CN113447159A - Wind tunnel cold leakage monitoring method based on global normal distribution and local outlier factor - Google Patents

Abstract

The invention discloses a wind tunnel cold leakage monitoring method based on global normal distribution and local outlier factors, which comprises the steps of obtaining temperature measuring point data of a temperature measuring optical fiber laid on the outer surface of a wind tunnel; performing global anomaly detection on all temperature measuring points based on normal distribution to obtain global anomaly points; local anomaly screening is carried out on all global anomaly points based on the local outlier factors, and the screened local anomaly points are used as stable cold leakage points; and (4) establishing a cold leakage area by taking all stable cold leakage points as supports. The invention provides a wind tunnel cold leakage monitoring method based on global normal distribution and local outliers, which aims to solve the problems that the false alarm rate is high and the situation of cold leakage and liquid leakage on the surface of a wind tunnel is difficult to effectively acquire when the surface temperature of the wind tunnel is monitored by adopting distributed optical fibers in the prior art, and achieves the purposes of reducing the false alarm rate and effectively detecting the situation of cold leakage and liquid leakage on the surface of the wind tunnel.

Description

Wind tunnel cold leakage monitoring method based on global normal distribution and local outlier factor

Technical Field

The invention relates to the field of wind tunnels, in particular to a wind tunnel cold leakage monitoring method based on global normal distribution and local outlier factors.

Background

In the wind tunnel test, simply speaking, according to the principle of relativity of motion, a model or a real object of an aircraft is fixed in a ground artificial environment, and airflow is artificially made to flow through, so that various complex flight states in the air are simulated, and test data are obtained. In order to meet the temperature requirements of some special tests, a heat insulation layer is required to be paved on the surface of the wind tunnel, and the surface temperature of the heat insulation layer is not lower than the local dew point temperature under the normal condition according to the design requirement of the heat insulation layer. The failure of the thermal insulation layer can cause the temperature of the surface of the hole to drop. In the operation process of the wind tunnel, the state of the heat-insulating layer needs to be monitored so as to judge whether the heat-insulating layer fails or not. However, the conventional infrared temperature measurement technology has the following disadvantages: precision is reduced due to the influence of factors such as distance difference between the infrared temperature measuring equipment and the surface of the wind tunnel, formed observation angle and even surface smoothness difference; moreover, as the wind tunnel equipment has complex field environment and the wind tunnel of the object to be measured is large and irregular in volume, the ladder platform is arranged on the periphery of the tunnel body, and equipment such as a measurement and control cabinet is arranged on the periphery of the tunnel body, so that the ladder platform, the measurement and control cabinet and the like can shield infrared imaging. Therefore, the monitoring of cold leakage from the surface of a wind tunnel has been a technical problem in the art.

In order to overcome the technical problems, the applicant of the present application develops a distributed optical fiber temperature measurement system used in the field of wind tunnel temperature measurement to cooperate with a vision temperature measurement system to monitor the surface temperature of the wind tunnel of the measured object. The core of the distributed optical fiber temperature measurement system is that temperature measurement optical fibers are distributed at the bottom of the wind tunnel and used for measuring the surface temperature of the tunnel body which cannot be covered by the vision temperature measurement system. However, this system encounters the following difficulties in the design process: (1) because the temperature measuring optical fiber can only obtain the relationship between the length of the optical fiber and the temperature, the relationship between the temperature measured by the optical fiber and the position of the wind tunnel cannot be known, and the condition of cold leakage and liquid leakage on the surface of the wind tunnel is difficult to effectively obtain; (2) the false alarm rate of the system is high, and the prior art cannot effectively eliminate the false alarm rate.

Disclosure of Invention

The invention provides a wind tunnel cold leakage monitoring method based on global normal distribution and local outliers, which aims to solve the problems that the false alarm rate is high and the situation of cold leakage and liquid leakage on the surface of a wind tunnel is difficult to effectively acquire when the surface temperature of the wind tunnel is monitored by adopting distributed optical fibers in the prior art, and achieves the purposes of reducing the false alarm rate and effectively detecting the situation of cold leakage and liquid leakage on the surface of the wind tunnel.

The invention is realized by the following technical scheme:

the wind tunnel cold leakage monitoring method based on the global normal distribution and the local outlier factor comprises the following steps:

acquiring temperature measuring point data of a temperature measuring optical fiber paved on the outer surface of the wind tunnel;

performing global anomaly detection on all temperature measuring points based on normal distribution to obtain global anomaly points;

local anomaly screening is carried out on all global anomaly points based on the local outlier factors, and the screened local anomaly points are used as stable cold leakage points;

and (4) establishing a cold leakage area by taking all stable cold leakage points as supports.

Aiming at the problems that the false alarm rate is high and the leakage condition of cold on the surface of a wind tunnel is difficult to effectively obtain when the surface temperature of the wind tunnel is monitored by adopting a distributed optical fiber in the prior art, the invention provides a wind tunnel leakage monitoring method based on global normal distribution and local outlier factors, the method comprises the steps of firstly obtaining temperature measurement point data of the temperature measurement optical fiber laid on the outer surface of the wind tunnel, then preliminarily judging abnormal points with the temperature lower than all the measurement points by a global abnormal detection method, and defining the abnormal points as global abnormal points; and then, acquiring a final abnormal detection point by a local abnormal detection method, and defining the final abnormal detection point as a stable cold leakage point. And finally, taking all stable cold leakage points as supports to obtain a finally determined cold leakage area. The global anomaly detection method is realized based on normal distribution, and the local anomaly detection method is realized based on local outlier factors. The method detects candidate possible abnormal points through detection based on global normal distribution abnormal points, then constructs local abnormal temperature region characteristics through a method based on local outlier factors according to the layout of actual heat insulation blocks and the layout of optical fibers, and realizes screening and identification of the candidate global abnormal points, so that false alarm can be effectively eliminated, accurate identification of cold leakage of a tunnel body is further realized, and cold leakage and liquid leakage conditions of the surface of the wind tunnel in a temperature measurement optical fiber monitoring region are rapidly and accurately obtained.

Further, the temperature measurement point data includes coordinates and temperature.

Further, the temperature measurement point data is a three-dimensional scattered point coordinate system, and the establishment method of the three-dimensional scattered point coordinate system comprises the following steps:

acquiring basic data of temperature measuring pointsX,Y,Z,T) WhereinX,Y,ZIs the three-axis coordinate of the temperature measuring point in the wind tunnel coordinate system,Tthe temperature measured by the temperature measuring point;

projecting the temperature measurement point toX,Y) Forming a grid point on the plane;

constructing a temperature value from the grid pointsTThe three-dimensional scatter coordinate system is formed.

Temperature measurement optical fiber can accurately measure temperature, but the obtained temperature data is single-point, so the temperature measurement point data in the scheme is a three-dimensional scattered point coordinate system, the establishment process firstly acquires the temperature and coordinate information of the temperature measurement point as basic data, and (a) is collectedX,Y,Z,T) In which coordinate information is includedX,Y,ZAnd temperatureT(ii) a Then projecting the optical fiber temperature measuring point toX,Y) Forming a grid point on the two-dimensional plane, and combining the temperature valueTIs composed of (X,Y, T) Is a three-dimensional scatter coordinate system ofSubsequent global anomaly detection is fully prepared.

Further, the method for detecting global anomaly of all temperature measurement points comprises the following steps:

temperature point composition obtained by measuring with temperature measuring optical fibernDimension vector (a)T ₁,T ₂,…,T _n )；

Calculate thenMean of dimensional vectorsuVariance, varianceδ；

Measured temperature data less thanu-cδ) The temperature measuring point is used as a global abnormal point; whereincIs a constant.

The scheme refines the process of global anomaly detection, wherein the temperature of each temperature measuring point is extracted from a three-dimensional scattered point coordinate system established in advance to formnDimension vector (a)T ₁,T ₂,…,T _n ) Then calculating to obtain the average value of the global temperatureuVariance, varianceδ：

Under the assumption of normal distribution, the temperature data is considered to be less than: (A)u-cδ) And marking the corresponding optical fiber temperature measurement point as a global abnormal point if the optical fiber temperature measurement point is an abnormal value.

Further, the method for screening local anomalies of all global anomaly points comprises the following steps: extracting local outlier factors for global outliersLOFScreening ofLOFThe point where the value is greater than the set threshold is taken as the stable cold leakage point.

According to the scheme, abnormal points obtained by a normal distribution-based global abnormal point detection algorithm are screened, and abnormal points with drastic abnormal changes are screened. The principle is to locally detect points after global abnormal pointsLOFExtracting and screeningLOFThe point with a large value is taken as a stable cold leakage point. By means of game-playIn the case of detection of partial abnormalityLOFThe detection precision of cold leakage points can be controlled and the false alarm rate can be reduced by adaptively setting the threshold value; the threshold value is selected according to the specific conditions of different wind tunnels. The scheme introduces local outlier factorsLOFThe method has the advantages that the optical fiber temperature measurement data are further screened according to the method, so that the points with strong local change in the points subjected to global abnormal screening are finally screened, the elimination of the false alarm of the system is fully ensured, and the detection precision of the cold leakage and liquid leakage conditions on the surface of the wind tunnel is effectively improved.

Further, local outlier factors of global outliersLOFCalculated by the following formula:

in the formula (I), the compound is shown in the specification,pis the calculated global anomaly point;kis as followskA neighborhood;LOF _k (p) is a pointpIn the first placekLocal outlier factors within the neighborhood;N _k (p) is a pointpIn the first placekNeighborhood points within a neighborhood;sis a pointpIn the first placekSampling points in the neighborhood;lrd _k (s) is a dotsLocal achievable density of;lrd _k (p) is the local achievable density of point p.

lrdThe local reachable density is represented by the inverse of the average reachable distance from a point in the neighborhood of the point.

Local Outlier Factor (Local Outlier Factor)LOF) Is to measure a certain pointpNeighborhood point ofN _k (p) local reachable density and pointpAverage of the locally achievable density ratio of (a), in this applicationLOFThe calculation formula of (a) is as follows:

；

if the above ratio is close to 1, it indicates thatpAs much as its neighborhood point density,ppossible sibling with the neighborhoodA cluster of; if the ratio is less than 1, this indicatespIs higher than the density of its neighborhood points, v is a density point; if the ratio is greater than 1, this indicatespIs less than its neighborhood point density,pthe more likely it is an outlier. As can be seen from the above, the present invention,LOFby comparisonpJudging whether the point is an abnormal point or not by the density of the point and the neighborhood point, if so, judging whether the point is an abnormal point or notpThe lower the density of (a), the more likely it is to be identified as an anomaly. The farther the distance between the points is, the lower the density, and the closer the distance is, the higher the density.LOFThe density is calculated by the first of the pointskNeighborhood computation, rather than global computation, it is a local anomaly detection method, extracted for each pointLOFThe features are local features.

Further, the method for establishing the cold leakage area comprises the following steps:

eliminating temperature values in the temperature measurement point data, and forming two-dimensional grid points under an O-XY two-dimensional coordinate system;

solving for a von neumonian map for the two-dimensional grid points;

solving the association between the cooling leakage points in the Von Rogoyi graph to obtain a two-dimensional cooling leakage area;

and mapping the two-dimensional cold leakage area from an O-XY two-dimensional coordinate system to an O-XYZ three-dimensional coordinate system.

After the anomaly detection algorithm, stable cold leakage points are screened out, and the points can be used as cold leakage supporting points to establish a cold leakage area. According to the scheme, a cold leakage area is established by adopting an idea in triangulation, a three-dimensional coordinate formed by optical fiber temperature measuring points is used for removing a temperature value and performing a cold leakage abnormity detection algorithm, grid points formed by the optical fiber temperature measuring points in an O-XY coordinate system are used as two-dimensional grid points, a Voronoi diagram (Von-Ronoui diagram) is obtained for the two-dimensional grid points, each temperature measuring point is associated with a certain temperature measuring area in the Voronoi diagram (Von-Ronoui diagram), association relations among the temperature measuring points need to be solved, the two-dimensional cold leakage area is obtained, and finally the two-dimensional cold leakage area is mapped to an O-XYZ three-dimensional coordinate system from the O-XY two-dimensional coordinate system.

Further, a method of solving for associations between stable cold leakage points within a von neumoniae map comprises:

discarding the rest global abnormal points in the two-dimensional grid points, reserving stable cold leakage points, carrying out triangulation, and establishing a Delaunay graph; the cooling leakage region is defined as an aggregate of a von roughy region in a von roughy map and a delaunay region in a delaunay map.

In the scheme, in order to solve the association between the temperature measuring points, triangulation is carried out by discarding the remaining global outliers to establish a delaunay graph (delaunay graph), and the final cold leakage area is formed by a Voronoi area and a delaunay area.

Further, triangulation is achieved using the delaunay triangulation algorithm.

Further, the cold leakage monitoring method is used in a wind tunnel cold leakage monitoring system; the wind tunnel cold leakage monitoring system comprises:

the vision temperature measurement subsystem is used for acquiring a temperature field image and a texture image of the surface of the wind tunnel;

the optical fiber temperature measurement subsystem comprises a temperature measurement optical fiber laid at the lower part of the wind tunnel in a rotating manner, and the temperature measurement optical fiber covers a measurement blind area of the vision temperature measurement subsystem on the surface of the wind tunnel; the optical fiber temperature measurement subsystem is used for measuring the temperature of a measurement blind area of the visual temperature measurement subsystem on the surface of the wind tunnel;

the temperature correction subsystem is used for correcting and evaluating temperature measurement data of the vision temperature measurement subsystem and the optical fiber temperature measurement subsystem;

and the 3D display subsystem is used for displaying the three-dimensional temperature field and the position of a cold leakage point on the surface of the wind tunnel in real time on the 3D model of the wind tunnel.

In the prior art, the following problems exist in monitoring the surface temperature of the wind tunnel: precision is reduced due to the influence of factors such as distance difference between the infrared temperature measuring equipment and the surface of the wind tunnel, formed observation angle and even surface smoothness difference; the wind tunnel equipment site environment is complicated, and the measured object wind tunnel is bulky and irregular, and the tunnel body periphery is equipped with the ladder platform and settles and have equipment such as survey and control cabinet, leads to ladder platform and survey and control cabinet etc. all to cause to shelter from to infrared imaging. Therefore, the monitoring of cold leakage from the surface of a wind tunnel has been a technical problem in the art.

The wind tunnel cold leakage monitoring system adopts the combination of vision and optical fiber temperature measurement positioning technology to monitor the surface temperature change of the tunnel body. And directly mapping the monitoring result to the surface of the wind tunnel 3D model, warning and displaying the abnormal area, and giving the number of the heat insulation block at the position of the leakage point. The method and the device ensure seamless, non-missing and non-error cold leakage monitoring, reduce the redundancy of the system and reduce the waste of the computing resources of the system. The temperature correction subsystem is used for correcting and evaluating the temperature measurement data of the vision temperature measurement subsystem and the optical fiber temperature measurement subsystem so as to guarantee that the temperature measurement is accurate and the temperature measurement data of the subsystems are kept consistent. The vision temperature measurement subsystem is also used for acquiring the texture image of the surface of the hole body. The texture image on the surface of the hole body can be realized by a visible light image acquisition technology, and the appearance change (such as frost condensation, icing and the like) of the hole body caused by cold leakage can be detected and identified by an image identification technology, so that the vision temperature measurement subsystem can provide a secondary judgment function. Preferably, the cold leakage monitoring process of the visual temperature measurement subsystem can be monitored through a deep learning algorithm, so that the self-adaptive capacity of the visual temperature measurement subsystem is improved.

Compared with the prior art, the invention has the following advantages and beneficial effects:

1. according to the wind tunnel cold leakage monitoring method based on the global normal distribution and the local outlier, candidate possible outliers are detected through detection based on the global normal distribution outliers, local abnormal temperature area characteristics are constructed through a method based on the local outlier, the candidate global outliers are screened and identified, false alarms can be effectively eliminated, and cold leakage and liquid leakage of the surface of the wind tunnel in the temperature measurement optical fiber monitoring area can be rapidly and accurately obtained.

2. The wind tunnel cold leakage monitoring method based on the global normal distribution and the local outlier factors detects and identifies global abnormal points through the normal distribution, ensures sufficient preliminary screening of abnormal temperature points, and provides stable and reliable samples for subsequent further screening.

3. The invention is based on global normal distribution and local outlier factorThe wind tunnel cold leakage monitoring method introduces local outlier factorLOFThe method has the advantages that the optical fiber temperature measurement data are further screened according to the method, so that the points with strong local change in the points subjected to global abnormal screening are finally screened, the elimination of the false alarm of the system is fully ensured, and the detection precision of the cold leakage and liquid leakage conditions on the surface of the wind tunnel is effectively improved.

4. The wind tunnel cold leakage monitoring method based on global normal distribution and local outlier adopts the idea of triangulation to establish cold leakage areas, adopts Voronoi (von Lonoui) graphs in triangulation to reflect the influence of each stable cold leakage point on the surrounding areas, and delaunay (Delaunay) graphs to reflect the association between each node, namely the association between the cold leakage points, and finally accurately obtains the cold leakage areas.

Drawings

The accompanying drawings, which are included to provide a further understanding of the embodiments of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the principles of the invention. In the drawings:

FIG. 1 is a schematic flow chart of an embodiment of the present invention;

FIG. 2 is a schematic diagram of three-dimensional scatter coordinates according to an embodiment of the present invention;

FIG. 3 is a diagram illustrating a global anomaly detection result according to an embodiment of the present invention;

FIG. 4 is a diagram illustrating a local anomaly screening result according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of two-dimensional grid points according to an embodiment of the present invention;

FIG. 6 is a von Neuroy diagram of an embodiment of the invention;

FIG. 7 is a Delaunay diagram of an embodiment of the present invention;

FIG. 8 is a schematic view of a two-dimensional cold leakage area according to an embodiment of the present invention;

FIG. 9 is a schematic diagram of a temperature leakage cold area mapping according to an embodiment of the present invention;

FIG. 10 is a schematic view of a wind tunnel cold leakage monitoring system according to an embodiment of the present invention;

fig. 11 is a schematic view of the installation of each set of two-camera according to the embodiment of the present invention.

Reference numbers and corresponding part names in the drawings:

1-wind tunnel.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to examples and accompanying drawings, and the exemplary embodiments and descriptions thereof are only used for explaining the present invention and are not meant to limit the present invention. In the description of the present application, it is to be understood that the terms "front", "back", "left", "right", "upper", "lower", "vertical", "horizontal", "high", "low", "inner", "outer", etc. indicate orientations or positional relationships based on those shown in the drawings, and are only for convenience in describing the present invention and simplifying the description, but do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed in a particular orientation, and be operated, and thus should not be construed as limiting the scope of the present application.

Example 1:

the wind tunnel cold leakage monitoring method based on the global normal distribution and the local outlier factor shown in fig. 1 specifically includes:

the method comprises the following steps of firstly, obtaining temperature measurement point data of a temperature measurement optical fiber paved on the outer surface of the wind tunnel.

The temperature measurement optical fiber can accurately measure the temperature, and the temperature data obtained by measuring the temperature measurement optical fiber is single-point. The point formed by the optical fiber temperature measurement is (X,Y,Z,T) WhereinX,Y,ZIs the coordinate of the temperature measuring point in the wind tunnel coordinate system,Tis the temperature.

Projecting the temperature measurement point toX,Y) The plane on which the optical fiber is located forms a grid point, and the grid point and the temperature value of the optical fiber form (X,Y,T) The three-dimensional scatter coordinates of (a) are shown in fig. 2:

and step two, detecting global abnormal points based on normal distribution.

Assuming the temperature point composition obtained by fiber measurementnDimension vector (a)T ₁,T ₂,…,T _n ). By calculating the mean and variance of this vector, a global mean is obtaineduSum variance informationδ：

；

Under the assumption of normal distribution, the measured temperature data is considered to be less than: (u-cδ) The temperature measurement point (where c is a constant) is an outlier, and the corresponding fiber temperature measurement point is designated as the global outlier, as shown by the hollow point in fig. 3.

Thirdly, based on local outlier factorLOFLocal outlier screening of

And (4) sending the global abnormal points obtained by the normal distribution-based global abnormal point detection algorithm to a local abnormal point screening module for screening abnormal points, and screening the abnormal points with drastic abnormal changes.

The Local Outlier Factor (LOF) is a measure of a certain pointpNeighborhood point ofN _k (p) local reachable density and pointpIs averaged over the locally achievable density ratio. The calculation formula is as follows:

；

in the formula (I), the compound is shown in the specification,pis the calculated global anomaly point;kis as followskA neighborhood;LOF _k (p) is a pointpIn the first placekLocal outlier factors within the neighborhood;N _k (p) is a pointpIn the first placekNeighborhood points within a neighborhood;sis a pointpIn the first placekSampling points in the neighborhood;lrd _k (s) is a dotsLocal achievable density of;lrd _k (p) is the local achievable density of point p.

If the above ratio is close to 1, it indicates thatpAnd its neighborhood point densityIn the same way as much as possible,ppossibly in the same cluster as the neighborhood; if the ratio is less than 1, this indicatespIs higher than its neighborhood point density,pis a density point; if the ratio is greater than 1, this indicatespIs less than its neighborhood point density,pthe more likely it is an outlier. From the above, LOF is determined by comparisonpJudging whether the point is an abnormal point or not by the density of the point and the neighborhood point, if so, judging whether the point is an abnormal point or notpThe lower the density of (a), the more likely it is to be identified as an anomaly. The farther the distance between the points is, the lower the density, and the closer the distance is, the higher the density. LOF is calculated to pass the pointkNeighborhood calculation, not global calculation, is a local anomaly detection method, and the LOF features extracted for each point are local features.

Fig. 4 is a result of performing local LOF extraction on the global outlier after the global outlier detection, and screening a point with a large LOF value as a cold leakage point. Through the judgment of the global and local anomaly detection algorithms, the finally generated stable cold leakage points are points represented by squares in fig. 4. It can be seen that the use of LOF features allows the points that vary strongly locally among the points after undergoing global anomaly screening to be screened out finally, i.e., stable cold leakage points.

In the working process of the embodiment, the global anomaly detection is carried out by the pairs (A and B)u-cδ) In (1)cThe value is set, and the LOF threshold value is set during local anomaly detection, so that the detection precision of cold leakage points can be effectively controlled, and the false alarm rate can be reduced.

Fourthly, cold leakage area is established based on triangulation

After the anomaly detection algorithm, stable cold leakage points are screened out and can be used as cold leakage supporting points to establish a cold leakage area. In the embodiment, a cold leakage area is established by adopting the idea in triangulation. A Voronoi (von rounity) graph in triangulation is used to reflect the influence of each point on its surrounding area, and a delaunay (delaunay) graph reflects the association between the nodes, i.e. the connection between cold leakage points.

Removing the temperature value from the three-dimensional coordinates formed by the optical fiber temperature measuring points, and performing a leakage cold anomaly detection algorithm, wherein grid points formed by the optical fiber temperature measuring points in an O-XY coordinate system are shown in FIG. 5, solid points in FIG. 5 represent common temperature measuring points, hollow points represent residual global anomaly points, and squares represent stable leakage cold points.

Voronoi (von roughy) map extraction is performed on the grid in fig. 5 to obtain Voronoi (von roughy) partitions as shown in fig. 6, and each temperature measuring point in fig. 6 is associated with a small area. In order to solve the correlation between each temperature measurement point, in this embodiment, the remaining global singular points are discarded, triangulation is performed, a delaunay (delaunay) graph is built, and delaunay (delaunay) partitioning as shown in fig. 7 is obtained, and the final cold leakage region, that is, as shown by a shaded portion in fig. 8, is formed by a Voronoi (von loney) region and a delaunay (delaunay) region at the stable cold leakage point.

Fig. 8 shows that the leakage cooling region of the present embodiment is formed by a Voronoi (von rounity) region of the fiber temperature measurement leakage cooling anomaly point and a delaunay (delaunay) region between the key points. Through the projection relation among the optical fiber points, the areas of the O-XY coordinate system corresponding to the O-XYZ coordinate system one by one are obtained, and the cold leakage areas under the O-XYZ coordinate system are obtained, wherein the projection process is shown in FIG. 9. Fig. 9 shows the situation that the temperature measuring points of the arranged optical fibers are uniform, and the algorithm of the present application is also similar to the situation of non-uniform arrangement, except that the shapes of the finally obtained leakage cooling areas are irregular.

The arrangement of the optical fibers influences the two-dimensional distribution between the temperature measuring points and influences the Voronoi (von Lonouy) and delaunay (Delaunay) subdivision results, and theoretically, the more the temperature measuring points are arranged, the better the distribution is, and the final cold leakage area identification result is more accurate. The cold leakage monitoring scheme is carried out on the basis of three-dimensional temperature measurement data, candidate possible abnormal points are detected through detection based on global normal distribution abnormal points, local abnormal temperature region characteristics are constructed through an LOF (low order frequency) method according to the layout of an actual heat insulation block and the arrangement of optical fibers, the candidate abnormal points are screened and identified, and therefore misinformation is achieved. The temperature measurement and monitoring scheme can be used for accurately identifying cold leakage of the tunnel body and reliably reducing false alarm.

Example 2:

in the embodiment, an X wind tunnel is taken as an example, and the method described in embodiment 1 is used in a wind tunnel cold leakage monitoring system; the X wind tunnel is a continuous closed-loop back-flow transonic wind tunnel driven by an axial flow compressor, and the power of the compressor is 60 MW. The wind tunnel is composed of a tunnel body main loop, an auxiliary system, a factory building and the like. The temperature range of the wind tunnel operation medium is-196 ℃ to 50 ℃, the inner layer of the tunnel body is a heat insulation layer, the outer layer of the tunnel body is a stainless steel shell, the tunnel body is insulated by the heat insulation layer, and the heat insulation layer is formed by splicing 5000 heat insulation blocks with the size of 1m multiplied by 1 m.

According to the design requirement of the heat insulation layer, the surface temperature of the heat insulation layer is not lower than the local dew point temperature under normal conditions. The failure of the thermal insulation layer can cause the temperature of the surface of the hole to drop. In the operation process of the wind tunnel, the state of the heat-insulating layer needs to be monitored so as to judge whether the heat-insulating layer fails or not. The wind tunnel faces the following problems in temperature measurement at present:

(1) the infrared temperature measurement precision is reduced due to the distance, the angle and the smoothness of the measured object. The infrared camera is affected by the distance of the measured object, the formed observation angle and even the surface smoothness and other factors to cause the precision to be reduced due to the measurement characteristics of the infrared camera. In the project, 18 dual-spectrum cameras are arranged in cooperation with a holder to perform preset position scanning type infrared temperature measurement, and when different positions of the wind tunnel with surface smoothness are determined by measurement, the temperature measurement precision can be influenced to a certain extent due to the difference of distance and angle, so that the failure of temperature measurement cold leakage monitoring is caused.

(2) The visual equipment and the optical fiber temperature measuring equipment are reasonably arranged. As the field environment of the project is complex, the wind tunnel of the measured object is large and irregular, the ladder platform is arranged on the periphery of the tunnel body, the measurement and control cabinet is arranged on the periphery of the tunnel body, and the infrared temperature measurement cameras are arranged on the walls around the wind tunnel, so that the ladder platform and the measurement and control cabinet can shield the infrared imaging, and the appearance of the measured object is shown in the following figure. Therefore, the surface of the cave body is difficult to accurately measure and completely covered by vision temperature measurement only, the auxiliary temperature measurement must be carried out by matching with an optical fiber temperature measurement system, and finally, the infrared thermometer is manually held for rechecking, so that the accurate temperature measurement and positioning are realized.

(3) Multi-source heterogeneous data fusion problem. The fusion of infrared image data and optical fiber temperature measurement data is needed to be realized in the project, the temperature measurement data is uniformly mapped to a wind tunnel 3D model, the 3D display effect is carried out after grid division is carried out according to a wind tunnel three-dimensional model, a wind tunnel surface temperature field is displayed in a heat map mode, and a cold leakage area is marked.

The visual data has higher resolution but relatively lower temperature measurement accuracy, the optical fiber temperature measurement data has lower resolution but relatively higher temperature measurement accuracy, and the output modes of the two data have larger difference, so that the fusion is more difficult; in addition, the number of system devices is large, the fusion data volume is large, and the fusion difficulty is high.

In order to solve the above problems, the wind tunnel cold leakage monitoring system based on vision and distributed optical fiber combined temperature measurement as shown in fig. 10 is provided in this embodiment, and the system mainly includes:

the vision temperature measurement subsystem is used for acquiring a temperature field image of the surface of the wind tunnel;

the optical fiber temperature measurement subsystem is used for measuring the temperature of the measurement blind area of the visual temperature measurement subsystem on the surface of the wind tunnel;

the temperature correction subsystem is used for correcting and evaluating temperature measurement data of the vision temperature measurement subsystem and the optical fiber temperature measurement subsystem;

and the 3D display subsystem is used for displaying the three-dimensional temperature field and the position of a cold leakage point on the surface of the wind tunnel in real time on the 3D model of the wind tunnel.

The working flow of the system is as follows:

a system preparation stage: firstly, laying software and hardware, and carrying out system connectivity test after laying is finished. And if the connectivity test is not problematic, starting to calibrate the system, including calibrating the camera and calibrating the temperature measuring optical fiber.

A data acquisition stage: the double-optical camera and the distributed optical fiber start to synchronously acquire data, and the data are respectively transmitted from the corresponding ten-gigabit network switch and the corresponding gigabit network switch through the high-speed optical fiber. Image data collected by the double-optical camera is transmitted into the NVR network video storage server through the switch, and is stored in the storage server, compressed, subjected to image preprocessing and the like and then transmitted into the data processing server. The optical fiber data is directly transmitted to the data processing server.

And (3) a data fusion stage: the data processing server fuses the visible light and the infrared data and maps the two-dimensional infrared image to the surface of the three-dimensional model. And then, carrying out cold leakage monitoring through a deep learning algorithm, and carrying out algorithm acceleration through a GPU. And the optical fiber data is used for positioning the surface of the three-dimensional hole body by a method of installing a cooperation mark on the temperature measurement optical fiber, and mapping the optical fiber temperature measurement data to the surface of the three-dimensional hole body.

A display stage: and fusing the visual temperature measurement data and the optical fiber temperature measurement data, directly mapping the monitoring result to the surface of the wind tunnel 3D model, warning and displaying the abnormal area, and giving the number of the heat insulation block at the position of the leakage point.

In one or more preferred embodiments, the visual thermometry subsystem is further configured to obtain an image of the surface texture of the cavity.

In one or more preferred embodiments, the vision temperature measurement subsystem comprises a plurality of groups of double-light cameras which are oppositely arranged outside the wind tunnel in pairs; the double light of the double-light camera is infrared light and visible light;

each double-light camera is provided with an effective temperature measuring area for measuring the infrared temperature of the surface of the wind tunnel, and the effective temperature measuring areas of each group of two opposite double-light cameras are partially overlapped;

the installation heights of all the double-optical cameras are higher than the top of the wind tunnel, and the measurement area of the vision temperature measurement subsystem at least covers the upper surface of the wind tunnel.

In one or more preferred embodiments, the effective temperature measurement area is calculated based on the johnson criterion; each dual-light camera has a matched pan-tilt.

In one or more preferred embodiments, the optical fiber temperature measurement subsystem comprises a temperature measurement optical fiber laid at the lower part of the wind tunnel in a rotating manner, and the temperature measurement optical fiber covers a measurement blind area of the vision temperature measurement subsystem on the surface of the wind tunnel; and the top convolution part of the temperature measuring optical fiber enters the effective temperature measuring area.

In one or more preferred embodiments, a plurality of optical fiber measurement key points are distributed on the temperature measurement optical fiber at equal intervals, and redundant segments are distributed on the temperature measurement optical fiber at equal intervals.

In one or more preferred embodiments, the temperature correction subsystem comprises:

the blackbody correction module comprises a blackbody which is arranged in the visual field range of the vision temperature measurement subsystem and is used for correcting the measurement temperature of the vision temperature measurement subsystem;

the temperature correction and evaluation module comprises a plurality of first temperature sensors arranged on the surface of the wind tunnel and a plurality of second temperature sensors arranged around the wind tunnel; the first temperature sensor is used for correcting and evaluating temperature measurement data of the vision temperature measurement subsystem and the optical fiber temperature measurement subsystem, and the second temperature sensor is used for monitoring the ambient temperature in the wind tunnel workshop;

and the handheld thermal infrared imager is used for carrying out secondary detection on the position of the cold leakage point.

In one or more preferred embodiments, the 3D presentation subsystem comprises:

the visual temperature measurement processing module is used for processing the temperature field data acquired by the visual temperature measurement subsystem;

the optical fiber temperature measurement processing module is used for processing temperature data measured by the optical fiber temperature measurement subsystem;

the fusion module is used for fusing temperature data of an overlapped measurement area of the visual temperature measurement subsystem and the optical fiber temperature measurement subsystem;

the mapping module is used for mapping the fused temperature data to the wind tunnel 3D model;

and the display module is used for displaying the surface temperature field of the wind tunnel in a heat map mode.

In one or more preferred embodiments, a method for fusing temperature data of overlapping measurement regions of a visual thermometry subsystem and a fiber optic thermometry subsystem comprises:

extracting optical fiber data points obtained by optical fiber temperature measurement;

and obtaining temperature data among the optical fiber data points by adopting a data interpolation algorithm, wherein the curved surface adopted by the data interpolation algorithm is obtained by fitting the measured data of the visual temperature measurement subsystem.

In one or more preferred embodiments, the method for mapping the fused temperature data onto the wind tunnel 3D model is a bicubic interpolation method, including:

setting the P point as the position of the target image corresponding to the source image at the position (X, Y); the coordinate of the point P is P (x + u, y + v), wherein x and y respectively represent an integer part, and u and v respectively represent a decimal part;

and (3) calculating:

；

in the formula (I), the compound is shown in the specification, f(x,y) The pixel value of the P point after interpolation; x _i ，y _i pixel coordinates in a range of 4 × 4 around the point P;i =0,1,2,3； f(x _i ，y _i ) Pixel values corresponding to pixel coordinates within a range of 4 × 4;Wis an interpolation coefficient, and the value is as follows:

whereina=-0.5。

Example 3:

on the basis of embodiment 2, this embodiment optimizes the visual thermometry subsystem:

according to the requirements of the wind tunnel plant in the embodiment, the imaging equipment in the vision temperature measurement subsystem is arranged on the side wall of the wind tunnel plant so as to realize reasonable layout and realize the observation of the upper surface of the tunnel body to the maximum extent.

The camera layout takes into account several factors: (1) arranging a camera to cover the upper surface of the hole body; (2) when the cameras are arranged, infrared temperature measurement errors caused by the reason of the illumination angle of the double-optical camera need to be considered, and a basis is provided for optical fiber arrangement; (3) because this embodiment is backward flow formula wind-tunnel, when the camera was laid, because lay around wind-tunnel factory building wall, need consider near external surface of cavity (by the wall) and far away internal external surface of cavity (do not lean on the wall), how to use the reasonable monitoring of camera.

The wind tunnel outer surface that this embodiment is about 200m with the girth divides into 18 regions, carries out the infrared temperature measurement of corresponding region by 18 two optical cameras respectively, and 18 two optical cameras are all installed in factory building wall eminence all around, highly divide into two sets ofly according to its area of being responsible for:

10 cameras with numbers of C1, C2, C3, C5, C8, C10, C11, C13, C16 and C18 in total;

② the cameras numbered C4, C6, C7, C9, C12, C14, C15 and C17 total 8.

Wherein, because this embodiment is backward flow formula wind-tunnel, is responsible for highly lower to the two optical cameras of external surface pipeline monitoring, and is responsible for highly higher to the two optical cameras of internal surface pipeline monitoring to two sets of cameras are collocated each other, both have certain public visual field, make public visual field not too big again, thereby the visual surface of cover pipeline as big as possible. In order to avoid that the sight line of the opposite camera is shielded by the section of pipeline, the installation position of the corresponding camera is higher than that of other cameras for the thicker section of pipeline in the backflow type wind tunnel.

In the scheme, all the dual-optical cameras perform sequential cruise observation from high altitude through rotation of the holder according to a preset point position planned in advance at an overlooking angle; each camera is responsible for a designated area, and a corresponding preset position is set for cruise observation.

Because the surface to be measured is a curved surface target, when the dual-optical camera measures the target surface at a non-normal view angle, that is, the target surface is not perpendicular to the optical axis of the camera, the number of effective pixels of the surface to be measured on the infrared sensor has an equivalent proportionality coefficient (wherein, the included angle between the tangential plane of the camera surface and the optical axis) and the calculation is performed based on the recognition standard of 6 pixels in the johnson criterion, and the tubular surface with the imaging angle between the optical axis of the camera and the surface to be measured of 20 degrees or more is obtained as the optimal effective area for infrared temperature measurement, that is, the effective temperature measurement area.

Based on the effective temperature measurement area, two opposite dual-optical cameras in each group are installed as shown in fig. 11, wherein the reference number 1 in fig. 11 is a wind tunnel, the AB section minor arc is the effective temperature measurement area of the camera C1, the CD section minor arc is the effective temperature measurement area of the camera C2, and the clampCornerαThe limit angle is 20 degrees corresponding to the effective temperature measuring area of the camera.

In order to achieve the best measurement effect, two groups of cameras are matched with each other, a certain public view field exists, the public view field is not too large, the effective measurement areas of the cameras on two sides of the visible surface of the pipeline are covered as large as possible, the edge of the effective measurement area is an overlapping area with the length of about 0.5m, namely the BC section minor arc is the overlapping area, and the circumferential length of the effective measurement area is 0.5 m.

It should be noted that, for the back-flow wind tunnel, it is also necessary to verify whether the camera on the peripheral wall is shielded by the pipeline on the same side when monitoring the temperature of the inner wall of the pipeline on the opposite side.

Example 4:

on the basis of embodiment 2, this embodiment optimizes the optical fiber temperature measurement subsystem:

the optical fiber temperature measurement subsystem in the embodiment is composed of a distributed optical fiber temperature measurement host (such as ATDTS-MK), a temperature measurement optical fiber (such as MGTSV-2A 1), a gigabit network switch and a data processing server (which is the same as a data server of the visual temperature measurement subsystem), and is responsible for measuring the surface area which can not be covered by the tunnel body visual subsystem, detecting the cold leakage and liquid leakage conditions on the surface of the tunnel body and positioning the liquid leakage position of the tunnel body.

The optical fiber temperature measurement subsystem comprises a temperature measurement optical fiber laid at the lower part of the wind tunnel in a rotating manner, and the temperature measurement optical fiber covers a measurement blind area of the vision temperature measurement subsystem on the surface of the wind tunnel; and the top convolution part of the temperature measuring optical fiber enters the effective temperature measuring area. A plurality of optical fiber measurement key points are distributed on the temperature measurement optical fiber at equal intervals, and redundant sections are distributed on the temperature measurement optical fiber at equal intervals.

In the embodiment, optical fiber measurement key points (1 m in redundancy) are reserved at intervals of 100 m and used as temperature measurement calibration points, and the optical fiber system is ensured to overcome the thermal deformation influence of the hole body. And an optical fiber rotary laying scheme is adopted, the optical fiber laying interval distance is 0.5m, and the optical fiber laying interval distance is consistent with the connecting position of the heat insulation block with the size of 1m square, so that the optical fiber can effectively cover the gap of the connecting position of the heat insulation block, the detection accuracy of a cold leakage area is improved, and the detection area accuracy reaches the standard. And in the surface section of the hole body with other objects such as ladders/platforms, the winding height of the optical fibers exceeds the height of the ladders/platforms by 0.5 m.

According to the laying principle, the temperature measuring optical fiber in the embodiment is 0.5m higher than the minor arc section AD on the surface of the hole body which can not be covered by the dual-optical camera in FIG. 11, namely the laying position and the height of the optical fiber are the rough sections in FIG. 11.

In this embodiment, the functions of the optical fiber temperature measurement subsystem include:

(1) the data acquisition and transmission function is realized, the data acquisition and transmission module is used for acquiring the original temperature measurement optical signal data of the temperature measurement optical fibers distributed on the surface of the hole body and transmitting the data to the DTS host machine through the optical cable; the processed temperature measurement electric data is transmitted to a data processing server by a kilomega network switch; wherein in order to guarantee the expandability of the system, the exchanger is adopted to transmit temperature measurement electric data, and a plurality of DTSs and corresponding temperature measurement optical fibers can be accessed at any time.

(2) And the data storage and processing function is realized by the data storage and processing module in the DTS host to process the acquired temperature measurement optical signal data, obtain the temperature measurement electric signal and transmit the temperature measurement electric signal to the data processing server for storage.

(3) The optical fiber testing system has a calibration function, and the calibration of the length and the temperature measurement distance on the temperature measurement optical fiber is realized by the optical fiber testing system calibration module.

(4) And the measurement result and the wind tunnel 3D model are associated, and the calibration and association of the temperature measured by the optical fiber and the position of the wind tunnel model are realized on the data processing server by the optical fiber temperature measurement three-dimensional positioning module.

(5) And a cold leakage monitoring function, which is realized by the detection and positioning of the abnormal part of the optical fiber temperature in the DTS host after the temperature measurement electric signal is obtained by data processing of the cold leakage monitoring module, and the detection and positioning of the abnormal part of the surface temperature of the three-dimensional hole body on the data processing server, and is realized based on the method of the embodiment 1.

Example 5:

on the basis of embodiment 2, this embodiment optimizes the temperature correction subsystem:

the temperature correction subsystem is mainly used for correcting and evaluating temperature measurement data of the vision temperature measurement subsystem and the optical fiber temperature measurement subsystem and comprises a blackbody correction module, a temperature correction evaluation module and a handheld thermal infrared imager.

The blackbody correction module is mainly used for continuously correcting the infrared temperature measurement camera so as to ensure the temperature measurement precision. The black body is a constant temperature target and is used for continuously correcting the infrared temperature measurement of the double-camera so as to ensure the temperature measurement precision, and the black body is arranged in the visual field of the double-camera and is used for measuring the temperature of the black body and carrying out temperature measurement calibration in real time by taking the temperature measurement as a reference. This project adopts portable blackbody correction module, can place and install on the tripod. When the temperature is corrected, the black body is arranged at a fixed position, and after the camera is corrected, the black body is arranged at the calibration position of the next camera to correct the next double-camera.

The temperature correction and evaluation module mainly comprises a first temperature sensor arranged on the surface of a wind tunnel shell, 4 positions are arranged, online real-time monitoring is carried out on the surface temperature of the tunnel body and used for correcting and evaluating test data of a visual temperature measurement and optical fiber temperature measurement system, in addition, 2 positions of second temperature sensors are arranged on the periphery of the wind tunnel and used for monitoring the ambient temperature in real time and used as a local reference temperature when a cold leakage detection algorithm is constructed.

Example 6:

on the basis of embodiment 2, this embodiment optimizes the 3D display subsystem:

and the 3D display subsystem is mainly used for displaying a cold leakage detection result and displaying a three-dimensional temperature field and a cold leakage point position on the wind tunnel 3D model in real time.

The 3D display subsystem, the vision temperature measurement subsystem and the optical fiber temperature measurement subsystem share one data processing server to perform fusion processing on vision temperature measurement and optical fiber temperature measurement data; and uniformly mapping the temperature measurement data to a wind tunnel 3D model, carrying out data fusion on the overlapped measurement area, carrying out 3D display after carrying out grid division according to the wind tunnel three-dimensional model, displaying the wind tunnel surface temperature field in a heat map form, and marking a cold leakage area.

The 3D display subsystem fuses data of visual temperature measurement and optical fiber temperature measurement by adopting a data fusion technology. The fusion belongs to decision-level fusion, preliminary conclusions are obtained on the wind tunnel surface temperature information through an infrared camera and a distributed optical fiber respectively, then decision-level fusion judgment is carried out through correlation processing, and finally fusion type wind tunnel surface temperature information is obtained. And meanwhile, fusing and mapping the fused temperature image to a wind tunnel 3D model by adopting a data interpolation algorithm, so as to display the wind tunnel surface temperature field in a heat map mode.

The functions of the 3D display subsystem mainly comprise fusion of infrared temperature measurement and optical fiber temperature measurement data in an overlapping part and mapping of a three-dimensional temperature field.

Fusion of infrared temperature measurement and optical fiber temperature measurement data in an overlapping part:

in the blind area of infrared temperature measurement, the infrared camera can not shoot and obtain data. Therefore, the temperature measurement is performed by using the optical fiber. At the edge of the fiber layout, the infrared camera and the fiber region have a partial overlapping region, so data fusion is required.

The infrared camera can shoot temperature changes of continuous space, and measurement accuracy is low compared with an optical fiber. The optical fiber can measure the temperature accurately, but the limitation of the optical fiber layout causes the optical fiber to measure data of a single point.

The point value is the temperature measured by the optical fiber, and the curved surface is the temperature field measured by the infrared camera. Therefore, the temperature measured by the infrared camera has higher resolution, and the data map is smoother; and the temperature measured by the optical fiber is more accurate. Therefore, the optical fiber data are adopted at the position with the optical fiber data, and the point data are obtained at the points among the optical fiber data by adopting an interpolation method. The curved surface used for interpolation is obtained by data fitting of an infrared camera. The data of the fitted curved surface corresponding to the optical fiber is the temperature measured by the optical fiber, and the variation trend of the data is the same as that of the data of the infrared camera; equivalent to the optical fiber data providing accurate temperature values, the infrared camera provides a trend of temperature changes.

This embodiment is through effective, reasonable deployment infrared temperature measurement and optic fibre temperature measurement equipment, has guaranteed that the at utmost covers the hole body surface, only by sheltered from the department and the unable detection of hole body the latter half, has both ensured seamless no missing errorless cold leakage monitoring, reduces the redundancy of system simultaneously, reduces the waste of system computing resource.

(II) mapping of three-dimensional temperature field:

in this embodiment, the infrared image is required to be mapped onto the wind tunnel three-dimensional model, and since the resolution of the infrared image is only 640 × 512, the number of pixels inside a patch of the wind tunnel three-dimensional model is much higher than the resolution of the infrared image. Therefore, interpolation is required in the wind tunnel three-dimensional model patch to map the temperature field. Meanwhile, the interval of the optical fibers distributed at the bottom of the wind tunnel is 0.4m, and the grid size of the 3D model is 0.25m multiplied by 0.25 m. Therefore, data interpolation is also required at the time of model display.

The project adopts a bicubic interpolation method to perform data interpolation, and the obtained texture transition is more uniform, the sawtooth effect is minimum, and the image effect is best.

The main principle of bicubic interpolation is as follows:

assume that the source image a is M × N in size and the scaled target image B is M × N in size. Then the corresponding coordinate of B (X, Y) on a can be found from the scale item as:

A(x，y)=A(X×(n/M)，Y×(n/M))；

in bilinear interpolation, the item takes the nearest four points of A (x, y). In the bicubic interpolation method, the nearest 16 pixel points are selected as parameters for calculating the pixel values of the target image B (X, Y).

Let P be the position of the target image B in (X, Y) corresponding to the source image A, the coordinate position of P will appear in fractional part, so the project assumes that the coordinate of P is P (X + u, Y + v), where X, Y respectively represent integer parts and u, v respectively represent fractional parts. Then the position of the nearest 16 pixels can be obtained, using a_ij(i, j =0,1,2, 3).

The purpose of bicubic interpolation is to find out an influence factor of the 16 pixels on the pixel value at the P position by finding out a coefficient, so that the pixel value of the corresponding point of the target image is obtained according to the influence factor, and the purpose of image scaling is achieved. The calculation formula is as follows:

wherein the content of the first and second substances,f(x,y) Is the pixel value of the interpolated P point,x _i ，y _i pixel coordinates in a range of 4 x 4 around the P point,f(x _i ，y _i ) Is a pixel value corresponding to a pixel coordinate within a range of 4 × 4. It can be seen that the bicubic interpolation is weighted according to the 16 surrounding pixels.

Wherein the interpolation coefficientWThe calculation method of (2) is as follows:

wherein the content of the first and second substances,a=-0.5。

in the embodiment, mapping is performed by a bicubic interpolation method, finally, 3D display is performed after grid division is performed according to a wind tunnel three-dimensional model, a wind tunnel surface temperature field is displayed in a heat map mode, and cold leakage areas are marked. According to the camera layout scheme, the minimum pixel number of a target with the size of 0.25m in the camera is 15.2 pixels, the target number is far larger than that of the target with the size of 6 pixels, and the resolution of a threshold wind tunnel 3D display grid is enabled to reach 0.25m multiplied by 0.25m through three-dimensional temperature field mapping.

Example 7:

on the basis of any of the above embodiments, in order to accurately obtain the relationship between the measured temperature of the optical fiber and the position of the wind tunnel, the embodiment further implements the following optical fiber fast positioning method:

calibrating the length and the temperature measurement distance of the temperature measurement optical fiber: a plurality of optical fiber length calibration points are arranged on the temperature measuring optical fiber at equal intervals, and a calibration section is arranged at each optical fiber length calibration point; sequentially placing each calibration section in an environment with obvious temperature difference with room temperature, acquiring optical fiber temperature measurement data, and determining the position of the calibration section; collecting the positions of all calibration sections to finish the calibration of the length of the optical fiber and the temperature measurement distance;

temperature correction is carried out on the temperature measurement optical fiber: the method comprises the steps of carrying out temperature measurement consistency calibration, sensitivity calibration, temperature deviation calibration and position calibration in sequence;

installing a temperature measuring optical fiber in a set area on the surface of the wind tunnel;

positioning the position of the temperature measuring optical fiber on the surface of the three-dimensional wind tunnel model: setting a binocular vision system in the wind tunnel plant and calibrating a binocular camera; installing a cooperation mark; detecting each cooperation mark through a binocular camera; calculating the three-dimensional space coordinates of each cooperation mark; and converting the three-dimensional space coordinate into a wind tunnel model coordinate system.

In one or more preferred embodiments, the method for calibrating the temperature measurement consistency includes: placing the whole temperature measuring optical fiber in a constant temperature space, standing for at least 1 hour, obtaining temperature measuring data of the whole temperature measuring optical fiber, and correcting to be consistent;

in one or more preferred embodiments, the method of sensitivity calibration includes: placing the local part of the temperature measuring optical fiber in a constant temperature space, and standing for at least 5 minutes; measuring the internal temperature of the constant-temperature space and the ambient temperature outside the constant-temperature space, and acquiring temperature measurement data of the whole temperature measurement optical fiber; if the environment temperature is consistent with the temperature measurement data of the temperature measurement optical fiber positioned outside the constant temperature space, comparing the internal temperature of the constant temperature space with the temperature measurement data of the optical fiber section positioned in the constant temperature space, and correcting the temperature measurement data of the optical fiber section to be consistent with the internal temperature of the constant temperature space;

in one or more preferred embodiments, the method of temperature offset calibration includes: placing the local part of the temperature measuring optical fiber in a constant temperature space, and standing for at least 5 minutes; measuring the internal temperature of the constant-temperature space and the ambient temperature outside the constant-temperature space, and acquiring temperature measurement data of the whole temperature measurement optical fiber; if the environment temperature is different from the temperature measurement data of the temperature measurement optical fiber outside the constant temperature space, the temperature measurement data of the temperature measurement optical fiber is integrally corrected;

in one or more preferred embodiments, the method of position calibration includes: the length position and the temperature measurement distance are calibrated for any one or more point positions of the temperature measurement optical fiber: the method comprises the following steps of area boundary points of the surface of the wind tunnel, a measurement starting point of the temperature measurement optical fiber, a measurement finishing point of the temperature measurement optical fiber and a key focus point.

In one or more preferred embodiments, the cooperative identifier includes a global identifier point, a local identifier point; the global identification point is pasted on the surface of the wind tunnel, and the local identification point is pasted on the temperature measurement optical fiber; the local identification point is provided with a unique code.

In one or more preferred embodiments, the method of calculating three-dimensional space coordinates of a cooperative identifier includes:

defining a binocular camera as a left camera and a right camera;

calculating a projection matrix for a left cameraM _l Comprises the following steps:

in the formula (I), the compound is shown in the specification,K _l is an internal reference matrix of the left camera,R _l is the identity matrix of the left camera and,T _l is a zero matrix of the left camera and,mis a projection matrixM _l The calculation result of (2);

computing a projection matrix for a right cameraM _l ’Comprises the following steps:

in the formula (I), the compound is shown in the specification,K _l ’is an internal reference matrix of the right camera,R _l ’is the identity matrix of the right camera and,T _l ’is a zero matrix of the right camera and,m’is a projection matrixM _l ’The calculation result of (2);

two cameras synchronously shoot target points in spacePSetting a pointPThe coordinates in the left camera coordinate system are: (X,Y,Z) The undistorted imaging point coordinate in the left camera image is (x _l ,y _l ) (ii) a DotPThe coordinates in the right camera coordinate system are (X’,Y’, Z’)，The undistorted imaging point coordinate in the right camera image is (x _l ’,y _l ’)；

Obtaining the following overdetermined linear equation set according to the projection matrix and the projection principle:

solving the points according to the above equation setPThree-dimensional space coordinates of (X,Y,Z)。

The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are merely exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.

It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Further, the term "connected" used herein may be directly connected or indirectly connected via other components without being particularly described.

Claims (9)

1. A wind tunnel cold leakage monitoring method based on global normal distribution and local outlier factors is characterized by comprising the following steps:

acquiring temperature measuring point data of a temperature measuring optical fiber paved on the outer surface of the wind tunnel;

performing global anomaly detection on all temperature measuring points based on normal distribution to obtain global anomaly points;

local anomaly screening is carried out on all global anomaly points based on the local outlier factors, and the screened local anomaly points are used as stable cold leakage points;

and (4) establishing a cold leakage area by taking all stable cold leakage points as supports.

2. The wind tunnel cold leakage monitoring method based on the global normal distribution and the local outlier factor according to claim 1, wherein the temperature point data comprises coordinates and temperature.

3. The wind tunnel cold leakage monitoring method based on the global normal distribution and the local outlier factor according to claim 1, wherein the temperature measurement point data is a three-dimensional scattered point coordinate system, and the method for establishing the three-dimensional scattered point coordinate system comprises the following steps:

acquiring basic data of temperature measuring pointsX,Y,Z,T) WhereinX,Y,ZIs the three-axis coordinate of the temperature measuring point in the wind tunnel coordinate system,Tthe temperature measured by the temperature measuring point;

projecting the temperature measurement point toX,Y) Forming a grid point on the plane;

constructing a temperature value from the grid pointsTThe three-dimensional scatter coordinate system is formed.

4. The wind tunnel cold leakage monitoring method based on the global normal distribution and the local outlier factor according to claim 1, wherein the method for detecting global anomaly of all temperature measurement points comprises the following steps:

temperature point composition obtained by measuring with temperature measuring optical fibernDimension vector (a)T ₁,T ₂,…,T _n )；

Calculate thenMean of dimensional vectorsuVariance, varianceδ；

Measured temperature data less thanu-cδ) The temperature measuring point is used as a global abnormal point; whereincIs a constant.

5. The wind tunnel cold leakage monitoring method based on the global normal distribution and the local outlier factor according to claim 1, wherein the method for screening local anomalies of all global anomaly points comprises the following steps: extracting local outlier factors for global outliersLOFScreening ofLOFThe point where the value is greater than the set threshold is taken as the stable cold leakage point.

6. The wind tunnel cold leakage monitoring method based on the global normal distribution and the local outlier factor according to claim 5, wherein the local outlier factor of the global outlierLOFCalculated by the following formula:

；

in the formula (I), the compound is shown in the specification,pis the calculated global anomaly point;kis as followskA neighborhood;LOF _k (p) is a pointpIn the first placekLocal outlier factors within the neighborhood;N _k (p) is a pointpIn the first placekNeighborhood points within a neighborhood;sis a pointpIn the first placekSampling points in the neighborhood;lrd _k (s) is a dotsLocal achievable density of;lrd _k (p) is a pointpLocal achievable density.

7. The wind tunnel cold leakage monitoring method based on the global normal distribution and the local outlier factor according to claim 1, wherein the method for establishing the cold leakage area comprises the following steps:

eliminating temperature values in the temperature measurement point data, and forming two-dimensional grid points under an O-XY two-dimensional coordinate system;

solving for a von neumonian map for the two-dimensional grid points;

solving the association between the cooling leakage points in the Von Rogoyi graph to obtain a two-dimensional cooling leakage area;

and mapping the two-dimensional cold leakage area from an O-XY two-dimensional coordinate system to an O-XYZ three-dimensional coordinate system.

8. The wind tunnel cold leakage monitoring method based on the global normal distribution and the local outlier factor according to claim 7, wherein the method for solving the association between the cold leakage points in the von neumoniae graph comprises the following steps:

discarding the rest global abnormal points in the two-dimensional grid points, reserving stable cold leakage points, carrying out triangulation, and establishing a Delaunay graph; the cooling leakage region is defined as an aggregate of a von roughy region in a von roughy map and a delaunay region in a delaunay map.

9. The wind tunnel cold leakage monitoring method based on global normal distribution and local outlier factor of claim 8, wherein the triangulation method is delaunay triangulation algorithm.

Images