KR101026351B1 - Railway rail wear measurement system using wavelet and its method - Google Patents

Abstract

By using the line laser and high-speed camera, images of the railroad rail vertical cuts are obtained, and the spatial information on the railroad rail vertical cuts is extracted and compared with the design section information of the railroad rails. Provided are railroad rail wear measurement systems and methods using wavelets that can be improved. The railway rail wear measurement system includes at least one camera for acquiring a raw image of the railway rail to which the line laser is irradiated; An image processor which detects a laser region at high speed from the raw image acquired by the camera; An outline coordinate extractor for extracting two-dimensional coordinate data corresponding to the obtained laser region; An outline spatial converter that obtains spatial information on a vertical cut plane of a three-dimensional railroad rail by mapping the extracted two-dimensional coordinate data to a vertical cut plane image of a railroad rail; A main processor that controls the driving of the line laser and controls operations of the image processor, the outline coordinate extractor, and the outline space converter; A coordinate data memory for storing spatial information on the obtained three-dimensional railway rail vertical cut surface; And a host computer receiving the spatial information on the vertical rail cut and analyzing the spatial information with the onboard analysis program.

Railway rail, wear measurement, wavelet, 3D spatial information, line laser, image processing

Description

System for measuring railway abrasion wear using wavelet and method thereof

The present invention relates to a railway rail wear measurement system, and more particularly, by applying a wavelet image processing algorithm to the railroad rail image obtained by using a line laser and a camera, an image for measuring the surface wear of the railroad rail A railroad rail wear measurement system using wavelets to obtain data.

After a certain period of time, railway rails become irregularly worn, creating grooves and producing noise. Grooves or cracks in these railway rails are caused by the failure to properly manage the wear of the railway rails. If this is left unattended, the wear of the railway rails must be accurately measured periodically because of the risk of safety accidents such as derailment of the train.

However, the railway rail wear measurement system according to the conventional technology using ultrasonic waves, etc., exhibits mechanical and functional limitations, and there is a problem in that accurate measurement is difficult due to the influence of the measurement state and the measurement environment. In addition, the railway rail wear measurement system according to the prior art had a problem that the measurement efficiency is low compared to the cost.

In general, conventional railway rail wear measurement system can be largely divided into high-speed detection equipment and self-propelled detection vehicle.

As the high speed detection equipment, equipment for measuring the vibration acceleration of the vehicle and the bogie is used, and when the train travels at a high speed, it is estimated that the track caused by the vibration of the vehicle is incorrect. However, the measurement by the vibration has a problem of low accuracy and error of the measurement value.

In addition, the self-propelled detection vehicle is used to identify more precise geometrical state of the track and to establish a maintenance work plan, which performs track linear detection, measurement of railway rail cross-section, measurement of railway rail surface defects, and wave wear measurement. . At this time, detection is possible at a detection speed of 160 km / h. These self-propelled detection vehicles can improve the detection accuracy by adopting inertial and optical non-contact measurement methods using accelerometers and cameras, but in particular, it is impossible to detect high-speed railways of 300 km / h and can accurately analyze railway rail wear phenomenon. none. In addition, in the case of severely worn or stagnant railway rails, an error occurs because the detection car proceeds out of the center of the rail rail head, and thus there is a problem in that accurate meter reading is impossible.

Accordingly, there is a demand for the development of a railway rail wear measurement system that can quickly and accurately inspect and manage the data measured in a short time at a detection speed of up to 300 km / h using a laser and a camera.

The technical problem to be solved by the present invention is to apply a wavelet image processing algorithm to the railway rail image obtained by using a line laser and a camera, the image for measuring the surface wear of the railway rail To provide a railway rail wear measurement system and method using a wavelet, which can obtain the data.

Another technical problem to be solved by the present invention is to accurately obtain the three-dimensional railway rail vertical cutting surface information on the surface of the railway rail and compare it with the cross-sectional information of the railway rail stored in the host computer, so that the degree of wear of the railway rail can be quickly and accurately corrected. To provide a railway rail wear measurement system and method using a wavelet that can be measured efficiently and efficiently.

Another technical problem to be achieved by the present invention is a wavelet capable of quantitatively determining the wear rate of a railroad rail at a speed of up to 300 km by acquiring spatial information on a vertically divided plane of a three-dimensional railroad rail surface. To provide a railroad rail wear measurement system and method thereof.

As a means for achieving the above-described technical problem, at least one railway rail wear measurement system using a wavelet according to the present invention, to obtain a raw image of the railway rails irradiated with a line laser (Line laser) More cameras; An image processor which detects a laser region at high speed from an original image acquired by the camera; An outline coordinate extractor for extracting two-dimensional coordinate data corresponding to the laser region obtained by the image processor; An outline spatial converter that obtains spatial information on a vertical cut plane of a three-dimensional railroad rail by mapping the extracted two-dimensional coordinate data to a vertical cut plane image of the railroad rail; A main processor controlling driving of the line laser and controlling operations of the image processor, an outline coordinate extractor, and an outline space converter; A coordinate data memory for storing spatial information on a three-dimensional railway rail vertical cut plane obtained from the outline space converter; And a host computer receiving the spatial information on the vertical cut plane of the three-dimensional railway rail and analyzing the spatial information with a mounted analysis program.

Here, the railway rail wear measurement system using a wavelet according to the present invention, a line laser for irradiating a laser to the rail rail vertical cut surface; And a laser driver configured to control the laser intensity of the line laser to obtain an optimal raw image, and the line laser may be a laser light source having a wavelength in an infrared band separated from an external light source.

In this case, the image processing unit has two redundant structures for alternately processing the raw image obtained by the camera into an odd frame and an even frame.

The image processor may include: an image data classifier configured to rearrange image data by applying a high speed algorithm; A raw image memory configured to store and manage image data information rearranged by the image data classifier; A laser region detector for accurately detecting an outline of a line laser by applying a wavelet-based image processing algorithm from an image stored in the raw image memory; And an image processing memory configured to store and manage line laser outline data information of the laser area detector.

In this case, the outline spatial converter is characterized in that the perspective transformation and the camera model are applied to obtain spatial information on the vertical cutting plane of the three-dimensional railway rail from the line laser region extracted from the two-dimensional coordinate data.

In this case, the host computer displays the spatial information on the vertical cutting surface of the three-dimensional railway rail, and the degree of wear of the railway rail compared with the cross-sectional information of the pre-stored railway rail to accurately analyze the extent of the railway rail wear It provides a quantitatively.

Here, the railway rail wear measurement system using a wavelet according to the present invention, the coordinate data memory and the host computer so as to communicate with each other, the real-time measured spatial information on the three-dimensional railway rail vertical cutting plane measured by the host It may further include a communication module for transmitting to a computer.

On the other hand, as another means for achieving the above technical problem, a railroad rail wear measurement method using a wavelet according to the present invention, a) obtaining a raw image of the railroad rail line irradiated; b) rapidly detecting a laser region from the obtained raw image; c) extracting two-dimensional coordinate data corresponding to the obtained laser region; d) obtaining spatial information on the vertical cut plane of the three-dimensional railway rail by mapping the extracted two-dimensional coordinate data to the vertical cut plane image of the railway rail; e) storing spatial information on the three-dimensional railway rail vertical cut plane obtained from the outline space converter; And f) receiving spatial information on the vertical cut plane of the three-dimensional railway rail, and analyzing the spatial information with a mounted analysis program.

According to the present invention, in terms of securing safety of railway operation, reducing maintenance costs, and securing vehicle diagnostic technology, an image measuring system capable of acquiring an image of a vertical section of a railway rail during high speed driving using a line laser and a high speed camera is measured. By analyzing the condition according to the wear level of the railway rail through the analysis data, it is possible to quickly and precisely obtain quantitative data on potential risk factors.

According to the present invention, it is possible to apply the same to all railroad rails of the same standard, it is possible to apply to the system that can replace the field of visual inspection by manpower such as fastener damage, sleeper crack, as well as defects on the railroad rail This saves time and money for maintenance inspections.

According to the present invention, by eliminating the vibration of the train caused by the excessive wear of the railway rail to improve the ride comfort and by the pre-detection of track damage caused by the impact of the railway rail and wheels caused by excessive wear and the defects present in the railway rail Railway operation can be done safely.

According to the present invention, it is possible to measure while mounted on a high-speed train at a speed of 300 km / h, and facility defects through systematic analysis and management of the data obtained by providing accurate data on the wear of the railway rail By observing the progress of the train and establishing the maintenance plan, systematic railroad rail management and operation stability can be maximized.

DETAILED DESCRIPTION Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art may easily implement the present invention. 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 the drawings, parts irrelevant to the description are omitted in order to clearly describe the present invention, and like reference numerals designate like parts throughout the specification.

Throughout the specification, when a part is said to "include" a certain component, it means that it can further include other components, without excluding other components unless specifically stated otherwise. Also, the term "part" or the like, as described in the specification, means a unit for processing at least one function or operation, and may be implemented by hardware, software, or a combination of hardware and software.

First, the concept of a wavelet applied to an embodiment of the present invention will be described.

Fourier Transform (FT), the most common technique for analyzing signals, is a technique that converts a value expressed as a function of time into a function of frequency. In other words, it is a method of calculating the magnitude of each sine wave by seeing that the sine waves having different frequencies overlap each other. This Fourier transform method is particularly useful for analyzing electrical signals in which several sine waves of different frequencies are mixed and are used to reduce noise by removing only signals of unwanted frequencies from the signal.

Wavelets are more advanced Fourier transforms (FTs). While the Fourier transform (FT) uses infinitely repeated sine waves as the fundamental waveform, only the frequency is changed and the correlation is revealed. Wavelet uses the waveform of one wavelength as the basic waveform and changes its magnitude and position to reveal the correlation. Technology. Here, changing the size is the same concept as changing the frequency of the FT, but changing the position is Wavelet's unique method.

However, in the case of Fourier transform (FT), since the infinitely repeated sinusoidal waveform is a fundamental waveform, the time information disappears when it is changed as a function of frequency. That is, it is possible to know which frequency components are many, but it is not known at which position the components appear in time. However, Wavelet has the advantage of knowing the time information as well as the frequency information because the wavelet of one wavelength changes its position along with its size. Analysis of a signal by Fourier transform (FT) shows a two-dimensional graph of the frequency and amplitude axes, while wavelet analysis shows three dimensions of the scale, translation, and value axes. Can be represented graphically.

These wavelets are mainly used for noise reduction or signal compression through signal analysis. For example, it is used for the analysis of signals whose waveforms are completely different from sinusoidal waveforms such as heart rhythm and brain waves, or where local signals have a significant influence on the analysis. Used for

As an embodiment of the present invention, by obtaining the three-dimensional railway rail vertical cutting surface information on the railway rail surface and comparing with the cross-sectional information of the railway rail stored in the host computer, the degree of wear of the railway rail can be accurately and efficiently Provided are railroad rail wear measuring systems using measurable wavelets and methods thereof.

1 is a block diagram of a railway rail wear measurement system using a wavelet according to an embodiment of the present invention.

Referring to FIG. 1, a railway rail wear measurement system 100 using a wavelet according to an embodiment of the present invention includes a camera 110, an image processor 120 and 120 ′, an outline coordinate extractor 130, and an outline space. The converter 140 includes a main processor 150, a coordinate data memory 160, a communication module 170, a host computer 180, a line laser 191, and a laser driver 192. Here, the image processing units 120 and 120 'are dualized into the first and second image processing units, respectively, the image data classifiers 121 and 121', the raw image memories 122 and 122 ', and the laser region detectors 123 and 123. ') And image processing memories 124 and 124'.

The line laser 191 irradiates a laser beam to the railroad vertical cutting plane, and the laser driver 192 adjusts the laser intensity of the line laser 191 to obtain an optimal raw image.

At least one camera 110 acquires an image of the railroad to which the line laser 191 is irradiated. In this case, the camera 11 has a band pass having a narrow pass band centering on the wavelength of the line laser 191 to effectively distinguish and effectively block light of various wavelengths incident from the line laser 191 of the infrared band and an external light source. It is configured to include an optical filter, mounted on a high-speed train running more than 300km to use the high speed camera having a shooting speed of 480 frames or more per second to measure the wear of the railway rail.

The first and first image processing units 120 and 120 'apply odd-numbered and even-numbered frames to detect a laser region at high speed by applying a wavelet-based image processing algorithm from the raw image acquired by the camera 110. Take turns.

Accordingly, the first and second image processing units 120 and 120 'are mounted on a train traveling at a high speed of 300 km or more and use the high speed camera 110 having a shooting speed of 480 frames or more per second in order to measure wear of the railway rail. In order to process the image data generated at high speed in real time, it is divided into two physically divided image processing areas so as to alternately process odd frames and even frames.

The outline coordinate extractor 130 extracts two-dimensional coordinate data of the laser region from the laser regions acquired by the first and first image processors 120 and 120 ′.

The outline space converter 140 applies a perspective transformation algorithm and a camera model to the two-dimensional coordinate data obtained by the outline coordinate extractor 130 and maps it to a vertical cut plane image of a railroad rail, thereby generating a space for a vertical cut plane of a three-dimensional railroad rail. Information is obtained.

The main processor 150 controls the first and first image processors 120 and 120 ′, the outline coordinate extractor 130, the outline space converter 140, and the laser driver 192.

Coordinate data memory 160 stores the spatial information on the three-dimensional railway rail vertical cut surface obtained from the outline space converter 140.

The communication module 170 connects the coordinate data memory 160 and the host computer 180 to transmit spatial information on the vertical cut plane of the three-dimensional railway rail measured in real time to the host computer 180.

The host computer 180 is equipped with an analysis program, and compares the spatial information on the vertical cutting plane of the three-dimensional railway rails received from the communication module 170 with the cross-sectional information of the stored railway rails to compare the degree of wear of the railway rails. Will be analyzed.

Railroad rail wear measurement system 100 using a wavelet according to an embodiment of the present invention, by irradiating the line rail 191 to the railroad rail, and photographing it with one or more cameras 110 according to the wear of the railroad rail Acquire a raw image including the surface state information, and from the raw image, the first and second image processing units 120 and 120 'and the outline coordinate extractor 130 The two-dimensional coordinate data is obtained by extracting the line laser irradiation area with information. Subsequently, the line laser irradiation area obtained by the two-dimensional coordinate data by the outline space converter 140 is converted into three-dimensional spatial coordinates to obtain spatial coordinates corresponding to the vertical cut planes of the railway rails. The wear rate of the railway rail is quantitatively measured by comparing the information with respect to the vertical cutting plane of the rail.

In addition, the railway rail wear measurement system 100 using a wavelet according to an embodiment of the present invention processes the image by applying a high-performance FPGA and DSP to perform the image processing for the high-speed image data, the laser region The detectors 123 and 123 ', the outline coordinate extractor 130, and the outline space converter 140 may be divided and disposed in the FPGA and the DSP so that parallel processing is possible. For example, if the entire image for each frame is divided into N regions and divided into image segments, an algorithm for image processing is configured in parallel to simultaneously process each image segment included in one frame at the same time. The efficiency and precision can be improved by distributing the amount of computation required for the image processing of the image and shortening the time required for the image processing.

FIG. 2 is a diagram illustrating a configuration in which the image processor of FIG. 1 is duplicated.

Referring to FIG. 2, two redundant image processing units 120 and 120 'may include image data classifiers 121 and 121', raw image memories 122 and 122 ', laser region detectors 123 and 123', and And image processing memories 124 and 124 '.

The image data classifiers 121 and 121 'of the first and second image processing units 120 and 120' respectively rearrange the image data in each image frame to effectively process the image data transmitted from the camera 110. The raw image memories 122 and 122 'temporarily store rearranged image data obtained from the image data classifiers 121 and 121'.

In addition, the laser region detectors 123 and 123 'of the first and second image processing units 120 and 120' may use wavelet-based image processing algorithms on the raw images stored in the raw image memories 122 and 122 ', respectively. By applying the method, the edge region of the line laser irradiated to the railroad is accurately detected, and the image processing memories 124 and 124 'store the laser region information obtained from the laser region detectors 123 and 123', respectively. In order to apply the continuous image processing process, the data memory function is performed.

In other words, the image processing units 120 and 120 'are dualized, and an odd frame (the image processing path of the first image processing unit 120) is obtained for the continuous frames of the raw image data obtained from the camera 110. ) And even frames (image processing paths of the second image processing unit 120 ') to proceed in parallel to ensure optimized system performance and to perform image processing efficiently and quickly.

In addition, the line laser region extraction by the laser region detectors 123 and 123 'is applied by applying a wavelet-based image processing algorithm to the FPGA, and the outline of the line laser region in the image uses a parameter notation technique. By converting the two-dimensional representation of the line laser region to a one-dimensional representation.

Therefore, the outline of the line laser 110 is expressed as a function of the parameter t, and it is expressed as a continuous function of two pairs of functions {X (t), Y (t)}, and thus the outline is represented by a one-dimensional wavelet transform. By dividing, outline information of all scales can be obtained.

As a result, it is possible to extract the characteristics of the line laser region by marking the outlines of all scales using the wavelet transform. That is, since the bi-orthonormal wavelet has more advantages than the orthonormal wavelet, the binary form of the third-order B-spline function that is the basic function of the bi-orthonormal is selected and used as the scale function. In addition, it utilizes the advantages of hardware and software that is simple to implement and easy to interpret, so that it is used as smoothing function by filtering image data with dilated B-spline function.

Here, Equation 1 and Equation 2 are used to establish a relationship between two scales between the scale function and the wavelet, and the signals approximated by Equation 1 and Equation 2 are decomposed as signals related to the wavelets. The wavelet transform is composed of two signals, and the wavelet transform is used to restore the approximated signal to the original signal for scale representation.

이 얻어질 수 있다.As such, because the B-spline function and its derivatives (the derivatives) are very well interpreted and require only simple operations, the approximated signal is the equivalent of the original signal because the B-spline function provides an appropriate approximated signal. You can think of it as interpolated with the B-spline function. In addition, a pair of wavelets are converted for each scale from the original object outline notation {X (t), Y (t)} as shown in Equation 3 below.

This can be obtained.

In addition, the curvature of the outline may be generally defined as the derivative of the tangent angle to the curve. Thus, using the parameter representation of the curve, the curvature of the outline of the line laser region is expressed as The outlines can be obtained at a expressed and lower resolution. Since the curvature function is represented by the wavelet transform coefficient of Equation 5, it becomes an important feature in the shape analysis of the curvature function, and zero-crossing points that can be used for a scale-independent notation method are applied to all scales. It can be found from the curvature function over and thus, it is possible to accurately extract the line laser area in the obtained image of the railway rail.

3 is a diagram illustrating a raw image obtained by the line laser and the camera of FIG. 1, and illustrates a raw image of the railway rail 300 obtained by the line laser 191 and the camera 110. And 320 indicates wear included in the raw image.

4 is a diagram illustrating result images of line laser region detection acquired by an image processor according to an exemplary embodiment of the present invention. FIG. 4 is a line laser obtained by an image processor 120 including a laser region detector 123. Results images 410, 420, 430, and 440 for region detection are illustrated.

Meanwhile, FIG. 5 is a diagram illustrating a result of implementing two-dimensional coordinate data acquisition for line laser area detection by a window-based program by the outline coordinate extractor 130 of FIG. 1.

Referring to FIG. 5, reference numeral 510 denotes an acquired raw image, and reference numeral 520 denotes a processed image. Reference numeral 530 denotes image information, and reference numeral 540 denotes coordinate data, respectively. Reference numerals 551 to 556 denote input buttons, respectively, an image acquisition button 551, an image processing button 552, an area capture button 553, an extraction button 554, a system disconnection button 555, and a program. Each end button 556 is shown. Therefore, it can be obtained as coordinate data in two dimensions according to the result of implementing the program based on the window.

Meanwhile, FIG. 6 illustrates that the perspective transformation for converting two-dimensional coordinate data of the outline of the line laser obtained from the outline coordinate extractor 130 of FIG. 1 into three-dimensional spatial coordinates by the outline space converter 140 is illustrated. It is a figure.

As shown in FIG. 6, for example, perspective transformation plays a central role in image processing because it is very close to the way a human composes an image by directly looking at a three-dimensional world. It provides a fundamentally different role from other transformations, including nonlinearities divided by coordinate values. In this case, the camera coordinate system (x, y, z) has an image plane coinciding with the xy plane, and has an optical axis (determined by the center of the lens) along the Z axis.

)이며, 만일 카메라가 멀리 떨어진 물체와 초점이 맞는 거리에 있고 카메라 좌표계가 실좌표계 (x, y, z)와 일직선상에 맞춰져 있다면,

를 렌즈의 초점거리라고 할 수 있으며, 투사된 3차원 점의 영상 평면 좌표를 수학식 6과 같이 구할 수 있다.Therefore, the center of the image plane is the origin and the lens center is the coordinate (0, 0,

), If the camera is at a focused distance from a distant object and the camera coordinate system is aligned with the real coordinate system (x, y, z),

It may be referred to as the focal length of the lens, and the image plane coordinates of the projected three-dimensional points may be obtained as shown in Equation 6.

,

,

,

,

,

)일 때, 수학식 7과 같이 벡터적인 표현이 가능해진다.In addition, Equation 6 has a nonlinear structure divided by the variable Z, and it is preferable to use homogeneous coordinates expressed in the form of a linear matrix such as rotation, movement, and scaling. Accordingly, the homogeneous coordinate of a point having Cartesian coordinates (x, y, z) is (

,

,

,

), A vector expression can be obtained as shown in Equation (7).

On the other hand, Figure 7 maps the line laser area converted to the three-dimensional spatial coordinates of the railway rail in the high-precision railway rail wear measurement system using a wavelet according to an embodiment of the present invention to the vertical cut plane of the railway rail To illustrate an example of applying a camera model.

만큼 좌우로, 각도

만큼 기울어져 철도레일을 비스듬한 각도에서 바라보게 된다. 여기서, Pan은 x와 X축 사이의 각이고, Tilt는 z와 Z축 사이의 각이라 할 때, 실좌표계의 원점에서 카메라 축의 중심까지 오프셋을 W■로 표기하고, 카메라 축의 중심에서 영상 평면 중심까지의 오프셋은 구성요소가 (r₁, r₃, r₃)인 벡터 r을 구하게 된다. 이후, 실제 점에 대해 영상 평면 좌표를 얻기 위하여 원근 변환을 적용하여 철도레일의 수직 절단면에 매핑하기 위한 공간좌표를 획득하게 된다.As shown in FIG. 7, for example, the camera model constitutes a basic mathematical model of the camera, and assuming that the camera and the real coordinate system are in a straight line, the camera model finds the camera and three-dimensional points (denoted by W). The camera angles using the actual coordinate system (X, Y, Z)

As long as left and right, angle

Tilt as much as you look at the rail from an oblique angle. Here, Pan is an angle between the x and X axes, and Tilt is an angle between the z and Z axes, and the offset is expressed as W ■ from the origin of the real coordinate system to the center of the camera axis, and the center of the image plane from the center of the camera axis. The offset to yields a vector r whose components are (r ₁ , r ₃ , r ₃ ). Subsequently, in order to obtain image plane coordinates with respect to actual points, spatial coordinates for mapping to the vertical cut plane of the railway rail are obtained by applying a perspective transformation.

Therefore, when the center of the camera axis and the origin of the image plane are at the origin of the real coordinate system and the camera is in the normal position in that all coordinate axes coincide, the displacement of the gimbal center at the origin, the pan on the x axis, the tilt of the z axis, and the gimbal center In the order of displacement of the image plane's displacement relative to, the sequence of points seen through the camera after being displaced from the normal position is very different. Because of this, it is possible to move back to the normal position if applied correctly in the same order of steps to virtually all real points, and the camera at the normal position satisfies the conditions for the application of perspective transformation.

) 및 Tilt Angle(

)에 대한 변환 행렬(R)은 수학식 8과 수학식 9에 의하여 표현될 수 있으며, 기하학적 배열을 만족하는 카메라에 의한 동차 실점은 카메라 좌표계에서 수학식 10과 같은 동차 표현을 갖는다.Accordingly, the transformation matrix C and the Pan Angle () for the origin displacement of the image plane by the vector r are

) And Tilt Angle (

Transformation matrix R for) may be represented by Equations 8 and 9, and the homogeneous actual point by the camera satisfying the geometric arrangement has the same homogeneous expression as in Equation 10 in the camera coordinate system.

의 네 번째 요소로 첫 번째와 두 번째 요소를 나누어서 영상점의 직교 좌표 (x, y)를 얻음으로써, 철도레일의 수직 절단면에 매핑하기 위한 공간좌표의 각 점에 대한 최종 값을 수학식 11과 수학식 12로 나타낸 바와 같이 획득할 수 있다.therefore

By dividing the first and second elements by the fourth element of to obtain the Cartesian coordinates (x, y) of the image points, the final value for each point in the spatial coordinates for mapping to the vertical cut plane of the railway rail It can be obtained as shown in equation (12).

Meanwhile, FIG. 8 is a diagram illustrating an image processing procedure for implementing high speed image processing in a railway rail wear measurement system using a wavelet according to an embodiment of the present invention.

As shown in FIG. 8, for example, a high-speed camera capable of acquiring an image of 480 frames per second to measure a wear rate of a railway rail mounted on a vehicle traveling at a speed of 300 km is used for each frame. When the size of the generated raw image is 1MB, the amount of raw image generated and processed per second amounts to 0.5GB. In this case, all image processing of 1MB raw image must be completed within 2㎳ (2 / 1000sec), so the entire image processing process focuses on the data memory in four areas to realize the function and performance of high-speed image processing in real time. Each area will work independently within the specified time.

Accordingly, the entire system is divided into four independent image processing regions by the raw image memories 122 and 122 ', the image processing memories 124 and 124', and the coordinate data memory 170, wherein the first region is The raw image input from the camera 110 may be converted into an image rearranged by the image data classifiers 121 and 121 'and stored in the original image memories 122 and 122'.

In addition, the second region detects the line laser region from the rearranged image obtained from the raw image memories 122 and 122 'through the laser region detectors 123 and 123' and stores the image in the image processing memories 124 and 124 '. It performs a function continuously.

In addition, the third region obtains the 2D coordinate data from the line laser region obtained from the image processing memories 124 and 124 'via the outline coordinate extractor 130, and then uses the outline space converter 140 to A function of acquiring spatial coordinates for mapping to the vertical cut plane and storing them in the coordinate data memory 160 is continuously performed.

In addition, the fourth region continuously performs a function of transmitting the final coordinate data acquired from the coordinate data memory 160 to the host computer 180 through the communication module 170.

Meanwhile, FIG. 9 is a diagram illustrating image processing for parallel computation processing for laser region detection, outline coordinate extraction, and outline spatial transformation according to an embodiment of the present invention.

As shown in FIG. 9, for example, if an entire image of each frame is divided into four regions and divided into image segments from the first image data region to the fourth image data region, each image included in one frame is included. By parallelizing the algorithm for image processing from the first image processing module 123a to the fourth image processing module such that the image segments are processed simultaneously at the same time, the amount of computation required for advanced image processing is distributed, and the image It is designed to increase efficiency and precision by reducing the time required for processing.

On the other hand, Figure 10 is a diagram illustrating a parallel processing process of the entire image processing for the high-speed image processing of the railway rail wear measurement system using a wavelet according to an embodiment of the present invention.

As shown in FIG. 10, for example, when the raw image from the camera 110 is incremented by 1T from the start time of 0T and starts continuously in the first frame and starts with the odd frame starting with A and starting with B The even frames can be independently processed without interference with each other.

Therefore, the raw image of the odd frame obtained from the camera 110 for a time from 0T to 1T is stored in the raw image memory 192 located on the image processing path for the odd frame by the A-1 process, and from 1T to 2T. The raw image memory of the even frame obtained from the camera 110 by the image processing unit 120, 120 'configured in a redundant manner is located on the image processing path for the even frame by the B-1 process ( When stored at 192, image processing processes A-2 and A-3 are performed by the laser region detector 123 and the outline coordinate extractor 130 in the image processing path for odd frames.

In addition, in the image processing path for the odd frames by the image processing units 120 and 120 'configured in a duplexing manner from 2T to 3T, the raw image of the new odd frame acquired from the camera 110 is stored in the raw image memory 122. And the new A-1 process, which is stored at < RTI ID = 0.0 > 122 ', < / RTI > and at the same time the image processing process B-2 and B by the laser region detectors 123 and 123' and the outline coordinate extractor 130 in the image processing path for even frames When -3 proceeds, the spatial coordinates of the railway rail vertical cut surface by the outline space converter 140 are performed for the odd frames. Therefore, a series of image processing processes are simultaneously performed in each region divided into independent image processing regions of the first region, the second region, and the third region by the memory storage means.

Accordingly, the A-5 process in which the final coordinate data for the original image of the odd frame first obtained during the time period from 3T to 4T is transmitted from the coordinate memory 160 to the host computer 180 through the communication module 170. Further proceeds. Subsequently, as shown in FIG. 9 after 3T, a series of image processing processes are simultaneously performed in each region divided into independent image processing regions of the first region, the second region, the third region, and the fourth region by the memory storage means. Proceeds simultaneously without interference.

On the other hand, Figure 11 is a flow chart of the railroad rail wear measurement method using a wavelet according to an embodiment of the present invention.

Referring to FIG. 11, in the method of measuring wear of a railroad rail using a wavelet according to an exemplary embodiment of the present invention, first, when a laser of which the intensity of the line laser is adjusted to obtain an optimal raw image is irradiated (S110), the line laser A raw image of the irradiated railway rail is obtained (S120). In this case, the line laser may use a laser light source having a wavelength of an infrared band separated from an external light source.

Next, a laser region is detected at a high speed from the obtained raw image (S130). The obtained raw image is divided into odd frames and even frames by two duplex structures, and alternately processed. Specifically, the image data is rearranged by applying a high-speed algorithm, and the stored image data information is stored and After that, the edge of the line laser is precisely detected by applying a wavelet-based image processing algorithm from the stored image, and the detected line laser outline data information is stored and managed.

Next, the two-dimensional coordinate data corresponding to the obtained laser region is extracted (S140), and then, from the extracted two-dimensional coordinate data to a vertical cut plane image of the railway rail, a three-dimensional railway rail vertical cut plane Obtain spatial information for (S150). In this case, a perspective transformation and a camera model technique are applied to obtain spatial information on a vertical cut plane of a three-dimensional railway rail from the line laser region extracted from the two-dimensional coordinate data.

Next, in operation 160, spatial information about the vertical cut plane of the three-dimensional railway rail obtained from the outline space converter is stored.

Next, the spatial information on the vertical cut plane of the three-dimensional railway rail is transmitted to the host computer (S170), and then, the host computer analyzes the spatial information with the analysis program mounted (S180). That is, the host computer displays the spatial information on the vertical cutting surface of the three-dimensional railway rail, and the degree of wear of the railway rail compared with the cross-sectional information of the pre-stored railway rail to accurately analyze the extent of the railway rail wear Can be provided quantitatively.

As a result, a railway rail wear measurement system using a wavelet according to an embodiment of the present invention is an image measuring system capable of acquiring an image of a vertical section of a railway rail during high speed driving using a line laser and a high speed camera. Through the data, it is possible to analyze the current and internal condition of the railway rail, so that quantitative data about potential risk factors can be obtained quickly and precisely, which will be widely used in the field of railway rail wear measurement system.

In addition, as the expansion of high-speed railways and the increase of new railway services such as urban railways and light rail trains in each region, the continuous expansion of railway rails is being carried out, but systematic management and maintenance of railway rails Inspection system for the whole depends on import income. Railroad rail wear measurement system using a wavelet according to an embodiment of the present invention will be a core infrastructure technology that can achieve an advanced railway culture and contribute to the increase of exports due to import substitution effect and export of advanced railway technology .

The foregoing description of the present invention is intended for illustration, and it will be understood by those skilled in the art that the present invention may be easily modified in other specific forms without changing the technical spirit or essential features 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 type may be implemented in a distributed manner, and similarly, components described as distributed may be implemented in a combined form.

The scope of the present invention is shown by the following claims rather than the above description, and all changes or modifications derived from the meaning and scope of the claims and their equivalents should be construed as being included in the scope of the present invention. do.

1 is a block diagram of a railway rail wear measurement system using a wavelet according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating a configuration in which the image processor of FIG. 1 is duplicated.

FIG. 3 is a diagram illustrating a raw image obtained by the line laser and the camera of FIG. 1.

4 is a diagram illustrating result images for line laser region detection acquired by an image processor according to an exemplary embodiment of the present invention.

FIG. 5 is a diagram illustrating a result of implementing two-dimensional coordinate data acquisition for line laser area detection by a window-based program by the outline coordinate extractor of FIG. 1.

FIG. 6 is a diagram illustrating a perspective transformation for converting two-dimensional coordinate data of an outline of a line laser obtained from the outline coordinate extractor of FIG. 1 into a three-dimensional spatial coordinate in an outline space converter.

FIG. 7 is a diagram illustrating an example of applying a camera model to map a line laser region converted to a three-dimensional spatial coordinate in the outline space converter of FIG. 1 to a vertical cut plane of a railway rail.

FIG. 8 is a diagram illustrating an image processing procedure for implementing high speed image processing in a railway rail wear measurement system using a wavelet according to an embodiment of the present invention.

FIG. 9 is a diagram illustrating image processing for parallel computation processing for laser region detection, outline coordinate extraction, and outline space transformation according to an embodiment of the present invention.

FIG. 10 is a diagram illustrating a parallel processing process of all image processing for implementing high speed image processing of a railway rail wear measurement system using a wavelet according to an exemplary embodiment of the present invention.

11 is a flowchart illustrating a method for measuring railway rail wear using a wavelet according to an embodiment of the present invention.

<Brief Description of Drawings>

100: railway rail wear measurement system using wavelets

110: camera 120, 120 ′: first and second image processing units

121, 121 ': Image data classifier 122, 122': Raw image memory

123, 123 ': laser range detector 124, 124': image processing memory

130: outline coordinate extractor 140: outline space converter

150: main processor 160: coordinate data memory

170: communication module 180: host computer

191: line laser 192: laser drive unit

Claims (15)

At least one camera for obtaining a raw image of a railroad to which a line laser is irradiated; An image processor which detects a laser region at high speed from the raw image acquired by the camera; An outline coordinate extractor for extracting two-dimensional coordinate data corresponding to the laser region obtained by the image processor; An outline spatial converter that obtains spatial information on a vertical cut plane of a three-dimensional railroad rail by mapping the extracted two-dimensional coordinate data to a vertical cut plane image of the railroad rail; A main processor controlling driving of the line laser and controlling operations of the image processor, an outline coordinate extractor, and an outline space converter; A coordinate data memory for storing spatial information on a three-dimensional railway rail vertical cut plane obtained from the outline space converter; And a host computer receiving the spatial information on the vertical cut plane of the three-dimensional railway rail and analyzing the spatial information with a mounted analysis program. A line laser for irradiating a laser to the vertical rail cut of the railway rail; And a laser driver for adjusting and driving the laser intensity of the line laser to obtain an optimal raw image. delete The method of claim 1, The line laser is a railway rail wear measurement system using a wavelet, characterized in that the laser light source having a wavelength of the infrared band is separated from the external light source. The method of claim 1, The image processing unit of the railway rail wear measurement system using a wavelet characterized in that it has two dual structure for alternately processing by dividing the raw image obtained by the camera into an odd frame and an even frame. The method of claim 1, wherein the image processing unit, An image data classifier for rearranging image data by applying a high speed algorithm; A raw image memory configured to store and manage image data information rearranged by the image data classifier; A laser region detector for accurately detecting an outline of a line laser by applying a wavelet-based image processing algorithm from an image stored in the raw image memory; And Image processing memory for storing and managing line laser outline data information of the laser area detector. Railway rail wear measurement system using a wavelet comprising a. The method of claim 1, The outline spatial converter is a wavelet characterized in that a perspective transformation and a camera model technique are applied to obtain spatial information on a vertical cutting plane of a three-dimensional railway rail from a line laser region extracted from two-dimensional coordinate data. Used railway rail wear measurement system. The method of claim 1, The host computer displays spatial information on the vertical cut of the three-dimensional railway rail, and compares the degree of wear of the railway rail with the cross-sectional information of previously stored railway rails in order to accurately analyze the extent of the railway rail wear. Railway rail wear measurement system using a wavelet, characterized in that provided by. The method of claim 1, Communication module for connecting the coordinate data memory and the host computer so as to communicate, and transmits the spatial information on the vertical cutting plane of the three-dimensional railway rail measured in real time to the host computer Railway rail wear measurement system using a wavelet further comprising. a) acquiring a raw image of the railway rail irradiated with the line laser; b) rapidly detecting a laser region from the obtained raw image; c) extracting two-dimensional coordinate data corresponding to the obtained laser region; d) obtaining spatial information on the vertical cut plane of the three-dimensional railway rail by mapping the extracted two-dimensional coordinate data to the vertical cut plane image of the railway rail; e) storing spatial information on the obtained three-dimensional railway rail vertical cut surface; And f) receiving spatial information on the vertical cut of the three-dimensional railway rail, and analyzing the spatial information with a mounted analysis program; Railway rail wear measurement method using a wavelet comprising a. 10. The method of claim 9, In the step a), the railway rail wear measurement method using a wavelet, characterized in that for irradiating the laser is adjusted the intensity of the line laser to obtain an optimal raw image. 10. The method of claim 9, The line laser of step a) is a railway rail wear measurement method using a wavelet, characterized in that the laser light source having a wavelength of the infrared band separated from the external light source. 10. The method of claim 9, B) the railway rail wear measurement method using a wavelet, characterized in that for processing the obtained raw image alternately divided into odd frames and even frames by the two duplicated structure. The method of claim 9, wherein b), b-1) rearranging the image data by applying a high speed algorithm; b-2) storing and managing the rearranged image data information; b-3) precisely detecting an outline of a line laser by applying a wavelet-based image processing algorithm from the stored image; And b-4) storing and managing the detected line laser outline data information Railway rail wear measurement method using a wavelet comprising a. 10. The method of claim 9, Step d) is a wavelet characterized in that a perspective transformation and a camera model technique are applied to obtain spatial information on a vertical cutting plane of a three-dimensional railway rail from a line laser region extracted from two-dimensional coordinate data. Measuring method of railway rail wear used. 10. The method of claim 9, In step f), the spatial information on the vertically cut railway rail in three dimensions is displayed, and the degree of wear of the railway rail is compared with the cross-sectional information of the previously stored railway rail to precisely analyze the extent of the railway rail wear. Railroad rail wear measurement method using a wavelet characterized in that it provides quantitatively.

Images