing Quadcopter 

1 Introduction

Ensuring comfortable living and working conditions for people in the settlements of Ukraine during the heating season is carried out by heat supply systems, where purified water is used as the heat carrier. Supply of heat carrier from boiler houses to consumers is carried out by heating networks, the total length of which in Ukraine in two-pipe equivalent is 35700 km. Due to the fact that the total deterioration of heating networks is about 70%, the heat loss during transportation by networks reaches more than 15%, and the loss of water is more than 30%, which significantly exceeds the standard level [1]. In this situation, the state and consumers incur large losses, spending significant funds on the additional acquisition of energy carriers and the maintenance of poor-quality and unreliable heating networks [2]. In addition, the old heating systems require frequent repairs and partial replacement of damaged areas, which leads to constant disconnections of consumers from the heat supply [3]. All this entails additional financial costs, which within the state amount to millions of dollars, and their size is constantly growing.

Monitoring the technical condition of the heating network makes it possible to determine the presence of damage to pipelines and their insulation. Basically, the consequences of damage of the heating network during operation is the destruction of metal pipelines as a result of internal and external corrosion, as well as deterioration of the characteristics of thermal insulation and waterproofing [4, 5].

The main defects of metal pipelines are: crack, rupture of the metal, thinning of the wall due to mechanical stress, the effects of corrosion or delamination. Defects such as metal rupture also include fistulas occurring in welded joints of pipelines. Defects of thermal insulation is moisture insulation and partial or complete destruction of it. Defects of metal pipelines and thermal insulation can be local or extended.

In the urban landscape, leaks from heating networks are reflected in the thermal field of the earth’s surface even in the presence of a solid road surface [6, 7].

After analyzing the features of the functioning of underground heating networks, we can distinguish two types of leaks. The first is the accidental leakage that occurs during the heat conductor ruptures and are accompanied by the outpouring of large amounts of mains water. Water fills the channel, goes out through the cracks, erodes the soil, forming cavities in it, sometimes goes to the surface. Of course, such leaks have a significant impact on the geological environment, being the main cause of flooding of urban areas. In addition, they lead to disruption of heat supply and cause significant damage to the urban economy, and sometimes cause accidents with tragic consequences. There are cases when people and equipment fell into the voids with hot water (Fig. 1).
Fig. 1.

Breakthrough of the main pipeline (November 2017, Kyiv)

Therefore, operational services are trying to quickly localize the emergency sections of the network and eliminate leaks [8]. The occurrence of an emergency leak is recorded by a sharp drop in pressure in the supply line, and therefore is determined quite quickly [9, 10].

Leaks of the second type, which can be called permanent, have a different character. Even with the normal operation of heating networks, there are leaks, the magnitude of which does not exceed technically permissible limits. They are associated with errors in the connections of pipelines, in seals, compensators, regulating devices, with the occurrence of small end-to-end violations in the walls – fistulas. Until recently, there was no reliable way to detect such leaks, as they practically do not manifest themselves in control systems (unlike emergency ones).

There are many systems for diagnosing heating equipment. Some of them are given in [11, 12, 13, 14]. Further thermal imaging of pipelines will be considered.

The task of identifying permanent leaks is of great practical importance. They are not only long-lasting sources of environmental impact, but also nascent foci of destruction of pipeline walls (accidents) due to the acceleration of the corrosion process. In addition, small leaks lead to a sharp increase in heat loss through damp insulation, losses of heating water, which have to be replenished. According to data from domestic heat power companies, the total actual amount of leakage in large cities is on average 2 l/s per 1 km2.

2 Research Methods

2.1 Thermal Aerial Photography Features

Thermal infrared (IR) aerial photography is the only remote method that allows, like X-rays, not only to “see” almost all underground heating networks (pipes with a diameter of 50 mm or less), but also to qualitatively evaluate their condition. This is explained as follows. A conductive heat flux spreads from a hot heat conductor, due to which a “thermal trace” forms on the surface of the earth, where the intensity of IR radiation is higher than the background. It is quite obvious that the geometrical parameters and expressiveness of such a trace depend on the diameter of the heat pipe, the temperature of the heat carrier, the method and the depth of the laying.

At the stage of experimental and methodical work, a complex of ground-based thermometric observations synchronous with aerial surveys was performed. Comparison of aerial survey infrared materials and a priori information about design features, laying depth, temperature in networks and condition of heat pipes allowed to get an idea of the nature of the manifestation in the thermal field of tracks of different diameters in different states – from normal to emergency.

The data obtained in the future are used as standards in the interpretation of thermal will and assess the status of thermal networks. Naturally, this is a qualitative assessment, since remotely measuring the absolute thermodynamic temperature is impossible in principle.

The data obtained in the future are used as standards in the interpretation of the thermal field and the assessment of the state of thermal networks. Naturally, this is a qualitative assessment, since remotely measuring the absolute thermodynamic temperature is impossible in principle.

2.2 Processing of Thermal Images

Next will be considered the input thermal image of rank 0, obtained from the thermal imaging device. If its dimensions are 640 × 480 (307200) pixels, then transformation to rank 1 will give an image with dimensions of 320 × 240 (76800) pixels.

The color image is brought to grayscale by any of the known methods:
  1. 1.

    according to the CCIR-601 standard

    $$ F[y,x] = 0.299R + 0.587G + 0.114B; $$
  2. 2.

    by the arithmetic mean value of the color components of the three channels

    $$ F[y,x] = {{\left( {R + G + B} \right)} \mathord{\left/ {\vphantom {{\left( {R + G + B} \right)} 3}} \right. \kern-0pt} 3} ; $$
  3. 3.

    fast (using an algorithm with green pixels)

$$ F[y,x] = G . $$
where R – red, G – green, B – blue (digital image components).
Further, it is rational to treat an image as a rectangular matrix of n × m elements whose values lie in the range from 0 to 127. Matrix filters are used to solve image pre-processing tasks that perform the convolution operation, which allows obtaining response values, taking into account the values of the surrounding pixels, within the dimension kernels. For highly noisy thermal images, it is necessary to use masks of relatively high resolution [15]. For the analysis of thermal images, it is proposed to conduct a discrete Laplacian with a filter of 11 × 11 dimensions
$$ D_{y,x}^{2} = \begin{array}{*{20}c} { - 1} & { - 2} & { - 3} & { - 3} & { - 3} & { - 3} & { - 3} & { - 3} & { - 3} & { - 2} & { - 1} \\ { - 2} & { - 4} & { - 6} & { - 6} & { - 6} & { - 6} & { - 6} & { - 6} & { - 6} & { - 4} & { - 2} \\ { - 3} & { - 6} & { - 9} & { - 9} & { - 9} & { - 9} & { - 9} & { - 9} & { - 9} & { - 6} & { - 3} \\ { - 3} & { - 6} & { - 9} & 0 & 9 & {18} & 9 & 0 & { - 9} & { - 6} & { - 3} \\ { - 3} & { - 6} & { - 9} & 9 & {27} & {45} & {27} & 9 & { - 9} & { - 6} & { - 3} \\ { - 3} & { - 6} & { - 9} & {18} & {45} & {72} & {45} & {18} & { - 9} & { - 6} & { - 3} \\ { - 3} & { - 6} & { - 9} & 9 & {27} & {45} & {27} & 9 & { - 9} & { - 6} & { - 3} \\ { - 3} & { - 6} & { - 9} & 0 & 9 & {18} & 9 & 0 & { - 9} & { - 6} & { - 3} \\ { - 3} & { - 6} & { - 9} & { - 9} & { - 9} & { - 9} & { - 9} & { - 9} & { - 9} & { - 6} & { - 3} \\ { - 2} & { - 4} & { - 6} & { - 6} & { - 6} & { - 6} & { - 6} & { - 6} & { - 6} & { - 4} & { - 2} \\ { - 1} & { - 2} & { - 3} & { - 3} & { - 3} & { - 3} & { - 3} & { - 3} & { - 3} & { - 2} & { - 1} \\ \end{array} $$
Further, the central element of the filter is superimposed on the studied pixel. The remaining elements are also superimposed on neighboring pixels [16]. Next, the sum is calculated, where the terms are the multiplied values of the pixels and the values of the cell of the core that covered the given pixel:
$$ G[y,x] = \left( {\sum\limits_{dy = - 5}^{5} {\sum\limits_{dx = - 5}^{5} {\left( {\begin{array}{*{20}l} {F^{i} \left[ {y + dy,x + dx} \right] \times } \hfill \\ { \times D_{y,x}^{2} \left[ {dy,dx} \right]} \hfill \\ \end{array} } \right)} } } \right)shr2 , $$
where \( F^{i} \) – rank matrix I ϵ [0,1], shr2 – operation of logical shift right by two digits.

A logical shift to the right by two bits (division by 4) of the result of the convolution is made to bring it into the range [−32768, 32767], what twice reduces the amount of necessary memory and allows to get rid of the reaction to noise in the image. The greater the dimension of the convolution kernel, the more accurate the response can be expected from the current pixel, since the set of neighboring pixels are also involved in the convolution operation, which leads to a large number of calculations.

Image processing by 11 × 11 filter implies a large number of multiplication operations, and there are two-digit numbers in this matrix, which will lead to an increase in time to calculate the value of one response. As a result, to reduce the computational cost, approximately 3 times, instead of the 11 × 11 operator, you should use an operator with a kernel size of 7 × 7 elements and then double the operator with a core of 3 × 3 elements, which will give an equivalent result when processed by one 11 × 11 operator.

The method of competitive analysis gives a good result for pattern recognition on non-noisy images [17]. However, in the conditions of noise, poor visibility or the presence of foreign objects it is also possible to apply the contour method using masks of a higher dimension, as well as introducing the definition of “singular points” – extreme response values on such images.

To determine the specific points, the algorithm of the Moravec detector [18] is used as a basis, which compares the extremes at the corners of the image using local detectors. A black-white image arrives at the detector input. At the output, a matrix is formed with elements whose values determine the degree of plausibility of finding the angle in the corresponding pixels of the image. The threshold value allows to “drop” the pixels, the degree of likelihood of which is less than the threshold. The remaining points are special or extremes. The Moravec detector is a simple angle detector, estimating the change in pixel intensity (y, x) by offsetting a square window centered in the current pixel (y, x) by one pixel in each of the 8 directions [19]. This method is implemented as follows:
  • for each direction of displacement \( F\left( {u,v} \right) \in \left\{ {\left( {1,0} \right),\left( {1,1} \right),\left( {0,1} \right),\left( { - 1,1} \right),\left( { - 1,0} \right),\left( { - 1, - 1} \right),\left( {0, - 1} \right),\left( {1, - 1} \right)} \right\} \), the change in intensity is calculated

    $$ V_{u,v} (y,x) = \sum\nolimits_{\forall a,b} {\left( {I\left( {x + u + a,y + v + b} \right) - I\left( {x + a,y + b} \right)} \right)}^{2} , $$
    where I(x + a) is the intensity of a pixel with coordinates (y, x) in the source image;
  • builds a map of the probability of finding the angles in each pixel of the image by calculating the estimated function. Essentially, the direction is determined, which corresponds to the smallest change in intensity, because the corner must have adjacent edges;

  • pixels are cut off, in which the values of the evaluation function are below a certain threshold value;

  • recurring corners are removed using the NMS procedure (non-maximal suppression);

  • non-zero elements correspond to the angles in the image.

The use of the Moravec detector makes it possible not to calculate the change in intensity, but to immediately perform an analysis on the generated response matrix. The table of directions will determine the maximum and minimum values of the response value located in the center of the window:
$$ f = \left\{ {\begin{array}{*{20}c} {G\left[ {y,x} \right] > 0\;AND\;G[y,x] > G[y + dy,x + dx]} \\ {G\left[ {y,x} \right] < 0\;AND\;G[y,x] < G[y + dy,x + dx]} \\ \end{array} } \right. $$
for all extremums \( dy = dx \in \left[ { - 1;0} \right)\;AND\;\left( {0,1} \right] \). Extremes can be in maximum proximity to each other through one element of the response, therefore, in order to reduce the number of iterations when an extremum is found, the following, in the direction of the sweep, the value is excluded from the comparison procedure.
Approaching the solution of the segmentation problem, namely the definition of related groups of elements in the image, it is necessary to select a common feature or threshold value that will allow to divide the desired signal into classes. The threshold separation operation is to compare the brightness value of each pixel of the image with the specified threshold value, and can be represented by a filter:
$$ B[y,x] = \left\{ {\begin{array}{*{20}c} {1,if\;A[y,x] \ge \Delta T;} \\ {0,if\;A[y,x] < \Delta T,} \\ \end{array} } \right. $$
where ΔT is a threshold value.

On images with monochromatic objects, the binarization threshold is selected from the histogram of brightness, examples of which are given in Subsect. 2.4. As a rule, light (warmer) areas appear as objects or search areas on thermal images. Depending on the temperature, areas can be a part of the wall, a staircase, and other geometric figures of a certain size, which fall, for example, under a warm stream of air or have heat-generating elements in their physical structure. The difference in brightness can lie in the whole range [0, 255]. That is why there is a need to get rid of noise, reducing the maximum value to 128.

In some cases, the Otsu method can be used to find the threshold, which determines the threshold that minimizes the variation of the pixel brightness in an object. The threshold is chosen between the highest pixel values in the histogram.

Having a generated table of extrema GE and image matrix F1, which is formed as a result of applying the Laplacian mask, the binarization threshold can be found by the formula:
$$ \Delta T = \frac{{\sum\limits_{i = 1}^{n} {G^{E} R_{i} \left| \sigma \right|} }}{{n\bar{X}}}, $$
where Ri is the i-th element from the extremum list. The coefficient of variation is necessary in order to characterize the one-sidedness of data, values and stability of processes, reflect the degree of variation of values regardless of the scale of measurements. This method of processing thermal images is described in more detail in [20].

One pixel in the image that is not tied to a coherent group of similar pixels of the analyzed segment is estimated by the system as noise and will not be an area due to which it was necessary to reduce the brightness range by one row to the right. Due to this, the Laplacian mask will not react to not so significant single pixels. Theoretically, the input image of 320 × 240 pixels allows to contain 8480 areas (the minimum size is 2 × 2 pixels with a minimum distance between them of one pixel), for a finite number of areas it can be take the value 0xFFFD.

Binarization in the response matrix is performed by cutting the response over the threshold and generating two matrices of domains of the same dimension for responses with positive (GP) and negative (GN) values with the conditions:

$$ \begin{aligned} G^{P} \left[ {y,x} \right] = \left\{ {\begin{array}{*{20}c} {0xFFFD,\;if\;G[y,x] \ge \Delta T;} \\ {0x0000,\;if\;G[y,x] < \Delta T,} \\ \end{array} } \right. \hfill \\ G^{N} \left[ {y,x} \right] = \left\{ {\begin{array}{*{20}c} {0xFFFD,\;if\;G[y,x] \le \Delta T;} \\ {0x0000,\;if\;G[y,x] > \Delta T.} \\ \end{array} } \right. \hfill \\ \end{aligned} $$

2.3 UAV Based Monitoring System

To carry out experimental studies of the monitoring system of the technical condition of heat pipelines, a multi-rotor type unmanned aerial vehicle, model MJX BUGS 3, was used. This quadcopter will also be used to study the effect of air dilution on the concentration of air components [21, 22] in the following works.

The MJX BUGS 3 quadrocopter (Fig. 2) is a world-renowned radio control toy manufacturer Meijiaxin Toys. The company is positioning this drone designed for both aerial photography and dynamic flights.
Fig. 2.


The features of this quadrocopter include:
  • support for 3S batteries;

  • control of battery charge and flight distance;

  • long flight time;

  • axle suspension;

  • control at 2.4 GHz;

  • 360o flip;

  • LED-backlight.

The quadcopter case MJX BUGS 3 is made of nylon fiber, has established itself as a reliable and durable material, while the supports are made of ordinary plastic.

The quadcopter MJX BUGS 3 is equipped with brushless motors of type MT1806 with a capacity of 1800 kV. The manufacturer describes them as economical and efficient among the same type of brushless motors. Each motor provides 230 grams of traction.

Also in quadcopter available speed controllers ESC with automatic anti-jamming, eliminating the possibility of burnout engines.

Included with the drone is a axle-free suspension with manual vertical adjustment, adapted for installing a small load. The distance from the ground to the suspension is 80 mm.

The quadcopter is powered by a 2S Li-Po battery with a capacity of 1800 mAh with a discharge current of 25C and an XT30 connector. According to the specification, the battery provides 19 min of continuous flight.

The quadcopter kit also includes hardware operating at 2.4 GHz. Its distinctive feature is the function of intelligent remote control, reports a low battery or a long distance of the drone from the equipment. It is powered by 4 AA batteries. The maximum distance of the drone from the equipment is 300–500 m [23].

During flight tests, the MJX BUGS 3 shows good flight performance even on the type 2S battery included in the kit. The 6-axis gyroscope works smoothly. In practice, the flight time of the quadcopter with a maximum load was 8 min. At a distance of 300 meters, the quadcopter clearly performs the specified flight directions [24].

The main advantages and disadvantages of the quadrocopter MJX BUGS 3 are shown in Table 1.
Table 1.

Advantages and disadvantages of the model MJX BUGS 3



• ease of using

• power

• feedback function with low charge and critical distance

• suspension for cargo

• compatible with 3S-batteries

• LED backlight

• price

• lack of dynamism

• charged battery as standard

To perform experiments with thermal imaging of heat supply pipelines on the basis of the UAV, a compact thermal imaging camera manufactured by Seek Thermal ™ (USA) (Fig. 3) was installed, which has a wide-angle lens with a total size of 2.5 × 4.4 × 2.5 cm resolution of 320 × 240. The greatest shooting distance is 610 meters, and the closest distance is 15 cm.
Fig. 3.

XR Compact thermal imaging camera (Seek Thermal)

Thermal Compact XR was used as a thermal imaging camera. Its technical characteristics are given in Table 2.
Table 2.

Thermal camera parameters


206 × 156

Working distance

to 550 m

Viewing angle


Pixel size

12 μ

Spectral range

7.5–14 μ



To implement the method of thermal aerial photography at the Institute of Engineering Thermophysics of the National Academy of Sciences of Ukraine an equipment complex was developed for diagnosing the state of heat equipment (boilers [25, 26] and main heat pipelines). Its appearance is shown in Fig. 4, where 1 – UAV, 2 – a thermal imaging camera, 3 – a communicator (smartphone).
Fig. 4.

Hardware complex for monitoring the technical condition of main pipelines (left – top view; right – bottom view): 1 - UAV, 2 - thermal imaging camera; 3 - communication device

2.4 The Results of Experimental Studies

During the initial stage of the research, studies were carried out on the ground parts of the main pipelines of heating networks. Fluke Ti50FT was used as a thermal imaging camera on a UAV. Figures 5 and 6 show a section of a land trunk pipeline of a heating network obtained in the visible and infrared range.
Fig. 5.

Photo of the main pipeline of the heating network in the visible range

Fig. 6.

Photo of the main pipeline of the heating network in the IR range (Fluke Ti50FT)

Figure 7 shows the histogram of pixel distribution of Fig. 6 by temperature. Figure 8 depicts a 3D model of Fig. 6 taking into account the temperature characteristics. The parameters of Fig. 6 are shown in Table 3.
Fig. 7.

The histogram of pixels distribution of the Fig. 4 by temperature ranges

Fig. 8.

3D model of Fig. 4 taking into account the temperature characteristics

Table 3.

Parameters of thermographic image (Fig. 6)

Background temperature

7.00 °C

Radiation coefficient


Transmission coefficient


Average temperature

14.87 °C

Image borders

from 6.38 °C to 30.06 °C

Camera model


IR sensor size

320 × 240

File name


Humidity setting

0 RH % 0 m

For monitoring extended objects (in our case, pipelines of heat networks) it is proposed to fly around the test object using a multi-rotor type UAV [27]. This allows to get high-quality photos and thermal images of the site of the heating system as an object of control for further analysis. The software allows you to use any topographic basis as a map. Binding can be done at two or more points. It is also possible to use as a topological basis of electronic maps [28, 29]. The program provides input, automatic control and editing of the route of the flight. An elevation can be specified for each waypoint.

The results of measuring the thermal state of the heat grids were carried out using a thermal imaging camera, mounted on a UAV.

Figure 9 shows thermal images of sections of the heat network where experimental studies were conducted. The shooting was performed on November 22, 2018 at 18 p.m. on a cloudless sky at an air temperature of −4 °C.
Fig. 9.

The working environment of the software complex based on Seek Thermal

Figure 9 clearly shows the possibility of identifying defective areas of main pipelines using low power UAVs.

The use of UAVs allowed us to distinguish 4 technical conditions of the main pipelines of the heating networks shown in Table 4.
Table 4.

Assessing the state of the main heating networks with UAVs


% heat losses




Dry and integral insulation of pipelines, minimum heat flow from the heat carrier to the earth’s surface



Wet or broken insulation of pipelines, contributes to the nucleation of corrosive damage; in the thermal field can be displayed by a clear anomaly of the average brightness level and an increased width of the thermal trace



Disturbed and damp isolation of pipelines, the canal is often filled with water from neighboring water pipelines, groundwater or melt water; in the thermal field is displayed as a high-contrast anomaly with a width several times larger than normal



Violation of the integrity of the pipeline with the heat carrier spill. Thermal field anomalies have a very high contrast and a broad, diffuse shape, due to the microrelief features

3 Conclusions

Today we can state the rapid development of UAVs, which are mainly used in military operations. The list of subject areas of the use of UAVs for various other studies, operations, conventionally called non-military, is essentially limited. First of all, this limitation is due to the lack of created, developed and manufactured technical tools for conducting diverse studies, primarily measuring tools. It can be predicted that such an imbalance will soon be broken and the process of creating appropriate equipment for the UAV will be adjusted to conduct a wide range of research in various subject areas, among which the defense industry of the states will be priority and prospective.

The creation of mobile information-measuring systems based on UAVs makes it possible to diagnose the state and dynamics of the characteristics in time and space of the studied environment, both in on-line modes and other modes. The on-line mode is especially effective in case of accidents in areas of spatially branched heat networks. With normal operation of the objects under study, the current remote control is the most economical compared with other means of control. This allows to use such measuring tools to create the necessary databases for diagnostics the characteristics of the thermal state of heat networks to predict their dynamics.


  1. 1.
    Babak, S., Babak, V., Zaporozhets, A., Sverdlova, A.: Method of statistical spline functions for solving problems of data approximation and prediction of object state. In: CEUR Workshop Proceedings, vol. 2353, pp. 810–821 (2019).
  2. 2.
    Babak, V.P.: Hardware-Software for Monitoring the Objects of Generation Transportation and Consumption of Thermal Energy. Institute of Engineering Thermophysics of NAS of Ukraine, Kyiv (2016)Google Scholar
  3. 3.
    Prudyus, I., Shkliarskiy, V., Turkinov, G., Storozh, V., Kril, Y.: Analysis and choice of methods of monitoring technical condition of the heating systems. In: International Conference on “Modern Problems of Radio Engineering, Telecommunications and Computer Science (TCSET)”, p. 154 (2008)Google Scholar
  4. 4.
    Johnson, B., Barth, R., House, P., Kuntscher, J.: Controlling pipe and equipment operating temperatures with trace heating systems. In: PCIC Europe, pp. 1–10 (2013)Google Scholar
  5. 5.
    Friman, O., Follo, P., Ahlberg, J., Sjökvist, S.: Methods for large-scale monitoring heating systems using airborne thermography. IEEE Trans. Geosci. Remote Sens. 51(8), 5175–5182 (2014)CrossRefGoogle Scholar
  6. 6.
    Shang-bin, J., Le, F., Peng-yue, W., Lei, Q., Yong-ze, J., Qian-kun, S.: Assessment and prediction of leakage degree of expansion joints in underground network heating pipe network based on LS-SVM and ARIMA model. In: 13th IEEE Conference on Industrial Electronics and Applications (ICIEA), pp. 1697–1702 (2018)Google Scholar
  7. 7.
    Jiao, S.-B., Fan, L., Wu, P.-Y., Qiao, L., Wang, Y., Xie, G.: Assessment of leakage degree of underground heating primary pipe network based on chaotic simulated annealing neural network. In: Chinese Automation Congress (CAC), pp. 5895–5900 (2017)Google Scholar
  8. 8.
    Babak, V.P., Zaporozhets, A.A., Kovtun, S.I., Sergienko, R.V.: Diagnosing methods analysis of bulk heating systems technical condition. Sci. Heritage 1(14), 59–65 (2017)Google Scholar
  9. 9.
    Teng, Q., Wang, W.: The optimization and management research for central heating system. In: IEEE Workshop on Advanced Research and Technology in Industry Applications (WARTIA), pp. 175–177 (2014)Google Scholar
  10. 10.
    Fukuoka, M., Tang, L.: Wireless pipe inspection control system. In: The 5th International Conference on Automation, Robotics and Applications, pp. 497–502 (2011)Google Scholar
  11. 11.
    Babak, V., Zaporozhets, A., Kovtun, S., Serhiienko, R.: Methods and means of heat losses monitoring for heat pipelines. Int. J. “NDT Days” 1(2), 213–221 (2018)Google Scholar
  12. 12.
    Zaporozhets, A.A., Eremenko, V.S., Serhiienko, R.V., Ivanov, S.A.: Development of an intelligent system for diagnosing the technical condition of the heat power equipment. In: IEEE 13th International Scientific and Technical Conference on Computer Sciences and Information Technologies (CSIT), pp. 48–51 (2018)Google Scholar
  13. 13.
    Zaporozhets, A., Eremenko, V., Serhiienko, R., Ivanov, S.: Methods and hardware for diagnosing thermal power equipment based on smart grid technology. In: Shakhovska, N., Medykovskyy, M. (eds.) Advances in Intelligent Systems and Computing III, vol. 871, pp. 476–489. Springer, Cham (2019)Google Scholar
  14. 14.
    Eremenko, V., Zaporozhets, A., Isaenko, V., Babikova, K.: Application of Wavelet Transform for Determining Diagnostic Signs. In: CEUR Workshop Proceedings, vol. 2387, pp. 202–214 (2019).
  15. 15.
    Kosar, O., Shakhovska, N. An overview of denoising methods for different types of noises present on graphic images. In: Shakhovska, N., Medykovskyy, M. (eds.) Advances in Intelligent Systems and Computing III, vol. 871, pp. 38–47. Springer, Cham (2019)Google Scholar
  16. 16.
    Gauci, J., Falzon, O., Formosa, C., Gatt, A., Ellul, C., Mizzi, S., Mizzi, A., Delia, C.S., Cassar, K., Chockalingam, N., Camilleri, K.P.: Automated region extraction from thermal images for peripheral vascular disease monitoring. J. Healthc. Eng. 5092064 (2018)Google Scholar
  17. 17.
    Duarte, A., Carrao, L., Espanha, M., Viana, T., Freitas, D., Bartolo, P., Faria, P., Almeida, H.A.: Segmentation algorithms for thermal images. Procedia Technol. 16, 1560–1569 (2014)CrossRefGoogle Scholar
  18. 18.
    Chen, X., Liu, L., Song, J., Li, J., Zhang, Z.: Corner detection and matching for infrared image based on double ring mask and adaptive SUSAN algorithm. Opt. Quant. Electron. 50, 194 (2018)CrossRefGoogle Scholar
  19. 19.
    Wang, X., Song, H., Cui, H.: Pedestrian abnormal event detection based on multi-feature fusion in traffic video. Optik 154, 22–32 (2018)CrossRefGoogle Scholar
  20. 20.
    Loginov, I.D.: Processing and segmentation of thermal images. Young Sci. 13(147), 62–71 (2017)Google Scholar
  21. 21.
    Babak, V.P., Mokiychuk, V., Zaporozhets, A., Redko, O.: Improving the efficiency of fuel combustion with regard to the uncertainty of measuring oxygen concentration. Eastern-Eur. J. Enterp. Technol. 6(8(84)), 54–59 (2016)CrossRefGoogle Scholar
  22. 22.
    Zaporozhets, A.O., Redko, O.O., Babak, V.P., Eremenko, V.S., Mokiychuk, V.M.: Method of indirect measurement of oxygen concentration in the air. Naukovyi Visnyk Natsionalnoho Hirnychoho Universytetu 5, 105–114 (2018)CrossRefGoogle Scholar
  23. 23.
    Devitt, D., Morozov, R., Medvedev, M., Shapoval, I., Konovalov, G.: Implementation of the hybrid technology for quadcopter motion control in a complex non-deterministic environment. In: 18th International Conference on Control, Automation and Systems (ICCAS), pp. 451–456 (2018)Google Scholar
  24. 24.
    Talha, M., Asghar, F., Rohan, A., Rabah, M., Kim, S.H.: Fuzzy logic-based robust and autonomous safe landing for UAV quadcopter. Arab. J. Sci. Eng. 44(3), 2627–2639 (2018)CrossRefGoogle Scholar
  25. 25.
    Zaporozhets, A.: Development of software for fuel combustion control system based on frequency regulator. In: CEUR Workshop Proceedings, vol. 2387, pp. 223–230 (2019).
  26. 26.
    Zaporozhets, A.: Analysis of control system of fuel combustion in boilers with oxygen sensor. Periodica Polytechnica Mech. Eng. (2019).
  27. 27.
    Gomez, C., Green, D.R.: Small unmanned airborne systems to support oil and gas pipeline monitoring and mapping. Arab. J. Geosci. 10, 202 (2017). Scholar
  28. 28.
    Allred, B., Eash, N., Freeland, R., Martinez, L., Wishart, D.: Effective and efficient agricultural drainage pipe mapping with UAS thermal infrared imagery: a case study. Agric. Water Manag. 197, 132–137 (2018). Scholar
  29. 29.
    Zaporozhets A., Kovtun S., Dekusha, O.: Determination of the technical condition of heating networks based on the processing of thermal imaging. In: Proceedings of International Scientific Conference Computer Sciences and Information Technologies (CSIT-2019), vol. 2, pp. 5–8 (2019)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  1. 1.Institute of Engineering Thermophysics of NAS of UkraineKyivUkraine

Personalised recommendations