DE102014206851A1 - Arrangement and method for detecting a one-dimensional position and speed of an object with a camera - Google Patents

Abstract

The invention provides an arrangement for detecting a one-dimensional position and a one-dimensional speed of a movable object (1). The arrangement comprises: a marker device (2) formed in the object (1) or arranged on the object (1) and designed to make the position and the speed detectable, a camera unit detecting the marker device (2) ( 3) which is designed to record images of the marker device (2), - an evaluation unit (4) which is designed to determine the position and the velocity from the acquired images, - a strip-shaped first gray value wedge (5) of the marker device ( 2) and - a strip-shaped second gray scale wedge (6) of the marker device (2) parallel to the first gray scale wedge (5), wherein the first gray scale wedge (5) has a periodic first gray level profile and the second gray scale wedge (6) a periodic second gray level profile having. An associated method is also given. The invention offers the advantage of being able to measure the position and the velocity of moving objects one-dimensionally with a simple and inexpensive construction using inexpensive webcams and standard image evaluation methods.

Description

Field of the invention

The invention relates to an arrangement and a method for detecting a one-dimensional position and a one-dimensional speed of a movable object by means of image acquisition and image evaluation of a marker device.

Background of the invention

• – schienengeführten Fahrzeugen,
• – Fördertechnik oder
• – bildgebende Geräten, z. B. in der medizinischen Diagnostik.

High-precision one-dimensional localization of objects is an important prerequisite for many applications. Examples are the position detection of

• - rail-guided vehicles,
• - Conveyor technology or
• - Imaging devices, eg. B. in medical diagnostics.

The localization must allow a high accuracy during movement of the arrangement. It must also be inexpensive to implement.

The metrological detection of a one-dimensional position was done so far, depending on the application, for example by the use of potentiometers, incremental encoders, camera systems or laser metrology. Until now systems could be realized with a high expenditure, which enable a very exact position detection even during a movement.

Applications of position capturing exist, for example, in X-ray imaging. In the post-published patent application DE 10 2013 210 543 A1 is an aperture plate of a depth stop of an X-ray imaging device provided with differently designed coding marks or reference marks, so that by means of sensors, the current displacement position of the aperture plate can be quickly determined. Any step losses of a stepping motor moving the aperture plates are detected and corrected if necessary.

Summary of the invention

It is an object of the invention to provide an arrangement and a method for detecting a one-dimensional position and a one-dimensional speed of an object, which are very accurate, robust and cost-effective.

According to the invention, the stated object is achieved with the arrangement and the device of the independent claims. Advantageous developments are specified in the dependent claims.

According to the invention, the absolute position and the absolute speed of a movable object are determined with the aid of image acquisition and the image evaluation of gray scale wedges connected to the object, which have periodic gray scale gradients. In this case, in particular for the position determination, the phase position of the gray value profiles is evaluated relative to one another and for the speed determination the amplitude reduction of the gray value profile caused by movement.

The invention claims an arrangement for detecting a one-dimensional position and a one-dimensional velocity of a movable object. The arrangement comprises a marker device embodied in the object or arranged on the object, which is designed to make the position and the speed detectable, a camera unit which detects the marker device and which is configured to generate images of the marker device An evaluation unit, which is designed to determine the position and the speed of the detected images, a strip-shaped first gray value wedge of the marker device and a strip-shaped, parallel to the first gray scale wedge second gray value wedge of the marker device, wherein the first gray value wedge a periodic first gray level profile and the second gray value wedge has a periodic second gray level profile.

The invention also claims an arrangement for detecting a one-dimensional position and a one-dimensional velocity of a movable object. The arrangement comprises a marker device which is designed to make the position and the speed detectable, a camera unit which detects the marker device and which is arranged to receive images of the marker device, an evaluation unit which is designed to record the position and the Determine speed from the captured images, a strip-shaped first gray value wedge of the marker device and a strip-shaped, parallel to the first gray value wedge second gray value wedge of the marker device, wherein the first gray value wedge has a periodic first gray level profile and the second gray value wedge a periodic second gray level profile.

The invention offers the advantage of being able to measure the position and the velocity of moving objects one-dimensionally with a simple and inexpensive construction using inexpensive webcams and standard image evaluation methods. The gray value wedges can be created with standard printers.

In a further development of the invention, the first gray value curve is formed sinusoidally with a first wavelength and the second gray level curve is formed sinusoidally with a second wavelength.

In a further embodiment, the number of first wavelengths forming the first gray value wedge is greater by one than the number of second wavelengths forming the second gray value wedge. As a result, an absolute position determination is possible.

In a further embodiment, the first wavelength is selected such that one to two first wavelengths can be displayed in the image of the camera unit. This ensures that a movement does not lead to non-reconstructable overlays.

In a development, the arrangement comprises a first reference strip parallel to the first gray value wedge and configured to indicate a darkest gray value, and a second reference strip parallel to the first gray value wedge and designed to indicate a brightest gray value.

In a further embodiment, the first reference strip is black and the second reference strip is white.

Furthermore, the arrangement may have a coding pattern strip parallel to the first gray value wedge, which has squares with a code formed by dots along its direction of movement.

In addition, the marker means may consist of a printed tape arranged along the direction of movement of the object.

The invention also provides a method for detecting a one-dimensional position and a one-dimensional velocity of a movable object, wherein from a arranged in the direction of movement of the object first and a second gray value wedge with a periodic first gray value progression or a periodic second gray level profile by means of image acquisition and image analysis, the position and determine the speed.

In a further development of the method, the first gray value profile is formed sinusoidally with a first wavelength and the second gray value profile is sinusoidal with a second wavelength.

Furthermore, the image analysis can have the Görtzel algorithm, by which the periodic first and second gray level characteristics are reconstructed.

In addition, the position can be determined from the phase position of the first gray level profile to the second gray level profile.

In a development, the speed is determined from the amplitude and phase of the first gray value profile and / or of the second gray value profile.

In a further embodiment, the speed is determined by means of the rolling shutter effect of a digital camera executing the image acquisition from the image distortion caused by movement.

Other features and advantages of the invention will become apparent from the following explanations of an embodiment with reference to schematic drawings.

Show it:

1 FIG. 3 is a block diagram of an arrangement with a marker device, FIG.

2 : a top view of a marker device,

3 FIG. 2: a diagram of the transfer functions of one-dimensional rectangular filters,

4 : a graph of gray scale gradients with and without movement,

5 Two graphs of measured grayscale gradients with and without motion and

6 : A view of a moving marker device with rolling shutter effect.

Detailed description of an embodiment

1 shows a block diagram of an arrangement for detecting a one-dimensional position and a one-dimensional velocity of a movable object 1 , In the simplest case, the object indicates 1 a camera unit 3 and an evaluation unit connected thereto in terms of data transmission 4 on. Optionally, the evaluation unit 4 via a radio interface from the object 1 be located away. The object 1 moves one-dimensionally in the direction of movement 11 ,

Parallel to the direction of movement 11 is a marker device 2 arranged, which is formed in the simplest case as a band. The marker device 2 has inventive pattern, the of the camera unit 3 as pictures are taken. By means of image processing in the evaluation unit 4 The patterns are analyzed and from this the absolute position and the absolute speed of the object 1 determined in one dimension. The object 1 and the marker device 2 For example, they may be part of an imaging medical device.

It is also possible to use the marker device on the movable object 1 to install and the camera unit 3 and the evaluation unit 4 to leave unmoved. In the further considerations, it depends only on the relative movement.

2 shows a section of a marker device 2 Following in an order 1 Can be used. The marker device 2 For example, it consists of a printed band in the direction of movement 11 of the object is arranged. The marker device has a strip-shaped first cooking step wedge 5 and a strip-like second gray scale wedge parallel thereto 6 which can be used to determine position and velocity in one dimension. The gray value wedges 5 and 6 encode the distance over the complete length of a movement.

The first grayscale wedge 5 has a sinusoidal first gray value curve (brightness curve). The second gray scale profile has a sinusoidal second gray level profile (brightness profile). The first gray scale profile has a first wavelength λ ₁ and the second gray scale profile has a second wavelength λ ₂ . The two wavelengths λ ₁ and λ ₂ are selected so that one to two wavelengths (= wavenumber) fit into a captured image of a camera. For example, if the image width is 300 pixels, then the wavelengths are 150 to 300 pixels.

In order to determine the absolute position clearly, the phase position between the first grayscale wedge 5 and the second grayscale wedge 6 be ambiguous over a whole measuring length. Therefore, the first and the second wavelength λ ₁ , λ _{2 are} slightly different, so that over the measuring length, the number of first wavelengths λ ₁ is one greater than the number of second wavelengths λ ₂ . In a superimposition of the first with the second gray scale profile then results in a beat whose beat wavelength λ _{S is} about as large as the measuring length.

In order to accurately determine the darkest and the brightest grayscale value in each captured image, the marker means 2 over the measuring length still a black first reference strip 8th and a white second reference strip 9 on, both parallel to the grayscale wedges 5 and 6 are formed. The white and black stripes 8th . 9 guarantee that minimum and maximum value at any time from one Any image acquisition are to be determined. The basic idea here is that a linear conversion between gray value and distance is possible.

Optionally, the position may also be with markers of a coding pattern strip 10 be determined. For reasons of computational efficiency, geometric primitives such as squares or circles are available as markers. For example, the markers consist of sixteen black square elements framed by a black frame. Each marker represents a binary information. With 16 elements this results in 65536 possible combinations. These can be detected without much computational effort. The calculation of the center and the alignment of squares is very easy to implement.

The absolute position of each square is known in advance and can be calculated using a formula. The system offset is a fixed distance and is determined during calibration of the measuring system. The markers are detected in the camera image and the pixel coordinates of the center points are determined. From the coordinates of the centers and the horizontal spatial resolution of the camera, the unique system position can be calculated. This is calculated for all markers detected in an image and then averaged. For plausibility checking, it is checked whether the detected markers lie in a continuous sequence.

Functions of an OpenCV library can be used for the detection of the markers. Here, the camera image is converted into a binary image and then the markers located in the image are searched for their shape (square). The decoding of the found markers takes place via an implemented, as simple as possible function. Here, in a part of the binary image, a search mask searches for the elements in the marker to identify it.

Moving while shooting with a camera creates motion blur. The movement model used is one-dimensional. In a continuous motion, the blur effect is similar to using a 1D rectangular filter. However, the noise reduction is not given because camera-specific sources of error (eg dark noise) can not be smoothed with it. Depending on the speed, the width of the rectangular filter is defined as: R _1xn = [1 1 1 ... 1] / n (2) where n is the number of pixels by which the image is shifted during recording.

The transfer function of this filter is obtained by means of discrete Fourier transformation to: R _1xn = sin (πnκ / 2) / (n · sin (πκ / 2) (3) where κ represents the wavenumber:

3 shows a graph of transfer functions of different speeds. The attenuation behavior of rectangular filters shows significant oscillations. The pixel offset during an aperture of the camera is simulated for four cases. A monofrequent signal with exactly three samples per period is completely suppressed.

For a reconstruction of a signal, therefore, its first zero crossing from the first zero crossing of the transfer function must be sufficiently spaced. The associated frame rate during a very fast movement with 50 pixels (pixel) offset during the aperture of the camera and a picture width b = 320 corresponds to: f = 320 / 50b ^-1 ~6.4b ^-1 . (4)

Therefore, the highest frequency of a pattern must be repeated a maximum of six times so that it can be reconstructed at the assumed speed.

For spectral analysis of a time series, the discrete Fourier analysis is usually used. In the present case, however, the frequency used is precisely known. We are only looking for the amplitude and the phase shift of a frequency. This can be calculated efficiently with the Görtzel algorithm. With N samples, a time series {x (k) | k = 1 ... N} is to be reconstructed as follows: w _n (k) = x (k) + 2 cos (2πn / N) × w _n (k - 1) - w _n (k - 2) (5) _n y (n) = w _n (N) - e ^{-f2πn / N} × _n w (N - 1) (6) where y _n (n) represents the reconstructed complex frequency-specific pointer. 4 shows this relationship in a diagram. You can see a sine signal without motion n = 0 with amplitude 1 and the same sine signal with a motion n = 50.

For a uniform motion that has n = 50 pixel skew during the aperture, the Görtzel algorithm yields an attenuation of 0.675. From this amplitude reduction and the phase shift, the velocity during recording can be reconstructed mathematically using equations (5) and (6).

5 shows two graphs A and B of measured sinusoidal gray scale gradients (brightness gradients) of gray scale wedges printed with different printers. The curve K1 shows the measured gray value profile and the curve K2 the gray value profile reconstructed from the curve K1 with the aid of the Görtzel algorithm. Diagram A shows the gray value curve of a laser printer and diagram B shows the gray level profile of an inkjet printer. The curves K1 show nonlinearities in the gray value curve due to the nonlinear characteristics of printers.

It can clearly be seen that the characteristic of a gray value wedge can be reconstructed on the basis of the gray value curve. Phases and amplitudes of both sinusoidal signals K1 can be determined robustly with the Görtzelfilter. Since the phase of a sinusoidal signal does not allow absolute localization, this method must be combined with other measures, such as 2 described by the use of a first and a second gray value wedge 5 . 6 ,

6 shows the consequences of a movement of the marker device 2 to 2 , The movement of the camera or the marker device 2 during a measurement makes pattern localization difficult. In addition to a motion blur occurs in some cameras, the so-called rolling-shutter effect.

The rolling shutter effect is a positional error in images that may occur due to a technical phenomenon in the line-by-line or column-wise photographic shooting of moving images. It first appears obvious that the exposure of all points of the photosensitive sensor of a digital camera begins at exactly the same time. But there are cameras where the design does not apply to the entire surface. There, one has to imagine a "line of simultaneous exposure starts", which moves either line by line or column by column over the sensor, for which a short but nevertheless not negligible time is needed.

As long as the subject is motionless, even with such cameras, all pixels are exposed at their correct position, regardless of when the corresponding points were exposed. However, when taking pictures of a moving subject, or moving camera, with consecutively exposed lines, objects are displayed in their current location. Since the mapping is done line by line, the object has already moved from one line to the next, so that in the case of a composition of all lines, an object is displayed that was not displayed as a whole at once, but at different times line by line. A straight line of the subject can be crooked or distorted with a camera pan or a moving subject (corresponding to the successive individual lines).

In the case of digital cameras, the rolling shutter effect typically occurs in the case of image sensors in CMOS sensor technology, since these generally allow only one line or column-wise signal evaluation.

6 shows the resulting distortion of the overall picture (line-dependent shift). The rolling shutter effect can be used to determine the speed. All that needs to be done is the distortion in the direction of movement 11 during the shutter opening time.

LIST OF REFERENCE NUMBERS

1

object

2

Marker device

3

camera unit

4

evaluation

5

first gray value wedge

6

second gray value wedge first / second gray value curve

8th

first reference strip (black)

9

second reference strip (white)

10

Codiermustersteifen

11

Direction of movement of the object 1

K1

Curve of the measured values of the gray value curve

K2

Curve of the reconstructed gray value curve

n

Offset of the pixels during the aperture of the camera unit 3

λ ₁

first wavelength

λ ₂

second wavelength

λ _S

Beat wavelength

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• DE 102013210543 A1 [0005]

Claims (15)

Arrangement for detecting a one-dimensional position and a one-dimensional velocity of a movable object ( 1 ), comprising: - one in the object ( 1 ) or on the object ( 1 ) arranged marker device ( 2 ), which is designed to make the position and the speed detectable, - the marker device ( 2 ) detecting camera unit ( 3 ), which is designed to store images of the marker device ( 2 ), and - an evaluation unit ( 4 ), which is designed to determine the position and the speed from the acquired images, characterized by: - a strip-shaped first gray value wedge ( 5 ) of the marker device ( 2 ) and - a strip-shaped, to the first gray value wedge ( 5 ) parallel second gray value wedge ( 6 ) of the marker device ( 2 ), - wherein the first gray value wedge ( 5 ) a periodic first gray value curve and the second gray value wedge ( 6 ) has a periodic second gray level profile. Arrangement for detecting a one-dimensional position and a one-dimensional velocity of a movable object ( 1 ), comprising: - a marker device ( 2 ), which is designed to make the position and the speed detectable, - the marker device ( 2 ) on the object ( 1 ) arranged camera unit ( 3 ), which is designed to store images of the marker device ( 2 ), and - an evaluation unit ( 4 ), which is designed to determine the position and the speed from the acquired images, characterized by: - a strip-shaped first gray value wedge ( 5 ) of the marker device ( 2 ) and - a strip-shaped, to the first gray value wedge ( 5 ) parallel second gray value wedge ( 6 ) of the marker device ( 2 ), - wherein the first gray value wedge ( 5 ) a periodic first gray value curve and the second gray value wedge ( 6 ) has a periodic second gray level profile. Arrangement according to claim 1 or 2, characterized in that the first gray value curve is sinusoidal with a first wavelength and the second gray value curve is sinusoidal with a second wavelength. Arrangement according to claim 3, characterized in that the number of the first gray scale wedge forming first wavelengths (λ ₁ ) is greater by one than the number of the second gray value wedge forming second wavelengths (λ ₂ ). Arrangement according to claim 4, characterized in that the first wavelength (λ ₁ ) is selected such that one to two first wavelengths (λ ₁ ) in the image of the camera unit ( 3 ) are representable. Arrangement according to one of the preceding claims, characterized by: - one to the first gray value wedge ( 5 ) parallel first reference strip ( 8th ), which is designed to indicate a darkest gray value, and - one to the first gray value wedge ( 5 ) parallel second reference strip ( 9 ) which is designed to indicate a brightest gray value. Arrangement according to claim 6, characterized in that the first reference strip ( 8th ) black and the second reference strip ( 9 ) is white. Arrangement according to one of the preceding claims, characterized by: - a first gray value wedge ( 5 ) parallel coding pattern strips ( 10 ), which along its direction of movement ( 11 ) Has squares with a code formed by dots. Arrangement according to one of the preceding claims, characterized in that the marker device ( 2 ) consists of a printed tape, which along the direction of movement ( 11 ) of the object ( 1 ) is arranged. Method for detecting a one-dimensional position and a one-dimensional velocity of a movable object ( 1 ), characterized in that from one in the direction of movement ( 11 ) of the object ( 1 ) arranged first and second gray value wedge ( 5 . 6 ) are determined with a periodic first gray level course and a periodic second gray level profile by means of image acquisition and image evaluation, the position and the speed. A method according to claim 10, characterized in that the first gray value curve is sinusoidal with a first wavelength (λ ₁ ) and the second gray level curve sinusoidal with a second wavelength (λ ₂ ). A method according to claim 10 or 11, characterized in that the image analysis comprises the Görtzel algorithm, with the help of which the periodic first and second gray value course (K2) are reconstructed. Method according to one of Claims 10 to 12, characterized in that the position is determined from the phase position of the first gray value profile to the second gray value profile. Method according to one of claims 10 to 13, characterized in that the velocity is determined from the amplitude and phase of the first gray level profile and / or the second gray level profile. Method according to one of claims 10 to 13, characterized in that by means of the rolling shutter effect of a digital camera unit which carries out the image acquisition ( 3 ) from the movement ( 11 ) caused image distortion, the speed is determined.

Images