WO2021043518A1

WO2021043518A1 - Method and measuring apparatus for measuring an object of a track

Info

Publication number: WO2021043518A1
Application number: PCT/EP2020/071946
Authority: WO
Inventors: Martin BÜRGER; Gerald Zauner
Original assignee: Plasser & Theurer Export Von Bahnbaumaschinen Gesellschaft M.B.H.
Priority date: 2019-09-05
Filing date: 2020-08-05
Publication date: 2021-03-11
Also published as: EP4025868A1

Abstract

The invention relates to a method for measuring an object (1) of a track by means of a laser stripe sensor, in which a scanning line (6) is projected onto a surface (2) of the object (1) by means of a laser (3), and the surface (2) is recorded by means of a digital image recorder (4). According to the invention, the scanning line (6) is projected at periodically varied light intensity, wherein the image recorder (4) records a series (11) of images (12) and wherein image positions (P(x, y)) corresponding in the image series (11) are evaluated together by means of a detection device (13), in order to detect a reproduction of the scanning line (6). The projection of the scanning line (6) is distinguishable from reflections of other light sources by means of the periodically varied light intensity.

Description

description

Method and measuring device for measuring an object on a track

Technical area

The invention relates to a method for measuring an object of a track by means of a laser light section sensor, a scanning line being projected onto a surface of the object by means of a laser and the surface being recorded by means of a digital image recorder.

The invention also relates to a measuring device for carrying out the method.

State of the art

[02] For example, DE 102015 119409 A1 describes a laser light section

Sensor for measuring the deformation of a rail is known. A laser projects a scanning line (laser line) onto the rail. A reflection image of the scanning line is recorded using imaging optics and a high-resolution digital image sensor. The output signals of the digital image sensor are evaluated in an evaluation device to determine rail deformation.

Presentation of the invention

[03] The invention is based on the object of improving a method of the type mentioned at the outset in such a way that disruptive influences from other light sources do not impair the measurement. Another object of the invention is to specify a corresponding measuring device.

[04] According to the invention, these objects are achieved by the features of independent claims 1 and 8. Dependent claims specify advantageous embodiments of the invention.

[05] The scanning line is projected onto the surface with periodically varied light intensity, the image recorder recording a series of images and using corresponding image positions in the image series a detection device are jointly evaluated in order to detect an image of the scanning line. Due to the periodically varied light intensity, the projection of the scanning line can be distinguished from reflections from other light sources. A correspondingly further developed laser light section sensor (3D laser profile sensor, light sectioning method) leads to increased measurement robustness when used outdoors, especially when exposed to direct sunlight.

[06] The recording of an image series ensures that the resulting changed brightness (reflection intensity) of the imaged scanning line is reproduced in the sequence of images. With the joint evaluation of corresponding image positions in the image series, this effect can be distinguished from other reflections with constant or differently changed intensity. In this way, interference from solar radiation or lighting fixtures can be avoided.

In a further development of the method it is provided that the from

The light received by the image sensor is filtered by means of a narrow-band optical filter. The optical filter is matched to a central wavelength of the laser. The delimitation according to the invention by evaluating the varied brightness is then only required from interfering light with the same wavelength. Light with a different wavelength does not reach the image sensor.

[08] Another improvement provides that the image series is recorded at least over a period of the periodically varied light intensity of the laser. This measure simplifies the subsequent evaluation, because in the series of images the difference between the light from the laser and the light from other sources becomes clearer.

In addition, the subsequent evaluation is simplified if matching image positions are evaluated together in at least three images. At least five brightness values are then available for the evaluation of each pixel. It can thus be clearly determined whether the respective pixel images a surface location of the object to be measured which is illuminated by means of the laser. A joint evaluation of only two images would be too few for a detection of a periodic change in the brightness values, because no progression curve can be modeled.

In a preferred variant of the method, the light intensity is varied sinusoidally, with a pixel-by-pixel Fourier transformation of an intensity signal (brightness signal) being carried out to evaluate the image series for the corresponding image positions, with an amount of a 1st harmonic of the Fourier signal for each pixel. Spectrum is determined and the amounts of the 1st harmonic are evaluated in order to detect imaging pixels of the scanning line. Such an evaluation can be carried out with fast algorithms, whereby FPGAs (Field Programmable Gate Array) or GPUs (Graphics Processing Unit) can be used as hardware. By means of such universal components, the sinusoidal modulation of the scanning line can be detected in an efficient manner.

[11] The jointly evaluated image series forms an image stack with corresponding image content, with sinusoidal brightness gradients only occurring at the points illuminated with the laser. It is therefore checked how high the absolute value of the 1st harmonic of each evaluated pixel is. Significantly increased values are present where the scanning line lies in the image stack. For the evaluation of measurement results, the image stack is merged into a common image, with the detected scanning line being highlighted.

[12] It is advantageous if a threshold value is specified for the evaluation of the 1st harmonic and if a uniform brightness value is specified for all pixels whose amount of the 1st harmonic is below the threshold value. For example, all pixels that do not image a point illuminated by the laser receive a brightness value equal to zero. Thus, in the merged image of the jointly evaluated image series, only the depicted scanning line is visible. Based on their course, the track object under consideration can be measured using optical triangulation.

In another variant, an electronic filtering of an intensity signal (brightness signal) is carried out pixel-by-pixel to evaluate the image series for the corresponding image positions, the filtered intensity signal of the respective pixel being evaluated for imaging pixels to detect the scanning line. An electronic filter is set up in such a way that only intensity signals that are varied in accordance with the laser light can pass. In this way, those pixels are recognized that depict surface areas illuminated by the laser.

In the measuring device according to the invention for performing one of the methods described, the laser is set up to project a scanning line with periodically varied light intensity, the image recorder being designed to record an image series and the detection device being designed to simultaneously evaluate image positions that match in the image series . Such a device can comprise known components of a laser light section sensor, a modified control of the laser and evaluation logic according to the invention being implemented.

[15] It is advantageous if an optical filter is arranged in front of the image recorder which is tuned to a central wavelength of the laser. With this arrangement, interference from light sources which illuminate the object with a light whose wavelength deviates from the central wavelength of the laser is eliminated in a simple manner.

In an advantageous development, the laser is set up to project a scanning line with sinusoidally varied light intensity, the detection device being set up to generate a pixel-by-pixel Fourier transformation of an intensity signal and to determine an amount of a 1st harmonic of the respective Fourier spectrum. Inexpensive and efficient components such as FPGAs or GPUs are available for implementing appropriate evaluation algorithms.

Brief description of the drawings

[17] The invention is explained below by way of example with reference to the accompanying figures. It shows in a schematic representation:

Fig. 1 laser and image recorder over track object

2 block diagram of a measuring device FIG. 3 image series with corresponding image positions. FIG. 4 Fourier transformation of an intensity signal

Description of the embodiments

The arrangement shown in FIG. 1 illustrates the measuring principle of a laser light section sensor. A rail resting on a plate forms an object 1 of a track to be measured. A laser 3 and a digital image recorder 4 with a common reference base 5 are directed onto a surface 2 of the object 1. For example, both components 3, 4 of the laser light section sensor are arranged in a common housing (not shown) with an alignment that is predetermined with respect to one another.

[19] A scanning line 6 (laser line) is projected onto the plate and the rail by means of the laser 3. This is done, for example, by means of suitable optics such as a cylinder lens that fans out a laser beam accordingly. A light plane 7 defined by the fanning out forms a first angle a with the reference base 5 and intersects the object 1 to be measured. The resulting probe line 6 is interrupted several times by shadowing of the rail head and the rail foot.

The image recorder 4 is designed, for example, as a black-and-white digital camera and includes imaging optics and a high-resolution digital image processor. When the digital image processor is exposed, a raster graphic is created with a predetermined number of pixels 8 (picture elements). Each pixel 8 depicts a point on the object surface 2. In the case of a black-and-white recording, a brightness value W (gray value) is assigned to each pixel 8. The respective brightness value W corresponds to the amount of light which falls from the recorded location of the object surface 2 through the imaging optics onto the image processor during an exposure time. In the case of color recordings, color information is added pixel by pixel, which can also be evaluated.

[21] An optical filter is advantageously arranged in front of the image processor. The filter is tuned to the frequency of the laser light so that Disturbing light reflections with different frequencies are already eliminated at this point.

A camera axis 9 forms a second angle β with the common reference base 5. The dimensions of the object 1 can be derived from the recorded probing line 6 via the known angles α, β by means of triangulation. Specifically, the height of the object 1 is deduced from a characteristic deformation of the projected scanning line 6.

[23] In conventional applications, it can happen that the scanning line 6 is not clearly recognized in a recording because disruptive reflections from solar radiation or other sources of interference occur. This problem arises in particular when using a laser light section sensor in track construction because there are numerous light sources for illuminating a track to be measured and processed. Then, in addition to the desired scanning line, further structures become visible in the image, which make an evaluation difficult or impossible.

In the solution according to the invention, the scanning line 6 is projected onto the surface 2 with periodically varied light intensity. For example, a control unit 10 is provided which controls the laser 3 accordingly (FIG. 2). It is advantageous if the control unit 10 outputs a sinusoidal control signal which varies the intensity of the laser light accordingly. An angle function thus determines the curve shape of a corresponding intensity signal l (z).

The image recorder 4 according to the invention records a series 11 of images 12 during a period of the varied light intensity, which are then evaluated in a detection device 13. For this purpose, the image recorder 4 is set up accordingly before the start of a measurement, or the image recorder 4 is controlled accordingly by means of the control unit. In any case, the periodic variation of the light intensity of the laser 5 and the series recordings by the image recorder 4 are coordinated over time.

The present measuring device comprises, for example, an electronic circuit with FPGAs or GPUs in which the control unit 10, the Detection device 13 and optionally an evaluation device 14 for determining dimensions of the object 1 are integrated.

In the recorded image series 11, matching image positions P (x, y) are jointly evaluated in order to detect the image of the scanning line 6. 3 shows such an image series 11 as an image stack with an assigned coordinate system xyz. The position of a respective pixel 8 within an image 12 is determined with respect to an x-axis and a y-axis. The chronological sequence of the recorded images 12 is plotted on a z-axis. Thus, each image position P (x, y) is assigned an intensity signal l (z) for the course of the respective brightness value W recorded during an image series 11.

[28] Corresponding image positions P (x, y) are those pixels 8 which reproduce corresponding locations of the detected object 2 in the individual images 12. For this purpose, the image recorder 4 is advantageously aligned immovably with respect to the object 2 during the series recording. For example, the laser 3 and the image recorder 4 are supported on the object 2 during a series recording.

In the case of relative movements between object 2 and image recorder 4, the image positions P (x, y) within an image series are stabilized before an evaluation. For this purpose, appropriate stabilization software is set up in the detection device 13, for example. In addition or as an alternative to this, optical image stabilization is provided in the imaging optics of the image recorder 4. In this way it is ensured that the images 12 of a recorded series 11 reproduce identical object locations at corresponding image positions P (x, y).

4 shows an advantageous pixel-by-pixel evaluation of the image series 11. The starting basis for the evaluation is a respective profile of the intensity signal l (z) (brightness profile) for each image position P (x, y) of the image series 11 Signal curve over time and thus shown along the z-axis according to FIG. 3. For example, a brightness value W is determined for a corresponding pixel 8 with each image 11, and the course of the intensity signal l (z) is obtained by interpolation. A corresponding one comes into the detector device 13 Interpolation algorithm used. As an alternative to this, the image recorder 6 can be set up for the pixel-by-pixel output of the intensity signal l (z).

[31] In the next method step, a 1-D Fourier transformation takes place pixel by pixel, so that a Fourier spectrum 16 is available for each image position P (x, y). A corresponding result is shown in the diagram on the right in FIG. 4. The 1st Flarmonic Fl of the assigned Fourier spectrum 16 is of interest across all image positions P (x, y). The amount of the 1st Flarmonic Fl is particularly high for those pixels 8 whose intensity signal 1 (z) is sinusoidal runs.

[32] In a further process step, the amounts of the 1st Flarmonic Fl are therefore compared over all image positions P (x, y). A threshold value is expediently specified and only those pixels 8 are subsequently taken into account, the 1st Flarmonic F1 is above the threshold value. Such pixels 8 reproduce those locations of the object 1 which were illuminated by the laser 3 with a sinusoidal varied light intensity. All other pixels 8 with the same or slightly changed maturity are not taken into account and are set equal to zero, for example. The result of this evaluation is a merged image 17 in which only the depicted probe line 6 appears.

In an alternative form, not shown, the pixel-by-pixel intensity signal l (z) is filtered by means of a digital filter. The filter suppresses those signals that are outside of a narrow frequency band. The frequency band is determined by the frequency with which the light intensity of the laser 3 is varied periodically. Thus, only the depicted probe line 6 appears in the merged image 17 here as well.

The merged image 17 is transmitted from the detection device 13 to the evaluation device 14 of the laser light section sensor. The result of the evaluation are dimensions of the object 1, as described at the beginning with reference to FIG. 1.

Claims

1. A method for measuring an object (1) of a track by means of a laser light section sensor, a scanning line (6) being projected onto a surface (2) of the object (1) by means of a laser (3) and the surface (2) is recorded by means of a digital image recorder (4), characterized in that the scanning line (6) is projected with periodically varied light intensity, that the image recorder (4) records a series (11) of images (12) and that in the image series (11 ) matching image positions (P (x, y)) are jointly evaluated by means of a detection device (13) in order to detect an image of the scanning line (6).

2. The method according to claim 1, characterized in that the light received by the image recorder (4) is filtered by means of a narrow-band optical filter.

3. The method according to claim 1 or 2, characterized in that the image series (11) is recorded at least over a period of the periodically varied light intensity of the laser (3).

4. The method according to any one of claims 1 to 3, characterized in that matching image positions (P (x, y)) are jointly evaluated in at least three images (12).

5. The method according to any one of claims 1 to 4, characterized in that the light intensity is varied sinusoidally that to evaluate the image series (11) for the corresponding image positions (P (x, y)) a pixel-by-pixel Fourier transformation of an intensity signal (l (z)) that an amount of a 1st harmonic (H) of the Fourier spectrum (16) is determined for each pixel (8) and that the amounts of the 1st harmonic (H) are evaluated to map pixels of the scanning line (6) to be detected.

6. The method according to claim 5, characterized in that a threshold value is specified for the evaluation of the 1st harmonic (H) and that a uniform value for all pixels (8) whose amount of the 1st harmonic (H) is below the threshold value Brightness value (W) is specified.

7. The method according to any one of claims 1 to 4, characterized in that for the evaluation of the image series (11) for the matching image positions (P (x, y)) an electronic filtering of an intensity signal (l (z)) is carried out pixel by pixel and that the filtered intensity signal (l (z)) of the respective pixel (8) is evaluated in order to detect imaging pixels of the scanning line (6).

8. Measuring device for performing a method according to one of claims 1 to 7, characterized in that the laser (3) is set up for projecting a scanning line (6) with periodically varied light intensity, that the image recorder (4) for recording a series of images (11 ) and that the detection device (13) is set up for the simultaneous evaluation of image positions (P (x, y)) that match in the image series (11).

9. Measuring device according to claim 9, characterized in that an optical filter is arranged in front of the image recorder (4) which is matched to a central wavelength of the laser (3).

10. Measuring device according to claim 8 or 9, characterized in that the laser (3) is set up to project a scanning line (6) with sinusoidally varied light intensity and that the detection device (13) for forming a pixel-by-pixel Fourier transformation of an intensity signal (l ( z)) and is set up to determine an amount of a 1st harmonic of the respective Fourier spectrum (16).