WO2009112260A1

WO2009112260A1 - Device and method for detecting relative movements of an object with respect to a reference surface

Info

Publication number: WO2009112260A1
Application number: PCT/EP2009/001783
Authority: WO
Inventors: Ubbo Ricklefs; Stefan Charisse
Original assignee: Fachhochschule Giessen-Friedberg
Priority date: 2008-03-12
Filing date: 2009-03-12
Publication date: 2009-09-17
Also published as: DE102008013947B4; DE102008013947A1

Abstract

The invention relate to a device and a method for detecting a relative movement of an object (4) with respect to a reference surface (7). The relative movement is detected by means of n photo sensors (18) disposed on the object (4) in a row (20), which detects the brightness values of the reference surface (7). The n photo sensors (18) of the row (20) are divided into groups m_i having m photo sensors (18) each, which define a path s_i. For this purpose two consecutive groups m_i, m_i+1 m-1 belong to photo sensors (18).

Description

University of Applied Sciences Gießen-Friedberg Wiesenstraße 14 35390 Gießen

Apparatus and method for detecting relative movements of an object to a reference surface

The invention relates to a device for detecting relative movements of an object to a reference surface according to the preamble of claim 1 and a method for detecting relative movements of an object to a reference surface according to the preamble of claim 8.

Various optical measuring methods are known from the prior art with which the relative movements of an object relative to a reference surface can be determined without contact, such as the correlation method or the spatial frequency filter method.

According to the correlation method, a reference surface having a random structure is imaged onto an optical receiver, such as a camera. When the camera is moved relative to the reference surface, the image content changes from the content of the previous image. Via a so-called correlation analysis, the images can be compared with respect to suitable, extracted features of the reference surface structure, to determine the distance traveled by an object relative to the reference surface and optionally a rotation of the object. From the determined distance Δx, the corresponding velocity v can then also be determined over the time Δt between the two images and the simple relationship v = Δx / Δt.

EP 1 067 388 A1 describes a very complex correlation method of this type in which the correlation between the images of two cameras is calculated taking into account the triangulation. Such a correlation method is known to be associated with a relatively large amount of computation and therefore also with a relatively large computation time. Since the complete calculation of a correlation is time-consuming, approximations of the correlation are usually calculated.

By contrast, according to the spatial frequency filter method, which in contrast to the correlation method is relatively easy to implement by measurement, a reference surface with a random structure is imaged onto a photosensor and superimposed with a lattice-shaped periodic structure which is relative to a moving object relative to a reference surface is stationary and acts as spatial frequency filter which serves to mask the scattered light in the measurement volume. Moving the photosensor relative to the reference surface produces a narrow-band, velocity-dependent signal which, in addition to low-frequency changes, also has a modulation whose frequency f ₀ (maximum frequency of the temporal frequency spectrum) is a measure of the velocity of the object relative to the reference surface and has a clear relationship with the velocity along the orientation of the lattice structure, a lattice constant of the lattice structure and the magnification of an imaging optic and is proportional to the speed. A disadvantage of this method is that it fails for low speeds.

It is therefore an object of the present invention to provide an optical measuring method which places minimal demands on the measuring technology with regard to the computational outlay and the computation time and thereby detects with sufficient accuracy the relative movements of an object relative to a reference surface. Furthermore, the optical measuring method should also be usable for low speeds.

In addition, the invention has for its object to provide an apparatus for performing the method according to the invention. Main features of the invention are specified in the characterizing part of claim 1 and claim 8. Embodiments are the subject of claims 2 to 7 and 9 to 23.

In an apparatus for detecting relative movements of an object with respect to a reference surface, with n in-line photosensors arranged on the object detecting the brightness values of the reference surface irradiated with light of a particular wavelength, the invention provides that the n photosensors of the Line in groups m, are subdivided with each m photosensors, which define a distance s, where each two consecutive groups (m "m, ₊₁ ) m-1 photo sensors belong.

The device according to the invention therefore comprises n photosensors arranged in a line which, arranged on an object, detect the brightness values of a reference surface irradiated with a light of a specific wavelength via an objective. According to the invention, the n photo sensors of the line are interconnected or subdivided in such a way that they form groups m, with in each case m photosensors, which define distances s. Two consecutive groups m "m, ₊₁ belong to m-1 photo sensors together.

In the individual groups m, at least m = 3, preferably m = 4 or m = 6 photosensors are provided.

If significantly more than m = 3 _{photo sensors} per group m, are provided, these can be subdivided into summation groups.

Suitable photosensors include all known and suitable prior art receivers for optical radiation measurement, such as photodiodes or so-called CCDs (Charge Coupled Devices) and the like. The geometric design of the individual photosensors can be arbitrary.

In addition to photosensors, other sensor arrays are also suitable with which spatially resolved information about the surface can be obtained. These include in particular magnetoresistive or capacitive sensor arrays for detecting local magnetic or electrical field distributions. The sensors may be arranged at a distance successively in the row. However, such a successive arrangement of the photosensors is not necessarily required due to the random structure of the reference surface. Rather, the photosensors within the line may have any arrangement.

In a preferred embodiment of the invention, a plurality of lines each having n photo sensors are combined to form at least one matrix. Such an arrangement of the lines not only allows the detection of the relative movement of the object to the reference surface in the longitudinal direction of the lines, but also the detection of a rotational movement of the object. For this purpose, different movements are detected in staggered lines, wherein in the individual lines in each case the evaluation according to the invention takes place. If the sensors are arranged in a matrix, the evaluation described here for rows can also be done column by column.

The photosensors are preferably integrated in a user-specific IC (ASIC = Application Specific Integrated Circuit) or digital component which contains a circuit designed according to customer requirements.

The method according to the invention for detecting relative movements of an object relative to a reference surface, in which the brightness values of the reference surface irradiated with a light are detected by means of n photo sensors arranged on the object in a row, is provided,

^■ that the photosensors (18) of the line (20) are subdivided into groups JTi ₁ with m photosensors (18) each, defining distances S ₁ , with two successive groups (nrii, m _{i + 1} ) m-1 photosensors (18) belong together,

^■ that in a first position of the line (20) by means of the individual groups rrij the brightness values of the reference surface (7) are detected, and

^■ that after the detection in the individual groups rrij for a selected frequency f a modulation function is determined which describes the brightness values along the distance Sj, wherein with the adaptation of the modulation function to the

Brightness values a parameter pair Ij, φι is determined so that a maximum of n-m + 1 parameter pairs I ₁ , φ, which define a vector field V ₁ for the first position of the line (20), are determined for the line (20),

^■ that in a second position of the line (20) in an analogous manner, a vector field V _{1 + 1} with a maximum of n-m + 1 parameter pairs Ij, φ, for the second position is determined, and ^■ that the relative movement of the object (4) to the reference surface (7) is determined from the comparison of the vector field V, the first position with the vector field V _{J + 1 of} the second position.

In a first position of the line relative to the reference surface, the brightness values of the reference surface along the line are consecutively detected by means of the individual groups m ", starting with a group defining a starting point. The reference surface should be based on a random structure. For the electrical signals generated in the individual groups m, a modulation function with a frequency f is then selected and used in order to determine a function l (m,) which approximately describes the brightness values along the distance s. In this context, reference should be made at this point to the Fourier theory, which will not be discussed in detail below, x is a point of the distance S ₁ . For the k-th photosensor of the group m, x results as x = k * d _s , if ds corresponds to the width of the photosensor.

The modulation is therefore preferably carried out with the periodic function

l (m _l ) = l ₀ + l ₁ ^* sin (2π ^* f ^* x + φ _l ),

where x indicates the relative position of the photo sensor in the line.

For the selected frequency f we obtain for each group m, a pair of parameters I ₁ , φ "which can be considered as describing the magnitude and direction of a vector. For the row, a maximum of n-m + 1 vectors (I ₁ , φ,) ((I ₁ , Cp ₁ ), (I ₂ , φ _Σ ). (I3, <P3) .-} is determined, which is a vector field Define V _j for the first position j = 1 of the line.

Then, in a second position of the line relative to the reference plane, a vector field V _{J + 1} having a maximum of n-m + 1 parameter pairs (I ₁ , φ,) defining the second position of the line is determined in an analogous manner.

The relative movement of the object to the reference surface is then determined from the correlation of the two vector fields V _Jf V _{J + 1} . As vectors to be correlated come here the vectors (l "φ,), their amounts I ₁ , and / or their directions φ, in question. Accordingly, the method according to the invention provides that the vectors I _jj , cp ^ of the first vector field V _{j are} vectorially subtracted from the vectors I _j + u, (Pj + u of the second vector field V _{i + 1} generated subsequently, wherein the Amounts D _j j of the difference vectors or the difference vectors themselves are added to a total error D ₀ , and that at the same time for differentiation at least two further differences between the first vector field V _j and the second vector field V _{j + 1 are} performed, the differences D _j j in an analogous manner between the vector pairs (IJJ, φjj) and (IJ + U + _A, <PJ + _I, J + _A) and on the other between the vector pairs (I _j j, ψji) and (lj ₊ i, _{i +} A2. <Pj + i, i + A2) are formed in order to determine further total errors D _A and D _A2 ZU analogously to the total error D ₀ .

In addition, one can normalize the amount of the total error D ₀ by the number of difference terms. It is also possible to determine the total errors D ₀ (A = 0), D _A (A = A), D _A2 (A = A ₂ ) and possibly further total errors D (A _k ), where the total errors D as a function from A and determines the point A, for which D (A) becomes minimal.

Furthermore, it is advantageous if only the amounts l _j, j of the first vector field V _j are subtracted from the amounts l _{j +} i, i _{(+ A} / A2) of the second vector field V _{j + 1} , the amounts D jj of the differences to A total error D _{O / A / AÜ be} added, and wherein the distances A, A2 are determined depending on the time between the two detections and a shift between the groups to be compared mi the vector fields V _j and V _{j + 1} by A and A2 Photo sensors correspond.

The invention further provides that an average relative movement of the object to the reference surface is determined from at least three total errors D ₀ , D _A , D _A2 . Of the

Total error will be minimal for the true shift. The aim is therefore to determine the location of the minimum. As a first approximation one can assume, for example, that D ₀ , D _A , DA ₂ and possibly further total errors in the environment of the true

Shift describes a parabola and that the position of the vertex corresponds to the parabola of the displacement.

In addition, starting with the group nrij, from the partial directions φ, of the individual groups Im ₁ , the total direction φ which belongs to the relative movement and is proportional to this can be determined. The time between two detections of two successive groups is preferably dimensioned such that the overall direction φ changes on average during the transition from the vector field Vj to the vector field V _{j + 1} by less than +/- π.

At least m = 3 photosensors may be provided in the individual groups nij. One can also provide m = 4 or m = 6 photosensors in the individual groups rrij or one uses significantly more than three photosensors in the individual groups (Ti ₁₎ .

The method according to the invention represents a combination of the two measuring methods mentioned at the outset (correlation and spatial frequency filter methods). However, in contrast to these two methods, the method according to the invention requires significantly less computational effort, which leads to a significantly lower computing time. Furthermore, in contrast to the spatial frequency filter method, the method according to the invention is also suitable for detecting low relative speeds of the object relative to the reference surface.

Furthermore, in contrast to the two known methods, the method according to the invention can be implemented in a metrologically simple manner. This circumstance predestines the method according to the invention for various technical applications, especially as its metrological implementation can be implemented cost-effectively.

From the correlation of the two vector fields V _j , V _{j + 1} , two basic procedures for determining the relative movement can be derived, the procedure according to the difference method and the procedure according to the phase method. These two approaches can also be combined.

According to the invention, at least m = 3 photosensors are used according to the invention in the individual groups m.sub.ι resulting from the consideration of the aforementioned sine function, with which the brightness values of the individual groups n.sub.rii along the path S.sub.j are described.

The distribution of the brightness values along the distance S ₁ can be understood as an oscillation of the brightness values with the amplitude Ij around the mean value I ₀ according to the sine function, but as such for the method according to the invention because of the difference method which will be described below. underlying difference formation is ignored or is irrelevant. These Sinusoidal oscillation also has a phase shift φ ₍ compared to a zero point, which can be evaluated by the phase method.

If, for example, only m = 2 photo sensors defined the distance Sj, brightness values Ij but no direction values φ could be formed by the difference method as a result of the difference formation. In the case of m = 3 photosensors per group rrii, there is theoretically a change in the phase φ, on average in the transition from a group nrii to a subsequent group m _{i + 1} of 360 ° / m = 3, ie 120 °. Correspondingly, if m = 4 photosensors per group rrii theoretically there is a change of the phase φ, on the average 90 ° and for m = 6 photosensors a mean change of 60 °. To avoid or reduce the number of phase jumps, it may be advantageous over several

Periods too mittein.

In a preferred embodiment of the invention, several lines, each with n photo sensors, are combined to form at least one matrix in order, as already mentioned, to permit detection of a rotational movement of the object in addition to the detection of the relative movement of the object relative to the reference surface in the longitudinal direction of the lines.

In a further embodiment of the invention, it is provided that the displacements are determined separately from the rows of the matrix. This makes it possible to determine a rotational movement from the difference of the measured displacements.

If several frequencies are selected, a plurality of modulation functions can be determined which describe the brightness values along the distance Sj, wherein a parameter pair Ij, ψi is determined for each of the modulation functions by the adaptation to the brightness values.

If necessary, other sensors can be used instead of the photosensors, via which a local variation of the magnetic or electric field can be determined.

Further features, details and advantages of the invention will become apparent from the wording of the claims and from the following description of exemplary embodiments with reference to the drawings. Show it: 1 is a schematic representation of a measuring device which is mounted on a schematically illustrated vehicle,

FIG. 2 shows a schematic illustration of a multiplicity of photosensors arranged in a row and corresponding to the sensor shown schematically in FIG. 1, and FIG

Fig. 3 is a schematic representation of a plurality of arrayed in a matrix photosensors, wherein the individual rows and / or columns of the matrix correspond to the line shown in FIG.

Fig. 1 illustrates the implementation of a measuring device 16 according to the invention in a vehicle 4, which moves relative to a lane 6 in one direction at a speed v (t).

The device 16 according to the invention comprises a light source, not shown here for the sake of simplicity, which irradiates an area 9 of the road surface 7 at an angle. A converging lens 12 images the radiation reflected by the region 9 via a mirror 10 onto a photosensor 14.

According to the embodiment of the invention shown in FIG. 2, the sensor 14 comprises n optical receivers arranged in a row, which are designed as photodiodes 18 and are arranged successively at a defined spacing. The individual photodiodes 18 are connected in such a way that they form groups nrij with in each case m = 6 photosensors 18. Two consecutive groups rrii, m _{i + 1} belong to m-1 in each case, ie 5 photodiodes 18 together. This corresponds to a shift from group nrii to group m _{i + 1} along line 20 around a photodiode 18.

According to the embodiment of the invention shown in FIG. 3, the sensor 14 comprises n rows 20 combined into a matrix 22.

The two embodiments have in common that the photodiodes 18 are integrated on a user-specific IC or analog and / or digital component, a so-called ASIC (Application Specific Integrated Circuit), which contains a circuit designed according to customer requirements. According to an embodiment of the invention not shown here, the sensor 14 comprises a plurality of rows combined into a matrix.

The m photodiodes of the groups m, take the brightness values along a distance s on. Assigned to them is a spatial frequency f = 1 / s. The highest representable spatial frequency is obtained for m = 3. If larger m values are chosen to form the distances s, then the f values greater than 1 / s can be selected. This corresponds to averaging over several periods along the distance s. In this case the photodiodes are the

Lines are divided into summation groups m _ls , wherein for the selected spatial frequencies f optimal parameters I, and φ, are sought, the brightness distribution along the

Route s, overall optimal describe.

In the following, the method according to the invention for detecting the relative movement of the vehicle 4 to the roadway surface 7 will be described.

According to a first embodiment of the invention, which is based on the difference method, are in a first position of the line 20 relative to the road surface 7 by means of the individual groups m, (In ₁ , m ₂ , m ₃ , ...) starting with the group ITi ₁ at the leftmost end of the line 20, which as such defines a starting point, sequentially detecting the brightness values of the reference surface 7 along the line 20. Subsequently, for the electrical signals generated in the individual groups m ₁ , which represent the brightness values of the reference surface, by optimizing the parameters I ₁ and φ, a modulation with the periodic function

l (m,) = lo + 1, ^* sin (2π * f * x + φ,)

determined, which describes the brightness values along the distance s. The parameter pair I ₁ and φ can be understood as the amplitude and direction or phase of a vector. The parameter I ₀ is not further investigated. For each group m, a vector (I ₁ , φ ") thus results for the selected modulation frequency f, so that along the line 20 a maximum of n-m + 1 vectors (I ₁ , φ,) ((I ₁ , Cp ₁ ) , (I ₂ , φ ₂ ), (Ia, φa), • ■ (lιwn + 1. Φn-m + i)}, which define a vector field V ₁ = V ₁ for the first position of the line 20 ,

Subsequently, in a second position of the line 20, a vector field V _{1 + 1} = V ₂ with again maximally n-m + 1 vectors (I ₁ , φ,) ((I ₁ , Cp ₁ ), (I ₂ , φ ₂ ), (I ₃ , φ ₃ ), ... (I _n _-m + i,

Φ _n - _m + ₁ )} for the second position. Finally, from the correlation of the two Vector fields V _j1 V _{j + 1 determines} the relative movement of the object 4 to the reference surface 7 according to the difference method as follows.

In order to identify the affiliation of the vector (I ₁ , φO to the vector field V _j , the notation ( _{1 jj} _j j) is used in the following: The magnitudes I _j j of the first vector field V _j are determined by the amounts l _{j + 1} j of the second vector field V _{j + 1} is subtracted. then the amounts are added _j1 D of the differences to an overall error D ₀ and the

Number of difference amounts or difference terms normalized. Usually you will for the

Difference formation the next time vector field V _{j + 1} select, but you could also select a temporally much more distant vector field V _{j + k} .

Alternatively, the D _j i can also be formed by vectorially subtracting the vectors (I _j j, ψ _j j) from the vectors (l _{j + 1 | i} , φ _{j + 1ii} ) and then D _j i as the magnitude this difference vectors is selected. A disadvantage of this method is the higher computational effort.

At the same time as this difference formation, further differences are formed between the first vector field V _j and the second vector field V _{j + 1} . While in the first subtraction the vector pairs (I _j1 , cp _j j) and (l _{j + 1ii} , φ _{j + 1} , j) were compared with each other, the differences D _j i according to the described methods now between the vector pairs (I _j j, φ _jti ) and (IJ + U + _A , ΦJ _{+ I} J _{+ A} ) to determine a total error D _A analogous to the total error D ₀ . This corresponds to a comparison of the group (Ti _{1 of} the vector field V _j with the group m _{i + A of} the vector field V _{j + 1} , which is shifted by the group nrii by A photodiodes.

In parallel, this calculation is repeated for at least one further shift A2 to determine an additional total error D _A2 ZU. Unless it is unknown which sign the shift has, then you will usually choose A2 = -A.

The distance A is determined as a function of the size of the relative movement performed between the two detection times and corresponds to a shift of the groups rτii of the first vector field V _j with respect to the vector field V _{j + 1} by A photo sensors 18.

From the total errors D ₀ , D _A , D _A2 and possibly further total error values D (A _R ), an average relative movement of the object 4 to the reference surface 7 finally becomes certainly. If one plots the total errors D ₀ (A = O), D _A (A = A), DA ₂ (A = A ₂ ) and possibly further total errors D (A _k ) as a function of A, one obtains a function which in the vicinity of the point A = O shows a parabolic course. The vertex of the parabola, which as such can be determined by means of at least three points, corresponds to said average relative motion in units of diode width. By means of the geometry of the photodiodes used and their optical imaging, the distance is known that corresponds to a diode width. This allows the displacement Δx to be calculated.

The relative movement determined in this way corresponds to the movement of the sensor 14 in the longitudinal direction of the line 20. If several lines are used, then from this movement after the vector calculation movements at an angle to the line length direction can be determined.

By means of the time Δt between the two positions V ₁ , V ₂ , the average velocity v in the longitudinal direction of the line 20 can then also be determined via the simple relationship v = Δx / Δt.

According to the embodiment of the sensor 14 in the form of a matrix shown in FIG. 3, not only relative movement of the object 4 to the reference surface 7 in the longitudinal direction of the lines, but also rotational movement of the object 4 can be detected.

After the phase method, the change of the phases ψ _j j is evaluated, which results from a shift. In the transition from a group nrii to the group m _{i + 1} , the magnitude of the vector l _j , j changes only slightly, since both groups m-1 photodiodes belong together. Its direction φ _jti, on the other hand, changes on _average during the transition from the group rτii to the group m _{i + 1} by the angle 2ττ / m. Due to the 2τr periodicity of the sinusoidal function phase jumps can also occur in the determination of the direction cpj.j, which can lead to ambiguities. The phase φjj is determined except for k * 2 ^* π (k = ... -1, 0, 1, 2, ...). By averaging over all groups nrii over all phases φ _{j: i} such that the mean phase change between two successive groups m ₍ and m _{i + 1} is taken into account as well as any phase jumps present, one can then use the vector field and thus the Position assign a total phase φ, which is a measure of the position.

If one measures the time span between the two successive vector fields V _j and V _{j + 1} so short that the displacement with respect to the reference surface is small, then also the phases φ _ui and φ _{j + 1ji differ} only slightly. The shift then corresponds to the phase difference φ _{j + 1,} j-cpy. In order to reduce the measurement errors, it is expedient to again agree on all phase differences. Again, ambiguities may occur again. Therefore, the time between two detections of two successive vector fields V _j and V _{j + 1 should be} such that the directions φ _ui and φ _{j +} i, i at the transition from the vector field V _j to the vector field V _{j + 1} on average to change significantly less like +/- π.

According to a further embodiment of the invention, which is based on the phase method, starting with the group In ₁ at the extreme left end of the line 20, which defines as such a starting point along the line 20, the partial directions φ, the individual groups nrij to the overall direction φ added up. Since the overall direction or

Total phase φ is proportional to the relative motion can from the overall phase φ the

Relative movement are determined, if the belonging to a relative movement phase change is known.

According to a further embodiment of the invention, the difference method and the phase method are combined with each other.

Both the difference method and the phase method are technically easy to implement.

The invention is not limited to one of the above-described embodiments, but can be modified in many ways. It can be seen, however, that the invention relates to a device and a method for detecting a relative movement of an object 4 relative to a reference surface 7. The relative movement is detected by means of n photosensors 18 arranged on the object 4 in a row 20, which detect the brightness values of the reference surface 7. The n photosensors 18 of the line 20 are subdivided into groups rτii with in each case m photosensors 18 which define a route S ₁ . In each case, two consecutive groups nrii, m _{i + 1} m-1 belong to photo sensors 18.

All of the claims, the description and the drawings resulting features and advantages, including design details, spatial arrangements and method steps may be essential to the invention both in itself and in various combinations. Reference number list

2 Arrangement of vehicle and measuring device

4 vehicle

6 lane

7 road surface

8 reflected light beam

9 picture area

10 mirrors

12 condenser lens

14 photosensor

16 measuring device

18 photodiode

20 photodiode lines

22 photodiode matrix

Claims

claims

1. A device for detecting relative movements of an object (4) to a reference surface (7), with n in a row (20) arranged photosensors (18) arranged on the object (4), the brightness values of the irradiated with a light Detecting reference surface (7), characterized in that the n photosensors (18) of the line (20) are subdivided into groups m, each having m photosensors (18) defining a distance s, wherein in each case two successive groups (m "M, ₊₁ ) belong to m-1 photo sensors (18).

2. Apparatus according to claim 1, characterized in that in the individual groups m, at least m = 3 photosensors (18) are provided.

3. Apparatus according to claim 2, characterized in that in the individual groups m, m = 4 or m = 6 photosensors (18) are provided.

4. Apparatus according to claim 2, characterized in that in the individual groups m, substantially more than m = 3 photosensors (18) are provided and that the photosensors (18) of the respective groups m, in summation groups m, _{s are} divided.

5. Device according to one of claims 2 to 4, characterized in that the photosensors (18) are arranged successively at a distance.

6. Device according to one of claims 1 to 5, characterized in that a plurality of lines (20) are combined with each n photosensors (18) to at least one matrix (22).

7. Device according to one of claims 1 to 6, characterized in that the photosensors (18) are integrated on a user-specific IC (ASIC).

8. A method for detecting relative movements of an object (4) to a reference surface (7), in which by means of n on the object (4) in a row (20) arranged photosensors (18) the brightness values of the irradiated with a light (8) Reference surface (7) are detected, characterized in that ■ the photosensors (18) of the line (20) in groups m, each with m photosensors (18) are divided, the distances s, define, each two consecutive groups (m "M _{l + 1} ) belong to m-1 photo sensors (18),

■ that in a first position of the line (20) by means of the individual groups m, the brightness values of the reference surface (7) are detected and ■ that after the detection in the individual groups m, a modulation function is determined for a selected frequency f, which the Brightness values along the distance s, describes, with the adjustment of the modulation function to the brightness values, a parameter pair l "φ, is determined, so that for the line (20) a maximum of n-m + 1 parameter pairs I ₁ , φ, are determined define a vector field V _j for the first position of the line (20),

■ that in a second position of the line (20) in an analogous manner, a vector field V _{1 + 1} with a maximum of n-m + 1 parameter pairs I ₁ , φ, for the second position is determined, and

■ that the relative movement of the object (4) to the reference surface (7) is determined from the comparison of the vector field V _{1 of} the first position with the vector field V _{J + 1 of} the second position.

9. The method according to claim 8, characterized in that the modulation with the function l (m,) = I ₀ + I, * sin (2ττ * f * x + φ,) is described, where x is the relative position of the photosensor in the line indicates.

10. The method according to claim 8 or 9, characterized in that the vectors I _{j 1} , (P _j1 , the first vector field V ₁ vectorially of the vectors I _{1 + 1 1} , φ _{J + 1ι} , the second vector field subsequently generated, the second vector field V _{1 + 1 are} subtracted, wherein the amounts D ₁ , the difference vectors or the difference vectors themselves are added to a total error D ₀ , and that at the same time for difference formation at least two other differences between the first vector field V ₁ and the second vector field V _{1 + 1} with the differences D ₁ , in an analogous manner between the vector pairs (I ₁ , φ _Jh ) and (I _{j + 1} , _{I + A} , <P _{J + I} , _I + _A ) and, secondly, between the vector pairs (I ₁ ,, φ _μ ) and (I _{1 + 1} , _{I + A} 2, φ, + i, ι + _A2 ) are formed in order to determine analogous to the total error D ₀ further total errors D _A and D ^.

11. The method according to claim 10, characterized in that the amount of the total error D ₀ is normalized by the number of difference terms.

12. The method according to claim 10 or 11, characterized in that the total error D ₀ (A = 0), D _A (A = A), DA ₂ (A = A ₂ ) and possibly further total errors D (AK) are determined , where the total error D is considered as a function of A and the point A is determined for which D (A) becomes minimal.

13. The method according to any one of claims 10 to 12, characterized in that only the amounts l _{Jfl of} the first vector field V _j of the amounts I _{J + I} , _{I (+ A /} A ₂ ) of the second vector field V _{1 + 1 are} subtracted , wherein the amounts D ₁ , the differences are added to a total error D _{0 /} A _/ A2.

14. The method according to any one of claims 10 to 13, characterized in that the distances A, A2 depending on the lying between the two detections

Time and a shift between the groups to be compared m, the vector fields V ₁ and V _{1 + 1} by A and A2 photosensors (18) correspond.

15. The method according to claim 14, characterized in that from at least three total errors D ₀ , D _A , D _A2 an average relative movement of the object (4) to the reference surface (7) is determined.

16. The method according to any one of claims 8 to 15, characterized in that, starting with the group m "from the directions φ, the individual groups m, the overall direction φ is determined, which belongs to the relative movement and is proportional to this.

17. The method according to any one of claims 8 to 16, characterized in that the time lying between two detections of two successive groups is such that the overall direction φ in the transition from the vector field V, to the vector field V _{J + 1} in the middle changes by less than +/- π.

18. The method according to any one of claims 8 to 17, characterized in that in the individual groups m, at least m = 3 photosensors (18) are provided.

19. The method according to claim 8 to 18, characterized in that in the individual groups m, m = 4 or m = 6 photosensors (18) are provided.

20. The method according to claim 8 to 19, characterized in that in the individual groups m, substantially more than three photosensors (18) are provided.

21. The method according to any one of claims 18 to 20, characterized in that the photosensors (18) are divided into summation groups m _ιs .

22. The method according to any one of claims 8 to 21, characterized in that a plurality of lines (20), each with n photo sensors (18) are combined to form at least one matrix (22).

23. The method according to any one of claims 8 to 22, characterized in that in the rows (20) of the matrix (22) separated from each other, the shifts are determined.

24. A method according to any one of claims 8 to 23, characterized in that a plurality of frequencies are selected to determine a plurality of modulation functions describing the brightness values along the distance s, wherein for each of the modulation functions by adjusting to the brightness values

Parameter pair I ₁ , φ, is determined.

25. The method according to any one of claims 8 to 24, characterized in that instead of the photosensors other sensors are used, via which the local variation of the magnetic or electric field is determined.