DE19500507C2 - Camera with lens and image carrier adjustment device and focusing method - Google Patents

Description

The invention relates to a focusing method of a camera and this itself, as known from DE 34 33 412 C2,

• - with a lens holder and an image holder with a connecting bellows that are relative to each other in a Bench direction are kept variable, the corresponding changes in distance fed to a processor will,
• - And in their lens holder a lens by two axes perpendicular to one another and perpendicular to the bank direction is held pivotable, both of which are accordingly Lens tilt angle values and the value of Lens focal length are fed to the processor,
• - And in their image carrier an image sensor level, with a film or a matte screen can be populated by two axes perpendicular to one another and perpendicular to the bank direction is held pivotable, both of which accordingly Image sensor swivel angle values are fed to the processor will,
• - and in their image sensor level Focus sensor that sends a focus signal supplies, can be positioned in two image sensor coordinates is the processor along with the Sharpness signal are supplied
• - According to the procedure for the simultaneous Focus at least three selected pixels, which are in an initial image plane,  
• - This on the image sensor level with the Focus sensor by changing the Image carrier distance from the lens one after the other in their focus taking the associated Image sensor coordinates and the Changes in image carrier spacing are detected in the position of the processor will,
• - And from the input data thus obtained from the processor Position parameters of the image sensor level and the Output image level can be determined mathematically and by Solution of mapping equations such target values for pivoting the lens holder and one Distance adjustment of the image carrier to the lens carrier calculated and output,
• - so that the set with these setpoint values new image level with the image pickup level if possible Cover comes.

In this known camera, the lens focal length, an imaging scale related to the imaging optics Space coordinates of three selected pixel areas a scene, for which a focus probe as an aid serves, adjustment angle values of the lenses and Image plane inclinations and swivels and the distance measure a main lens plane from the image plane to the processor either by keyboard or directly with measuring equipment supplied, which deviates from the subject matter of the invention this measurement data according to the Scheimpflug law, accordingly the lens plane, the plane of the scene points belong to and the level of the sharp pixels belong to have to cut in a straight line, Target values of the adjustment angle and the distance calculated so that after a made with it Setting the selected pixel areas of the Scene image with so-called double focus compensation, d. H. a focus compensation by swiveling the Lens or the imaging plane in two perpendicular  directions that are as sharp as possible were. So with this device the absolute Spatial coordinates of the pixels and the structural data of the Imaging optics as initial values for a closed Solving the mapping equations required to each such a location of the mapping level in relation to the Find lens setting in which the selected Pixel areas are in the imaging plane, if possible are clearly shown. This required a considerable amount Effort of measuring equipment with high absolute Accuracy requirements. In addition, this applies so far used Scheimpflug's law only for the illustration in paraxial area of a thin lens used in practice is meaningless.

The difficulties of double focus compensation conventional view cameras in many scene lens constellations to master is from: "Creative large format", Sinar Edition, Tillmans Urs, publisher "Photographie" AG, Schaffhausen, pp. 60-62.

It is an object of the invention to have an automatic Focusing method of a scene at double Focus compensation with the camera specified at the beginning reduced expenditure on measuring equipment and simplified device and to achieve with practical lenses.

The solution to the problem is specified in the main claim.

Advantageous embodiments of the method and the camera are described in the subclaims.

The new process eliminates the need for absolute Distance measurement between the lens holder and the Image carrier. It is sufficient that the Angle measuring devices are available and a relative  Distance measurement, i.e. the changes in the pull-out various focus settings as well as the Coordinate measurements in the image sensor plane and the Focus determination exist.

The setting procedure can be done as follows sum up:

• - The photographer sets the perspective by swiveling the focusing screen and then carries out a very rough pre-focusing of the picture, generally by changing the extract and in individual cases by swiveling the lens. The parameters a _m and b _{m of} the image sensor plane, which satisfies the plane equation z = a _m x + b _m y + dm, which therefore define their normal vector, but not their distance from the lens, which is not required here, and the set values the swivel angle Θ₀ and Φ₀ of the lens are read into the processor via angle sensors built into the standard.
• - Determination of the coordinates bi = (xi, yi, zi) of at least three pixels Bi by focusing with the aid of a focus sensor; the camera lens is pivoted through the angles Θ _o and Φ _o . The X-axis is here the axis parallel to the optical bench of the camera, which goes through the center of the lens, the Y-axis is perpendicular to it in the horizontal direction, the Z-axis in the vertical direction.
• - Fitting the level for which the sum of the squares of the distances between the measured pixels becomes minimal. Let the parameters of this output image plane be a _o , b _o and d _o , it is then given by the position z = a _o x + b _o y + d _o .
• - Determination of the image width b from the difference Δb each time focusing on a certain pixel a given change of lens swivel angle - according to Equations (5) to (7),
• - Determination of the image width b with the lens not pivoted according to equation (8),
• - Determination of the z-axis parameters d _o according to equation (9).
• - calculation of the target lens displacement angle Θ _to, Φ _soll and the focusing screen intersection x _Bsoll with the X-axis according to equations (1), (2) and (3).

The output image plane is mapped into the object space; the plane thus generated is mapped back into the image space with a varied lens pivoting with the aim of aligning the new image plane thus obtained parallel to the set focusing screen plane. The necessary lens Verstellwinkelwerte Θ and Φ _{is intended to} be set as well as the Schnittpunktkoordinate x _Bsoll of the image pickup Rebene and the X-axis to be found

The following nomenclature applies to equations (1) - (3):
The parameters a _m , b _m and d _{m of} the image pickup plane equation describe the starting position of the image pickup plane to which the initial swivel angle values Θ₀, Φ₀ belong. The parameter d _m is the value of the Z axis section between the lens and the point of intersection with the image sensor plane. Accordingly, the initial parameters a _o , b _o , d _{o are} those of the position of the image plane in which the selected pixels Bi lie. f is the focal length of the lens and the angles Θ and Φ are the target lens pan angle values of the final optimal setting to be made. X _b is the x-intercept after normalization - dm / am according to equation (9) in the final setting, ie the target setting with double focus compensation.

The only physical basis is the Mapping equation for thin lenses; it must be demonstrated that the process is easy for the Use of any lenses, their optical behavior always by their focal length and the location of their two Main levels is defined, can be modified.

The lens is now pivoted by the angles Θ and Φ calculated in this way (in this case Θ = Φ = 0 would correspond exactly to the situation that the lens is perpendicular to the optical bench of the camera). The image pickup plane is shifted in parallel so that it intersects the X axis at the position with the coordinate word x _b .

Overall, the following transformation of the Input data carried out:

Here are the location vectors of the selected pixels Bi.

The two mapping equations of the original image plane and the new image plane that lie in the ground glass plane are related to the invariant scene level, which is why the two object planes are directly correlated. Consequently there are equations that are not based on the principle of Scheimpflug are set up because no scene level is determined and not the equations of the intersection line between the scene and the lens level as well as the Objective and image plane can be determined.

In the relationships ( 1 ), ( 2 ) and ( 3 ) for the setting of the camera, the following parameters a _m , b _{o of} the output image plane, the parameters a _m , b _{m of} the image sensor plane and the parameters Θ₀, Φ₀ of the pivoting of the lens quite easy to determine: The parameters a _m , b _m , Θ₀ and Φ₀ can be determined directly by reading out angle sensors attached to the standards, a _o and b _o by measuring separation differences when focusing the image points Bi. The measurement of the further parameter d _{o of} the output image plane, however, poses considerable technical problems insofar as the absolute value of the extract for each of the image points Bi has to be determined very precisely (to an accuracy of approximately 50... 100 μm). Since the extension length can reach values of up to one meter and more, this means that the distance between the image point Bi and the objective must be measured directly with an error of a maximum of 0.1 per mille. Although this is possible, it is technically complex, which is why we have developed an indirect method for determining the extension length. Therefore, an indirect determination of this parameter is carried out according to the invention.

The fact that the image width b of an object point S imaged by the lens changes if  the lens is pivoted. The extent of this Image size change Δb clearly depends on the Object distance s of the object point S, so that from the Angle value Θ of the respectively used adjustment angle and the Value of the separation difference Δb, that of the change in image width corresponds in each case with the associated focusing Value of the object distance s can be determined. Is thereby the object distance s known, can over the Imaging equation directly the absolute value of the image width b be determined. The absolute error of value for that Image width b is of the order of magnitude of the measurement error the separation difference Δb between the focus positions according to the change in image size, which can be determined very precisely (better than about 50 µm, so that the relative error the permissible value does not exceed.

This method becomes particularly simple when you look at the lens only pivoted about the Y axis by the angle Θ (i.e. Φ = O) and draws an object point S that is in the XY plane lies. Then the distance of this point from the Objective level solely through its object distance s and the Adjustment angle Θ given according to

Here γ is the distance of the nominated for the focal length f Object point s from the lens.

The following applies to the image width b of the image point from the scene point S. then:

The separation difference Δb (s, Θ) is defined accordingly as

This results in the following expression for the Object distance s:

In formulas (5) to (8), the angle value Θ is the one to which the pull-out difference Δb corresponds. In formula (7), only the solution with is physically relevant the plus sign in front of the root expression, because only they unite Value s = f provides a real representation of the Point leads. The image width of the point for Φ = O, relative to which the pull-out difference Δb is measured can now using the mapping equation according to

be determined; it is now possible to specify all other measured relative excerpts (ie the x coordinates of the sharp image points Bi) in absolute terms without having measured any image width directly, since the X axis is now provided with a defined zero point; the y and z coordinates can easily be derived from the position of the movable CCD chip, the focus sensor in the image pickup plane, the x-axis position there and the swivel angles α, β of the image pickup plane against the x-axis. From the intersection of the output image plane with the X axis, x _o , which can now be determined absolutely, its z-axis section d _{o can be determined in} accordance with

d _o = -a _o x _o (9)

calculate and insert into relationships (1), (2) and (3).

The accuracy of the method can be increased even more by not just one, but several, differentiation at different adjustment angles Θ and that from values obtained for the object width s averages. If when swiveling the lens, the measuring CCD chip through suitable motor control always on the Connection line starting point of the lens center point (Θ = O) can be held a large number of Extraction difference values Δb for many different values of Θ be determined during the lens pivoting process, whereby the accuracy of the value obtained for the Object distance becomes very large.

Should no suitable object point be available that lies in the XY plane, any other object point can of course also be used; however, the calculation becomes a little more complicated because the distance of the point from the objective plane also depends on its z coordinate z _s . The following applies:

and thus

At least two values for the pull-out difference Δb must now be measured in focus at different adjustment angles Θ; The relationship (11) can then be fitted to the course of the pull-out difference from the swivel angle value Δb (Θ) thus obtained by varying the free parameters s and z _s , that is to say by stepwise approximation. The parameter z _s is no longer required; From the object distance s, the image distance b (Θ = O) can be determined according to (8) without pivoting. The accuracy of the fitted parameters naturally increases with the number of measured values; Here, too, it is possible to carry out the measurements very quickly for a large number of values of durch by automatically tracking the focus adjustment sensor, generally a CCD chip.

The fact that during the determination of the Differences in focus of the focus sensor, the Measuring CCD, always aligned to the same place in the picture can be benefited by looking at the complete Uses information that the focusing method provides. The The method is based on that for the considered Image region, for example according to DE 44 13 368 C1 The measured sharpness value is determined by the distance of the Focus sensor, the measuring CCD, (in the x-direction) from Focus depends and becomes maximum in focus. Out of focus their value gives a measure of the distance of the measuring CCD from sharp pixel. If you take advantage of this, you only need once in the beginning, to be passed through the vicinity of the focus the course of the Sharpness measurement value depending on the relative location on the Determine X-axis; from reducing the Sharpness values when swiveling the lens can then immediately on the focus and thus directly on the Movement difference can be closed without new focus. This approach enables faster  Measurement of the pull-out difference of Δb as a function of Lens swivel angle Θ.

With the methods shown here it is possible to do it alone all necessary adjustment parameters from the image information to determine; it also bypasses the image widths of the sharp pixels to be reproduced by direct Length measurement to determine what is required with the Accuracy is technically very complex and therefore high Would be associated with costs. Instead, the Object distance of a basically arbitrary object point by measuring the image width of its pixel calculated different lens positions and from this the Absolute value of the image width with the lens not swiveled determined. The X-axis of the "bench-fixed" camera coordinate system is therefore provided with a zero point, whose distance from the lens is known very precisely and relative to which all other move-out differences with high Accuracy can be measured; the used for this Length measuring distances of the displacement sensor only have to be a few cm To be long.

In addition, all problems that arise with a direct determination of the distance from the thermal Extension of the optical bench or from the not quite exact assembly of a split optical bench could result in several sections. It is not even important whether the lens and the image standard at all are interconnected. The distances in the Imager level can still be determined absolutely directly must (all lengths in the y * and z * directions, i.e. the Pixel coordinates in the image sensor plane), lie in the cm range and are therefore also about inexpensive Displacement transducers easily with the required accuracy determine.  

Advantageous configurations are shown in FIGS. 1 to 4.

Fig. 1 shows the camera in a simplified representation and without bellows with schematized control device,

Fig. 2 shows the functional dependence of the image distance from the extension difference at various lens Verschwenkwinkeln each focal length is normalized,

Fig. 3 shows the absolute error image distance at given measurement errors for a 65 mm lens at different angles of rotation,

Fig. 4 shows image size errors for a 1800 mm lens.

Fig. 1 shows a camera with an optical bench (OB), which can also be composed of several sections. A lens mount (OT) is mounted on it, which carries a lens (O) which is held pivotably about two axes (Y, Z). The swivel angles (Θ, Φ) are signaled to the processor (PR) via angle sensors. Possibly. electrically controlled lens swivel drives (HA, VA) are provided, which are controlled by the processor (PR).

There is also an image carrier (BT) on the bench (OB) arranged, which is also pivotable about two axes Image sensor level (BA), which during the adjustment in generally equipped with a focusing screen. Also whose swivel angle (α, β) are the angle encoder Processor (PR) reported. These pivots are also possibly by electrically controlled drives from the processor (PR) controllable.  

A CCD sensor (CD) is in the image sensor level positionable, with two relative position coordinates (Y1 *, Z1 *) of the same in the image sensor plane to the Processor (PR) are given. This relative Position coordinates can be from a positioning device Measuring probe (CD) can be removed. In addition, the CCD measuring probe (CD) sharpness-relevant pixel brightness signals (SS) of the respectively selected area in which the considered pixel is to the processor (PR) Evaluation for a sharpness criterion in a known manner using a Fourier analyzer (FA).

The image carrier (BT) is on the bench (OB) with a Extension adjustment device (AA) axially adjustable arranged, the adjustment signal to the processor (PR) as the Extension range (AW), which is included in the formulas as Δb, is forwarded. The pull-out adjustment device (AA) may have an electrically controlled drive by the processor (PR) is controlled.

To illustrate how the Focusing is a Cartesian Coordinate system drawn in, whose X axis through the Main point (H) of the lens (O) parallel to the bank (OB) runs. If a swiveling bench is provided, must their pivoting against the standards by a corresponding coordinate transformation is taken into account will. The Z axis is at the top and the Y axis to the side directed. In the simplest case, three scene points (S1-S3) sharply imaged in the image pickup plane (BA) and in a certain perspective to each other be arranged. When presetting by hand, the three pixels (B1-B3) belonging to the scene points (S1-S3) generally at one of the imager levels deviating initial image plane, which is therefore due to these three Pixels (B1-B3) is determined. In the example is the first  Pixel (B1) with the relative position coordinates (Y1 *, Z1 *) by means of the sensor (CD) on the image sensor level set. The second and third pixels (B2, B3) lie outside the image sensor level. From this The basic setting will then become the further focusing. To do this, each after positioning the CCD sensor (CD) on the second or third pixel, which is initially on the Image sensor level is still out of focus, Changes to the excerpt made up to focus. This results in differences in the extract (x1-x2, x2-x3) the position of the points in the x direction and their relative position coordinates Y *, Z * in the image pickup plane, based on the image carrier angle settings (α, β) in relative spatial coordinates, the inclinations of the Describe the initial image level, to be converted.

For the further calculations, it is provided that a new type of image width determination is carried out, for which purpose the lens (O) is pivoted by a predetermined angle (Θ) and a previously focused measuring point is brought into focus again, the change in separation (Δb) being measured. With the addition of the lens focal length (f), which are usually electrically encoded to be transmitted from the lens into the processor, can now be the output signals (Θ _should Φ _to x _Bsoll) for pivoting the image points in a new image plane parallel to the image pickup Rebene and their calculate and output the corresponding parallel shift in the X direction into this new image plane.

In FIG. 2 the dependence of the image distance (b) reproduced by the extension difference (.DELTA.b); both quantities are normalized to the focal length (f), which makes the graph independent of (f). The swivel angles are (from top to bottom) Θ = 2 °, 5 °, 7 °, 10 °, 15 °, 20 °, 25 ° and 30 °.

The larger the swivel angle, the better the image width can be determined, since then a small image width already causes a large separation difference. If a scene point located far away is selected, the following results as a special case: If the value b _{f is} known for which b _f cannot be distinguished from 1 in terms of measurement technology, the associated (very large) object width s _f applies.

b _{f is} the normalized image size and s _f the normalized object size in the special case mentioned.

From this object distance s _f , the focal length (f) can be set as the image width without any further measurement.

The new device therefore does not require any special Distance measuring means for measuring the lens-image plane distance, and such a distance measure or a Image scale or the like. Do not have to in the Processor can be entered as it is through the Pull-out difference measurement have been replaced; the existing ones Measuring equipment is also used for this. The novel Calculation methods do not require the measurement of the absolute Distances of the pixels to the lens.

The angle and extension settings for the Acquisition processes can be carried out manually, where the focus is displayed by the processor is, as well as possibly with electrical actuators automatically. Setting the target angle and the The target extension dimension can be adjusted by hand based on processor-controlled setting instructions or, if necessary, directly  with the electrical actuators by the processor respectively.

FIGS. 3 and 4 show error graph for a lens 65 to 1800 mm focal length, wherein a Auszugsmeßgenauigkeit of 10 microns and a Winkelmeßgenauigkeit of 1 'are accepted and pivot angle Θ of 5 °, 7 °, 10 °, 15 °, 20 °, 25 ° and 30 ° are selected as parameters. With large swivel angles, the error can be kept small even with large focal lengths. However, since relatively large extension lengths are to be measured, it has proven to be advantageous if the x coordinate determination is made from the extension distance (AW), which is measured with the specified accuracy by measurement technology, and combined with a coarsely screened, but exact basic dimension. For this purpose, the optical bench (OB), FIG. 1, has a mechanical grid, in particular a hole grid (L1-LN), which is used to position the object carrier (OT) and the image carrier (BT) on the bench with the aid of pinning by means of pinning. Several bench sections (OB1, OB2) are also connected to one another with grid accuracy by pinning in the hole pattern. The x coordinate determination is thus carried out first, as shown, with a relatively large error, after which the dimensions are classified in the next adjacent grid section and the remaining distance from the grid point found is added to the associated grid dimension, so that the absolute total error is only slightly larger than that the extension range is. This allows working with small swivel angles Θ with relatively small extension widths, while still achieving high accuracy in determining the image width.

The grid length is expediently smaller than the measuring section the extension width selected and several times larger than the biggest absolute error that occurs at a given Swivel angle occurs.  

The hole pattern provides another alternative for indirect image size or scale determination by the lens carrier (OT) by at least one grid length is shifted and before and after a focus of a scene point, the extension length is determined.

The grid on the optical bench and the associated Locking devices on the standards and bank details are advantageous designs of the camera arrangement for the Application of the new method to the known double Focusing, d. H. with one lens swivel each two axes perpendicular to the bank direction.

As already mentioned, the formulas given can also be extended to lenses with several main planes, and it can also be taken into account that the pivoting of the lens often does not take place about a main plane axis but parallel to it. The distance ie the main planes results in a parallel offset h of the image plane in the direction away from the lens from h = d _h · cosα *, where α * is the total angle of inclination of the lens to the X axis. The Z-intercept parameter of the shifted image plane d ^V _{m is derived} from this:

where a _m and b _m are the plane equation parameters of the imaging plane and Φ and Θ are the respective lens angle values and d _{m is} the Z coordinate parameter with respect to a thin lens.

By converting to the pivot coordinate system, whose Z-axis parameter is marked with the high index H,  and using the calculated target lens angle the Z-axis section results

where A₁ is an offset parameter, the distance from the object-side main point of the lens to the lens pivot point, and where the index p corresponds to the index "should" in other equations, and the associated target pull-out adjustment x _bp for double focus compensation is then

Claims (11)

1. Focusing procedure for a camera

• with a lens carrier (OT) and an image carrier (BT), which are held such that they can be changed relative to one another in a bank direction (X), the corresponding changes in distance (x1-x2; x2-x3, Δb) being fed to a processor (PR),
• - And in the lens carrier (OT) a lens (O) about two mutually perpendicular and perpendicular to the bank direction (X) axes (Y, Z) is held, both of which have the corresponding lens pivot angle values (Θ, Φ) and the value of the lens focal length (f ) are fed to the processor (PR),
• - And in the image carrier (BT) an image pick-up plane (BA) is held pivotable about two axes (Y, Z) perpendicular to each other and perpendicular to the bank direction (X), both of which corresponding image pick-up pivot angle values (α, β) are fed to the processor (PR) ,
• and in the image pick-up level (BA) thereof, a focus sensor (CD), which delivers a focus signal, can be positioned in two image pick-up coordinates (Y1 *, Z1 *), which are each fed to the processor (PR) together with the focus signal,
• - Whereby for the simultaneous focusing at least three selected pixels (B1-B3), which lie in an output image plane, these on the image pickup level (BA) with the focus sensor (CD) by changing the image carrier distance from the lens (O) one after the other in their focus while detecting the associated image sensor coordinates (Y2 *, Z2 *; Y3 *, Z3 *) and the image carrier distance change (x1-x2; x2-x3), the position in the processor (PR) is recorded,
• - And from the input data thus obtained from the processor (PR), positional parameters (a _o , b _o , a _m , b _m , Θ₀, Φ₀) of the image pickup level (BA) and the output image level are determined mathematically by solving first mapping equations, which are shown by mapping the output image plane into the object space and subsequent re-imaging into the image space with varied lens pivoting, a further necessary imaging parameter representing the absolute image width of a selected pixel being obtained by focusing on a selected pixel in the image pickup plane (BA) and then pivoting the lens by a predetermined lens swivel angle (Θ, Φ) or a lens shift in the bank direction by a predetermined lens shift path and then the selected image point is re-focused by a change in focus and the difference in focus (Δb) on the image carrier side between hen the two focus settings are measured and these are each used in further mapping equations, which contain a clear connection between the image width change (Δb) and the object distance of the selected image point, together with the associated lens swivel angle or the associated lens displacement path and the first and the further imaging equations together for wanted Solleinstellwerten (Θ _to, Φ _soll) can be solved for pivotal movements of the objective carrier (OT) and a target distance setting (x _Bsoll) of the image carrier (BT) relative to the lens carrier (OT) out.

2. The method according to claim 1, characterized in that the selected pixel at not around the horizontal axis (Y) pivoted lens (O) in the bank direction and the plane spanned perpendicular to it, horizontal swivel axis direction lies and the pivoting of the lens (O) to  Determination of the pull-out difference around the horizontal axis (Y) a predetermined swivel angle (Θ) takes place. 3. The method according to claim 1, characterized in that the selected pixel with at least two lens swivel angles is in focus and the separation differences (Δb) be measured and from the associated equations (11) an object width (s) and the image width (b) can be calculated from this. 4. The method according to claim 1, characterized in that during a change of statement continuously for the Partial extract changes from the respective pixel signals (SS) associated focus values for the selected pixel determined and associated with the partial statement changes be saved and after swiveling the lens by the specified swivel angle (Θ) or after the Lens shift by the specified displacement the then generated pixel signals (SS) the focus value is determined and with this from the stored Sharpness measurement values the associated change in pull-out (Δb) is selected, which serves the further calculations. 5. The method according to any one of the preceding claims, characterized in that the optical bench (OB) a Bank grid (L1, LN) carries the image carrier (BT) and the Lens carrier (OT) are to be fixed and from the Extraction difference (Δb) approximately determined the image width (b) and the next associated grid spacing of the Image carrier (BT) from the lens carrier (OT) of a bank grid (L1-LN) and the grid spacing Extraction difference (Δb) measured in a grid-related manner is added and this image carrier lens carrier distance calculated in this way as a more precise value of the image width (b) for the others Calculations is used.   6. The method according to any one of claims 1, 4 or 5, characterized characterized in that the displacement path at the Lens shift is one or more grid lengths. 7. The method according to any one of the preceding claims, characterized in that _(to Θ, Φ _should x _Bsoll) in the calculation of Solleinstellwerte dependent due to the pivoting of the objective (O), a parallel offset (h) of the image plane from a main plane distance (d _h ) of the lens (O) and an offset of the lens pivot axis to its main planes are taken into account (Equations (13) - (15)). 8. Camera that uses a focusing procedure one of claims 1 to 7 is to be operated, with an optical bench (OB) on which a Lens mount (OT) with one lens (O) and one Image carrier (BT) with an image pickup level (BA) in which a film or a focusing screen are to be arranged, each in Bench direction (X) with an intermediate bellows extension are positioned and each around vertical Axes Y, Z to bench direction X by swivel angle (Φ, Θ; α, β) are pivotable, and each associated Swivel angle values (Φ, Θ; α, β) can be fed to a processor (PR) are and in the image sensor level (BA) Focus sensor (CD) must be positioned so that it can be positioned, its focus signal (SS) and each set image sensor coordinates (Y *, Z *) to the processor (PR) can be fed and a pull-out difference (Δb) of the bellows extension in the bank direction (X) Focus of at least three pixels the processor (PR) can be fed and the associated absolute Image width of a pixel Measurement of the image width change (Δb) of the pixel at different Lens pivot angle is determined, wherein the optical bench (OB) with a grid is provided with mechanical catches (L1-LN), and the Lens carrier (OT) and the image carrier (BT) with Latching devices in the notches precisely spaced from each other are held.   9. Camera according to claim 8, characterized in that the optical bench (OT) from at least two bench sections (OT1, OT2), which is connected to a connector by means of a Latching device in the given grid with grid accuracy are connected. 10. Camera according to claim 8 or 9, characterized in that the grid length is shorter than a maximum Excerpt measurement length. 11. Camera according to one of claims 8 to 10, characterized characterized in that the catches as perforated catches (L1-LN) are formed and the locking devices are locking pins.