KR20190004173A - Apparatus and method for monitoring of pantograph in electric railway - Google Patents

Abstract

Provided are an apparatus and a method for monitoring a pantograph of an electric railway vehicle. According to the present invention, a pantograph image obtained by capturing a pantograph is collected, and is compared to a preset pantograph template image to detect a pantograph object for each image frame, and a tilt angle and displacement of a current collecting plate is measured from the detected pantograph objects to detect preset vibration characteristics. Also, the detected vibration characteristics are classified through a classifier which machine-learns the detected vibration characteristics of the pantograph image, and vibration characteristics are classified through the classifier which machine-learns the vibration characteristics of the pantograph image detected before identifying normal and abnormal states of a pantograph, thereby identifying the normal and abnormal states of the pantograph. Also, a pantograph search area of a next image frame is set based on a pantograph object detection result of a previous image frame when the pantograph object is detected.

Description

TECHNICAL FIELD [0001] The present invention relates to an apparatus and method for monitoring a pantograph of an electric railway vehicle,

The present invention relates to an apparatus and method for monitoring the pantograph vibration condition of an electric railway vehicle through image processing.

The pantograph is in contact with the electric wire supplying electric energy to supply electric energy to the electric railway vehicle. Such a pantograph is mechanically contacted with a catenary during operation, causing abrasion and vibration, and in the non-contact state, an arc phenomenon occurs in which electric energy is discharged.

As described above, abnormal pantograph vibration occurs due to various causes such as overspeed of an electric railway vehicle, line abnormality, arc generation, etc. Such pantograph abnormal vibration gives fatigue to a catenary line, causing damage to a catenary line and related parts.

As a method of detecting abnormal vibration, there is a method of inspecting the vibration of the pantograph through a sensor after inserting the pantograph in a state where a sensor for monitoring vibration such as an accelerometer is attached to the pantograph. However, since the pantograph transmits electric energy to the vehicle, high voltage and high current are always supplied, which makes it difficult to attach a sensor for vibration detection.

In addition, there is a method of monitoring the point of contact between the pantograph and the catenary by installing the video on the upper part of the vehicle so that the contact point between the pantograph and the catenary line can be monitored and visually confirming the contact point. However, since it is discriminated by the naked eye, it is difficult to derive the accuracy and the quantitative value.

Korean Patent No. 1058179 (entitled " Pantograph defect monitoring system ") discloses a method of comparing a current image of a pantograph, which has been obtained by removing noise of an image obtained from a camera, with a reference image to determine whether or not the defect is defective. And the like. Japanese Patent No. 5534058 (entitled " abrasion measuring apparatus and method thereof ") discloses an abrasion measuring apparatus and a method thereof that can obtain abrasion measurement values of a cadet line in consideration of the inclination angle of a pantograph.

SUMMARY OF THE INVENTION The present invention is directed to provide a pantograph monitoring apparatus and a method for monitoring the pantograph vibration characteristics of an electric railway vehicle using the results of previous image processing.

It is to be understood, however, that the technical scope of the present invention is not limited to the above-described technical problems, and other technical problems may be present.

According to an aspect of the present invention, there is provided an apparatus for monitoring an electric railway vehicle pantograph, including: a communication module for receiving a pantograph image of a pantograph; a memory for storing a pantograph monitoring program; And a processor for executing a program stored in the memory, wherein the processor compares the input pantograph image with a preset pantograph template image according to the execution of the pantograph monitoring program to detect a pantograph object for each image frame, Detecting the predetermined vibration characteristics by measuring inclination angles and displacements of the current collecting plate from the detected pantograph objects, classifying the detected vibration characteristics through a classifier that has previously learned the vibration characteristics of the pantograph image detected, Identify steady state and abnormal state. At this time, when the pantograph object is detected, the processor sets the pantograph search area of the next image frame based on the pantograph object detection result of the previous image frame.

According to another aspect of the present invention, there is provided a method of monitoring an electric railway vehicle pantograph, the method comprising: collecting a pantograph image of a pantograph; Comparing the pantograph image with a predetermined pantograph template image to detect a pantograph object for each image frame; Detecting predetermined vibration characteristics by measuring a tilt angle and a displacement of the current collecting plate from the detected pantograph objects; And a classifier that classifies the detected vibration characteristics through a classifier that has previously learned vibration characteristics of the pantograph image, and classifies the vibration characteristics of the pantograph image detected before identifying the steady state and the abnormal state of the pantograph And classifying the detected vibration characteristics to identify the steady state and the abnormal state of the pantograph. Upon detection of the pantograph object, a pantograph search area of the next image frame is set based on the pantograph object detection result of the previous image frame.

According to any one of the above-mentioned objects of the present invention, the pantograph state (i.e., vibration characteristic) of the electric railway vehicle can be detected and continuously monitored through image processing. In particular, when the pantograph image is extracted, It is possible to improve the accuracy of real-time monitoring by improving the calculation performance.

Further, according to any one of the tasks of the present invention, it is possible to automatically classify the normal state and the abnormal state of the pantograph by processing the machine learning based on the frequency characteristic of the pantograph image detected in real time, It is possible to accurately and quickly check the condition such as the current plate abnormality, the line abnormality, or the abnormality due to the change in the tension of the line of the electric line.

1 is a configuration diagram of an electric railway vehicle pantograph monitoring apparatus according to an embodiment of the present invention.
FIG. 2 is a flowchart for explaining a template matching process by reducing a search area according to an embodiment of the present invention. Referring to FIG.
3 is a diagram for explaining a pantograph object detection process according to an embodiment of the present invention.
4 is a conceptual diagram for explaining a pantograph vibration characteristic detection process according to an embodiment of the present invention.
5 is a view for explaining a method of measuring an angle between a horizontal line and a current collecting plate in a pantograph according to an embodiment of the present invention.
6 is a diagram for explaining a method of calculating a vibration period of a pantograph based on an angle according to an embodiment of the present invention.
FIGS. 7 and 8 are views for explaining a method of calculating the vibration strength of a pantograph according to an embodiment of the present invention.
9 is a view for explaining a method of measuring left and right displacements of a pantograph according to an embodiment of the present invention.
10 is a view for explaining a method of calculating a vibration frequency based on an average value of left and right displacements of a pantograph according to an embodiment of the present invention.
FIG. 11 is a diagram for explaining an abnormal situation recognition method for a vibrating circle of a pantograph contact surface according to an embodiment of the present invention.
12 is an exemplary view for explaining an optimal hyperplane according to the magnitude of SVM margin applied to an embodiment of the present invention.
13 is a flowchart illustrating a pantograph monitoring method according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, which will be readily apparent to those skilled in the art. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly illustrate the present invention, parts not related to the description are omitted, and similar parts are denoted by like reference characters throughout the specification.

Throughout the specification, when a part is referred to as being "connected" to another part, it includes not only "directly connected" but also "electrically connected" with another part in between . Also, when an element is referred to as "including" an element, it is to be understood that the element may include other elements as well as other elements, And does not preclude the presence or addition of one or more other features, integers, steps, operations, components, parts, or combinations thereof.

1 is a configuration diagram of an electric railway vehicle pantograph monitoring apparatus according to an embodiment of the present invention.

1, an electric railway vehicle pantograph monitoring apparatus 10 includes a communication module 100, a memory 200, and a processor 300. [

The communication module 100 performs data communication with the camera device for photographing the pantograph, and receives the pantograph photographed image in real time from the camera device. For reference, the electric railway vehicle pantograph monitoring apparatus 10 according to an embodiment of the present invention may include a camera apparatus itself.

Such a camera device can be installed at an upper position of an electric railway vehicle capable of securing an image of a pantagraph and an electric line during operation of the electric railway vehicle. The camera device may be a high-speed or general camera device, and the higher the resolution, the more precise the pantograph can be detected. The electric railway vehicle pantograph monitoring apparatus 10 according to an embodiment of the present invention can perform high speed image processing even for a high resolution pantograph image so that the numerical change of the contact surface between the pantograph and the electric line can be monitored in real time .

The memory 200 stores a pantograph monitoring program.

The memory 220 is collectively referred to as a flash memory, a nonvolatile storage such as an SSD, a DRAM, or an SRAM volatile storage device that requires power to maintain stored information even when power is not supplied.

Also, the memory 200 may include a storage unit (not shown) for storing the real-time pantograph image collected through the communication module 100.

The processor 300 executes a program stored in the memory 200, executes a pantograph monitoring program, and performs the following processing.

The processor 300 processes the pantograph object detection and the vibration characteristic detection through matching with the template image set in advance for the input image sequence (i.e., the pantograph image). The processor 300 analyzes the characteristic information of the vibration of the contact surface between the pantograph and the electric cable to perform post processing for distinguishing the normal state from the abnormal state.

First, the process of the processor 300 detecting a pantograph object in the pantograph image will be described in detail with reference to FIGS. 2 and 3. FIG.

FIG. 2 is a flowchart for explaining a template matching process by reducing a search area according to an embodiment of the present invention. Referring to FIG. 3 is a view for explaining a process of detecting a pantograph object according to an embodiment of the present invention.

As shown in FIG. 2, the processor 300 selects a template image having the same brightness or color information as the input image among a plurality of template images as a reference image (S210).

For reference, the processor 300 learns a plurality of learning images (a labeled moving picture and a KP (Knowledge Piece) file) to process the pantograph image before processing the photographed pantograph image in real time, Can be generated and stored.

The pantograph template image is a reference image for accurately detecting the pantograph position from the pantograph image in which the brightness of the image may be different according to the intensity of light, the exposure value of the camera lens, the focus, etc., and a plurality of dictionaries Means a designated image. In FIG. 3, a plurality of template images, for example, an in-tunnel image, a day image, and a night image are shown as an example.

For example, the processor 300 may process the texture feature extraction of a training image (i.e., a pre-designated image) and then perform a machine learning (e.g., SVM training) to generate a template classifier. The processor 300 processes texture feature extraction for an input image sequence and then selects a reference image among the template images through a template classifier (e.g., SVM classification).

3, the processor 300 matches (i.e., template-matches) the pantograph template image having the highest similarity among the pantograph template images (i.e., the set reference image) to the pantograph image, Detects an object. That is, by matching the pantograph template image having the highest similarity among the plurality of pantograph template images to the pantograph image in which the brightness of the image is variously photographed according to the environment such as the weather, the tunnel situation, and the obstacle state, the pantograph can be accurately detected can do.

At this time, the change in the acquired images is not large except for the case where the brightness change such as entering / advancing tunnel or arcing occurs rapidly during acquisition of the pantograph image. Accordingly, the processor 300 detects a pantograph object by processing a search range reduction in a pantograph image.

Referring back to FIG. 2, the processor 300 divides the collected real-time pantograph image into frames (S220). For example, if the collected pantograph image is a moving image, an image frame is extracted.

The processor 300 sets the search area within the image frame, and sets the search area by referring to the position of the pantograph object extracted from the image of the previous image frame. That is, since the possibility that the position of the pantograph is largely changed due to the continuity of the image is low, it is possible to search for the pantograph object in the next frame by applying the position of the pantograph object searched in the previous frame.

At this time, the processor 300 determines whether the pantograph position information is known (S230).

If it is determined in step S230 that the pantograph position information is not known, the entire area of the image frame is set as the search area in step S240.

That is, in the initial stage, since there is no pantograph object detected in the previous frame, the pantograph object is detected by setting the entire area of the image frame as the search area. In this case, at this time, the processor 300 stores positional information of the detected pantograph object (e.g., center point coordinates of the pantograph, coordinates of left and right upper corner points, and the like). Accordingly, when the search area is set in the next frame, the pantograph object can be searched by applying the position of the pantograph object detected in the previous frame, thereby greatly reducing the resources required for searching.

If it is determined in step S230 that the pantograph position information is known, it is determined whether the position of the pantograph is located within a predetermined search area in step S250.

If it is determined in step S250 that the position of the pantograph object detected in the previous frame is out of the specified search area, it is determined that there is an error in the position of the pantograph object and the search area is reset to the entire area in step S260.

If it is determined in step S250 that the position of the pantograph object detected in the previous frame is included in the specified search area, the previous search area is applied to the image frame in step S270.

Then, an environment identification check (e.g., color and brightness check) is performed on the input image frame in which the search area is set through either step S240, S260, or S270 (S280).

If the difference between the color information or the brightness information of the input image frame and the color or brightness of the preset template reference image is equal to or greater than the threshold value as a result of the environmental recognition test, the process returns to step S210 to change the template reference image .

If the input image frame and the template reference image are similar in environment (color or lightness) (i.e., the difference is less than the threshold) as a result of the environment identification test, the input image frame and the template reference image are subjected to template matching, The pantograph object is detected (S290).

Specifically, the processor 300 compares the pixel information of the image on an area-by-area basis, and detects a point that is most similar to the template as a pantograph object. The objective function is used to determine the pantograph detection position. The objective function can be expressed by Equation (1) below.

&Quot; (1) "

In Equation 1, I denotes an input image (i.e., a pantograph image), and T denotes a template image (i.e., a pantograph template image). W denotes a linear transformation performed to compare an input image with a template image, and denotes a matrix transformation operation. Also, x is a coordinate in the template image, and p denotes linear transformation parameters.

Next, with reference to FIG. 4 to FIG. 10, a process of the processor 300 detecting the pantograph vibration characteristic based on the pantograph object detected from the input image will be described in detail.

4 is a conceptual diagram for explaining a pantograph vibration characteristic detection process according to an embodiment of the present invention.

4, the processor 300 detects the position of the pantograph (i.e., the pantograph object) in the input image (S410), and then processes the detection of the center point of the collecting plate in the detected pantograph object (S420).

The processor 300 detects a contact strip (that is, a collecting plate) between the pantograph and the electric line (S430), measures the angle between the horizontal line extending from the center of the collecting plate of the pantograph and the extension line of the collecting plate, The inclination angle tilt of the portion is measured (S440).

Next, the processor 300 calculates the vibration characteristics of the pantograph based on the angles calculated for each input image frame. At this time, the processor 300 performs frequency measurement (S450) through zero crossing based on the contact portion slope, and measures the vibration period feature based thereon. In addition, the processor 300 performs gradient strength measurement (S470) through the pantograph displacement measurement based on the contact portion slope (S460), and based thereon, measures the vibration amplitude feature. In addition, the processor 300 performs Fourier transform (S480) on the result of the tilt intensity measurement (S470) to measure the vibration frequency feature.

5 and 6, the process of the processor 300 for measuring the oscillation period characteristic of the pantograph will be described in detail.

5 is a view for explaining a method of measuring an angle between a horizontal line and a current collecting plate in a pantograph according to an embodiment of the present invention. And FIG. 6 is a diagram for explaining a method of calculating a vibration period of a pantograph based on an angle according to an embodiment of the present invention.

The pantograph object 310 is detected from the input image frame as shown in Fig. 5 (a), and the center point 325 of the pantograph collecting plate is detected as shown in Fig. 5 (b).

Specifically, the processor 300 may perform Hough transform for each input image frame to detect a straight line 320 corresponding to the current collecting plate. At this time, the processor 300 can detect the straight line corresponding to the current collecting plate by estimating the straight line at the upper end of the pantograph object 310 obtained through the Hough transform to the current collecting plate. 5C, the processor 300 detects the center line 325 of the current collector plate and the horizontal line 330 of the current collector plate extending from the center point 325 of the current collector plate in the detected current collector plate, (320) can be detected.

The processor 300 then measures the angle between the horizontal line 330 extending from the center point 325 and the extension line 320 of the current collector plate while measuring the angle between the horizontal line 330 and the extension line 320 intersecting the center point 325 Is measured. The angle between the horizontal line 330 and the extension line 320 can be calculated for each frame of the photographed image.

The processor 300 calculates the vibration period of the pantograph based on the angles calculated for each image frame.

Referring to FIG. 6, the processor 300 may calculate the oscillation period based on the number of pantograph frames at which the measured angle is zero-crossed divided by the total number of pantograph frames (frames per second).

Specifically, the processor 300 arranges the angles between the horizontal line 330 measured in each image frame of the input image and the extension line 320 of the current collector plate in the order of the image frame numbers (i.e., in time order) Lt; RTI ID = 0.0 > zero crossing. ≪ / RTI > Then, the processor 300 calculates the number of zero-crossing frames in which the point where the angle is zero-crossed is detected from the total number of frames (frames per second) of the image photographed for one second. Then, the processor 300 calculates the frequency by dividing the calculated number of zero-crossing frames by 2 divided by the number of frames per second. As described above, the processor 300 can measure the oscillation period, which means that the tilting of the contact surface repeats once up and down, by dividing the total period by the frame rate, which is the image acquisition frequency, in units of seconds.

Next, with reference to FIGS. 7 and 8, a process of measuring the vibration intensity characteristic of the pantograph will be described in detail.

FIGS. 7 and 8 are views for explaining a method of calculating the vibration strength of a pantograph according to an embodiment of the present invention.

The processor 300 can calculate the vibration strength of the pantograph by measuring the pantograph displacement based on the magnitude of the angle between the horizontal line extending from the center point of the collecting plate of the pantograph and the extension line of the collecting plate.

As shown in FIG. 7, the processor 300 classifies the amount of change (i.e., displacement) of the magnitude of the angle between the horizontal line 330 and the extension line 320 measured in each image frame of the input image, So that the waveform of the vibration intensity can be detected.

Specifically, referring to FIG. 8, the processor 300 may detect the pantograph 310 separated from the background for each image frame. At this time, generally only the pantograph 310 can be detected through a binarization imaging technique used to separate the background of the image.

The processor 300 detects the lower end line 320a of the current collecting plate extending from the lower end face of the current collecting plate with respect to the center point 325 of the current collecting plate and detects the lower end line 320a from the lower end line 320a of the current collecting plate It is possible to detect the oscillation intensity reference line 330a parallel to the horizontal line 330 of the current collecting plate by rotating the oscillating plate 330 by the front plate inclination (the 'current collecting plate angle' shown in FIG. 5C).

Then, the processor 300 can measure the left displacement and the right displacement of the lower end line 320a of the current collecting plate and the vibration intensity reference line 330a. Here, the current collecting plate is a rigid body having almost no warping, and when the left side goes up, the right side goes down, so that the sizes of the left and right vertical displacements of the current collecting plate can be the same.

Accordingly, as shown in FIG. 8, the processor 300 calculates the difference between the left displacement and the right displacement of the current collecting plate to calculate the vibration intensity. At this time, the left and right sides of the current collecting plate can be obtained by a template matching technique, but the present invention is not limited thereto. In addition, the processor 300 can calculate the difference between the left and right reference of the current collecting plate in units of mm using ratio information per each image pixel (mm), and calculates the left displacement And the right displacement can be converted into the vibration intensity in mm.

Hereinafter, with reference to FIGS. 9 and 10, a process of the processor 300 for measuring the vibration frequency characteristic of the pantograph will be described in detail.

9 is a view for explaining a method of measuring left and right displacements of a pantograph according to an embodiment of the present invention. And FIG. 10 is a diagram for explaining a method of calculating a vibration frequency based on an average value of left and right displacements of a pantograph according to an embodiment of the present invention.

The contact surface between the pantograph and the caten grid is defined as a straight line positioned at the uppermost point with respect to the pantograph object 310, and the current collector is detected through Hough transform. The degree of inclination of the pantograph can be obtained by considering the horizontal plane based on the center point of the collector plate and the position of the camera device installed on the vehicle.

As shown in FIG. 9, the processor 300 can measure the left and right displacements of the horizontal line 330 and the extension line 320 of the current collector plate, which intersect at the center point 325, for each image frame.

Referring to FIG. 10, the processor 300 measures left and right average vertical displacements based on left and right pantograph displacements calculated for each image frame. At this time, the processor 300 can calculate the average of the vertical displacement of the left and right sides of the current collecting plate measured for each frame for one second in the input pantograph image. The processor 300 performs Fourier transform on the measured average vertical displacement of the left and right sides to calculate the vibration frequency. For example, the processor 300 may apply a Fast Fourier Transform (FFT) algorithm to convert the obtained signal into a frequency component and an intensity of a corresponding component. At this time, the sampling rate can be determined by the image acquisition frequency have.

Next, with reference to FIGS. 11 and 12, a process of monitoring the pantograph state by analyzing the pantograph vibration characteristic of the pantograph image collected by the processor 300 in real time will be described in detail.

FIG. 11 is a diagram for explaining an abnormal situation recognition method for a vibrating circle of a pantograph contact surface according to an embodiment of the present invention. And FIG. 12 is an exemplary view for explaining an optimal hyperplane according to the magnitude of SVM margin applied to an embodiment of the present invention.

The processor 300 analyzes the input pantograph image in real time and classifies the pantograph image into a normal state and an abnormal state. At this time, the processor 300 recognizes a case where the vibration of the pantograph is generated such that the vibration of the pantograph is greater than the threshold, as in the case of arcing, as an abnormal condition.

According to the Bayesian decision theory, there is an error in selection of the wrong selection when simply selecting the state of an input image from normal or abnormal. Accordingly, the processor 300 classifies the normal and abnormal states of the pantograph image through a maximum margin classifier (e.g., Support Vector Machine (SVM)) which is a technique that generally reduces the generalization error.

A linear crystal boundary (i.e., an optimal hyperplane) having an optimal margin can be obtained from a linear discriminant function as shown in Equation 2 below.

&Quot; (2) "

는 초평면의 법선벡터가 되고,

는 원점에서 직선까지의 거리를 결정하는 값이 된다. 이때, 한 점

에서 결정경계까지의 거리

는 아래 수학식 3과 같이 계산할 수 있다.In Equation (2)

Becomes a normal vector of the hyperplane,

Is a value that determines the distance from the origin to the straight line. At this time,

To the crystal boundary

Can be calculated as shown in Equation (3) below.

&Quot; (3) "

와

의 비율을 적절히 조절하여 다음의 수학식 4와 같은 조건이 성립되도록 결정하여, 주어진 데이터를 두 클래스로 구분할 수 있다.With respect to such a crystal boundary, as shown in FIG. 12, the support vector (that is, the position data closest to the hyperplane) is used as a reference

Wow

Is determined so as to satisfy the following equation (4), and the given data can be classified into two classes.

&Quot; (4) "

In Equation (4), a hyperplane passing through a support vector of class C ₁ is referred to as a "plus plane", and a hyperplane passing through a support vector of class C ₂ is referred to as a "negative plane". A distance from a crystal boundary to a support vector The margin can be calculated using the sum of distances from the crystal boundary to the support vector on the negative plane. This margin can be calculated by using the following equation (5).

Equation (5)

의 값을 최소화해야 한다. 이를 목적함수로 표현하면 아래 수학식 6과 같다.Therefore, in order to maximize the margin M

Should be minimized. This can be expressed by the following equation (6).

&Quot; (6) "

Equation (7) below must be satisfied in order to minimize the above Equation (6) as well as the training data.

&Quot; (7) "

를 최소화하는 동시에 수학식 7과 같은 또 다른 조건을 만족시키는 파라미터를 찾는 문제를 해결하기 위해서는 라그랑제 승수(Lagrange multipliers)를 이용한 최적화 방법을 적용해야한다. 이에 따른, 최적화 식은 아래 수학식 8과 같이 표현할 수 있다.As described above, one function formula

The optimization method using Lagrange multipliers should be applied to solve the problem of finding a parameter satisfying another condition such as Equation (7). Accordingly, the optimization equation can be expressed by the following equation (8).

&Quot; (8) "

The processor 300 mechanically learns the pantograph images through the maximum margin classification and optimization formula described above and inputs the vibration characteristic values calculated from the pantograph image input in real time to the classifier to identify the steady state or the abnormal state.

Then, the processor 300 outputs the pantograph state classification result through an output device (e.g., a monitor and an audio output device) set as a monitoring result.

Hereinafter, a method for monitoring a pantograph of an electric railway vehicle according to an embodiment of the present invention will be described in detail with reference to FIG. All of the procedures shown in FIG. 13 may be processed through the processor 300.

13 is a flowchart illustrating a pantograph monitoring method according to an embodiment of the present invention.

First, the pantograph object is detected based on the detection result of the previous frame for each frame of the photographed pantograph image (S1310).

The method of detecting the pantograph object using the reduction method of the search area is the same as that described above with reference to FIG. 2 and FIG.

Next, the vibration characteristic is detected from the pantograph object detected for each successive input image frame (S1320).

At this time, from the detected pantograph object, the center point of the portion where the pantograph and the catenary contact (that is, the collecting plate), the horizontal line extending from the center point, and the extension line of the current collecting plate are detected and the inclination angle between the detected horizontal line and extension line is measured . The vibration period, the vibration intensity, and the vibration frequency are detected based on the change of the angle of each input image frame.

The method of detecting the pantograph vibration characteristics is the same as that described above with reference to Figs. 4 to 10 above.

Then, the vibration characteristic detected for the pantograph image is input to the predetermined machine learning classifier, and the normal state and the abnormal state are classified (S1330).

In this case, the classifier may be an SVM classifier for processing the maximum margin classification, thereby identifying a case where a large amount of vibration occurs, such as an arc for the pantograph, as an abnormal state. Then, the result of classifying the steady state and the abnormal state is output so that the user can confirm it.

This pantograph state monitoring method is the same as that described above with reference to Figs. 11 and 12. Fig.

The apparatus and method for monitoring an electric railway vehicle pantograph according to an embodiment of the present invention described above may be implemented in the form of a recording medium including instructions executable by a computer such as a program module executed by a computer. Such computer-readable media can be any available media that can be accessed by a computer, and includes both volatile and non-volatile media, both removable and non-removable media. The computer-readable medium may also include computer storage media, which may be volatile and / or non-volatile, implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, Nonvolatile, removable and non-removable media.

It will be understood by those skilled in the art that the foregoing description of the present invention is for illustrative purposes only and that those of ordinary skill in the art can readily understand that various changes and modifications may be made without departing from the spirit or essential characteristics of the present invention. will be. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. For example, each component described as a single entity may be distributed and implemented, and components described as being distributed may also be implemented in a combined form.

The scope of the present invention is defined by the appended claims rather than the detailed description and all changes or modifications derived from the meaning and scope of the claims and their equivalents are to be construed as being included within the scope of the present invention do.

10: Vibration test device of pantograph of electric railway vehicle
100: communication module 200: memory
300: Processor 310: Detected pantograph
320: Extension line of the current collector plate 320a: Bottom line of the current collector plate
325: center point of the current collector plate 330: horizontal line
330a: Vibration strength baseline

Claims (12)

An electric railway vehicle pantograph monitoring apparatus comprising:
A communication module for receiving the pantograph image of the pantograph,
Memory in which the pantograph monitoring program is stored; And
And a processor for executing a program stored in the memory,
Wherein the processor, upon execution of the pantograph monitoring program,
Detects the pantograph object for each image frame by comparing the input pantograph image with a preset pantograph template image, detects predetermined vibration characteristics by measuring the tilt angle and displacement of the current collecting plate from the detected pantograph objects, And classifying the detected vibration characteristics through a classifier that machine-learned vibration characteristics of the detected pantograph image to identify a steady state and an abnormal state of the pantograph,
And when the pantograph object is detected, sets the pantograph search area of the next image frame based on the detection result of the pantograph object of the previous image frame.
The method according to claim 1,
The processor comprising:
A search area is set for each image frame of the input pantograph image,
If the position of the pantograph object detected in the previous image frame is unknown, the entire area of the image frame is set as the search area,
When the position of the pantograph object detected in the previous frame is known, if the position of the known pantograph object is included in the predetermined search area, the same search area as before is applied to the corresponding image frame, and if the position of the known pantograph object is And resets the search area to the entire area when the search area is out of the preset search area.
3. The method of claim 2,
The processor comprising:
Changing a pantograph template image when a difference between at least one of brightness and color of the set pantograph template image is not less than a predetermined threshold value for an image frame in which a search area is set based on the position of the pantograph object of the previous image frame, Electric railway vehicle pantograph monitoring device.
The method according to claim 1,
Wherein the classifier is a support vector machine (SVM).
The method according to claim 1,
The processor comprising:
Detecting a center point of the current collecting plate from the detected pantograph object,
Measuring an angle between a horizontal line extending from the center point and an extension line of the current collecting plate,
Detecting the number of image frames in which the angle is zero-crossed by arranging the angles according to a sequence of image frames, detecting a frequency by dividing the number of image frames to be crossed by two by a number of frames per second,
And calculates the oscillation period of the pantograph based on the detected frequency.
6. The method of claim 5,
The processor comprising:
Detecting a displacement, which is a magnitude variation of the angle of the image frames,
And calculates the difference between the left displacement and the right displacement of the current collecting plate to calculate the vibration strength of the pantograph.
The method according to claim 6,
The processor comprising:
The average value of the upper left and lower right displacements of the current collecting plate is calculated,
And calculates a vibration frequency of the pantograph by processing the average vertical displacement of the left and right sides of the current collecting plate.
A method of monitoring a pantograph of an electric railway vehicle pantograph monitoring apparatus,
Collecting a pantograph image of the pantograph;
Comparing the pantograph image with a predetermined pantograph template image to detect a pantograph object for each image frame;
Detecting predetermined vibration characteristics by measuring a tilt angle and a displacement of the current collecting plate from the detected pantograph objects; And
The detected vibration characteristics are classified through a classifier that has previously learned the vibration characteristics of the pantograph image, and the vibration characteristics of the pantograph image detected before the identification of the steady state and the abnormal state of the pantograph are machine- And classifying the detected vibration characteristics to identify a steady state and an abnormal state of the pantograph,
And when the pantograph object is detected, sets the pantograph search area of the next image frame based on the pantograph object detection result of the previous image frame.
9. The method of claim 8,
Wherein detecting the pantograph object comprises:
A search area is set for each image frame of the input pantograph image,
If the position of the pantograph object detected in the previous image frame is unknown, the entire area of the image frame is set as the search area,
When the position of the pantograph object detected in the previous frame is known, if the position of the known pantograph object is included in the predetermined search area, the same search area as before is applied to the corresponding image frame, and if the position of the known pantograph object is And resetting the search area to the entire area when the search area is out of the predetermined search area.
10. The method of claim 9,
Wherein detecting the pantograph object comprises:
Changing a pantograph template image when a difference between at least one of brightness and color of the set pantograph template image is not less than a predetermined threshold value for an image frame in which a search area is set based on the position of the pantograph object of the previous image frame, Electric Railway Vehicle Pantograph Monitoring Method.
9. The method of claim 8,
Wherein the classifier is a Support Vector Machine (SVM).
9. The method of claim 8,
Wherein the step of detecting the vibration characteristics comprises:
Detecting a center point of the current collector plate from the detected pantograph object; And measuring an angle between a horizontal line extending from the center point and an extension line of the current collector plate,
Detecting the number of image frames in which the angle is zero-crossed by arranging the angles in accordance with a sequence of image frames, detecting a frequency by dividing the number of image frames subjected to zero crossing by 2 divided by the number of frames per second, Calculates a vibration period of the pantograph based on the frequency,
Detecting a displacement which is a magnitude variation of the angle of the image frames and calculating a difference between a left displacement and a right displacement of the current collecting plate to calculate a vibration intensity of the pantograph,
And calculating the average value of the upper left and lower right displacements of the current collecting plate and processing the Fourier transform for each of the average vertical displacements of the left and right sides of the current collecting plate to calculate the oscillation frequency of the pantograph. Way.

Images