EP2686733A1

EP2686733A1 - Method and device for focusing a film camera

Info

Publication number: EP2686733A1
Application number: EP12711127.6A
Authority: EP
Inventors: Martin Waitz
Original assignee: Individual
Current assignee: Individual
Priority date: 2011-03-18
Filing date: 2012-03-19
Publication date: 2014-01-22
Also published as: AT511312B1; WO2012126868A1; AT511312A1; US9225894B2; US20140009665A1

Abstract

The invention relates to a method for focusing a film camera (1), wherein at least one auxiliary camera (6, 7) for producing an auxiliary representation of an object to be captured is provided, from which the desired focus setting is determined and a focusing signal is output to the film camera (1). Improved focusing can be achieved by virtue of the fact that the auxiliary camera (6, 7) is releasably connected to the film camera (1) and outputs the focusing signal to a servomotor fitted to the film camera (1).

Description

Method and device for focusing a film camera

The invention relates to a method for focusing a film camera, in which at least one auxiliary camera for producing an auxiliary representation of an object to be recorded is provided, from which the desired focus adjustment is determined and a focus signal is output to the film camera.

Depending on the application, different lenses are used for the different film rotation conditions. Usually, lenses are constantly changed on the film camera during a filming. Lenses can have three adjustment axes: a focus, aperture, and zoom adjustment axis. Each adjustment axis has a scale ring, which is provided by the lens manufacturer with engraved values. A mark indicates the currently set value.

For focusing a subject, adjusting the aperture or the zoom range of film cameras, it is known to flanging servo motors on the pinion of the adjustment axis on the lens. The control of the individual servo motor via a focusing device, which is often designed as a separate device. The focusing device can also be integrated in the servo motor. It is also the case that the servo motor and / or the focusing device are integrated in the lens, or are attached as a common unit to the lens.

It is possible for the operating personnel to specify desired positions of the focusing device with the aid of one or more manual control units. The devices are connected via cable or radio connection. Finally, the servomotor processes the default and adjusts the scale ring on the lens.

Manual control unit, focusing device and servo motor form the lens control system of a film camera.

When shooting movies, the correct focus of the film camera is a critical task. Especially with artistically valuable films very often worked with shallow depth of field to highlight certain objects or people or their parts accordingly. It is state of the art that the lens control system can be supported especially in the focus by other auxiliary devices.

One way of focusing, for example, is to aim the object to be focused with a laser-based rangefinder or an ultrasonic measuring device and to obtain therefrom the focusing information for the film camera. The disadvantage of this is that when moving objects continuous tracking is required and in many cases the spatial relationship between the film camera and the rangefinder is not clear. Another disadvantage is that these gauges only measure the distance from a certain point in space.

The data must be exchanged between these individual devices. For fast-moving objects, there is a tracking effect of the focus motor, since data exchange and the separate processing of the data lead to many dead times. A consequence of very fast objects is not possible.

From EP 1 084 437 B a method and a device for focusing a film camera are known in which auxiliary cameras are used, which are provided laterally next to the actual film camera. These auxiliary cameras are pivotally arranged so that the optical axes can be aligned with the object to be in focus. Triangulation calculates the distance of this object and obtains the focus signal. A disadvantage of this system is that the mechanical movement of the auxiliary cameras is susceptible to interference and even slight tolerances in the pivoting movement, in particular in the middle distance range, can result in great inaccuracies in determining the distance. Moving objects require continuous manual tracking. When changing the objective, the system has to be recalibrated.

The above problems are partially solved by the use of a camera as described in WO 2009/133095 A. In this case, an auxiliary representation is generated by a mounted on a camera auxiliary sensor, which can be used to obtain distance information. The disadvantage of this, however, is that it is not readily possible to make a distance determination before taking a picture, without using the camera and that the distance determination always depends on the particular lens used, which is constructed on the camera. Also, the calibration of the distance measurement is possible in the known solution always only with the camera in each state present.

Irrespective of the fact that the lens often has to be changed in film cameras, there is the following problem: Real optics for film cameras pump, ie the image section (the focal length) changes when focusing, thus there is no constant correlation of the image sections between the two cameras. To calculate a depth image, however, the image scales and therefore the image details of all participating cameras are to be known exactly. With zoom optics, the general problem is to know the current exact focal length range. Most optics for film cameras have no built-in electronics, which output the scale position and the focal length range.

Another problem is that real optics for film cameras - especially zoom optics - change the optical axis when zooming, or change the optical axis depending on how the optics is mounted on the camera. Therefore, a parallel alignment of the optical axes in reality is possible only by a very high effort, since the entire camera (camera including optics) would have to be moved while zooming. These two deviations (pumping and wandering the optical axis) would have to be given to an algorithm for depth calculation and are different for each optics. Without these values the measurement is inaccurate or not possible at all.

The biggest problem with filming is this: Filming often has a small desired depth of field. That is, the image of the main camera is blurred in large areas. A depth calculation can not take place in blurred image areas, i. The main camera can only be used to calculate a stereoscopic image if the depth of field is very large.

A less relevant, but often disturbing effect is that color images for a depth calculation (especially one that should run in real time) are not very suitable and provide comparatively inaccurate results. Images of grayscale cameras are therefore preferably used for the depth calculations.

Object of the present invention is to avoid these disadvantages and to provide a method in which a reliable distance signal for focusing the film camera can be obtained in a simple and robust manner. In particular, the distance determination should be made largely independent of the actual camera.

According to the invention, these objects are achieved in that the auxiliary camera is detachably connected to the film camera and outputs the focusing signal to a mounted on the film camera servo motor or to a lens control.

Preferably, a monitor for displaying the auxiliary representations is provided, which preferably also allows a superimposition of a plurality of auxiliary representations.

Auxiliary representation in this context means the real image of an auxiliary camera, from which a depth image is calculated.

It is therefore particularly preferred to use two auxiliary cameras which produce at least two auxiliary representations, from which a distance value is calculated for a plurality of pixels. The auxiliary cameras can look at the Optimized position of auxiliary representations, which are optimized for the depth calculation, for example, by working with strong closed aperture to achieve a large depth of focus. This allows the realization of a great freedom of design, since it is possible to select on the monitor at any time a displayed object to be focused on, even if this is the moment of selection far out of the focus range and thus is shown in the actual representation extremely blurred ,

Without jeopardizing the above advantages, if the spatial relationships, the optical properties of the lenses used and the tolerances are known, the video image of an auxiliary camera can be included in the image of the video camera so that the motif selection is also made via the image of the film camera can.

In the image calculation unit, it is possible to detect the removal of multiple subjects simultaneously. These areas or motifs can be defined before the movie is shot. The resulting distance values can be transmitted numbered to the focus setting of the film camera. This makes it very easy to change the focus setting between different subjects (for example by pressing a button) and to focus on these distances without having to observe or operate the display unit. The operator can focus on the action in front of the camera during the filming. He only needs to operate the focus setting, which he does anyway. But he can still take advantage of an automatic distance measurement.

In addition to the active measurement of a distance value by the user, the system is suitable for semi-automatic or automatically passive distance measurement. The following methods are used:

Restriction of the area of the auxiliary representation;

Setting distance limits to hide unwanted obstacles (such as columns);

Setting Focus Ramps / Fade: The monitor can mark two or more areas (with different distance information) between which the focus will run from area n to area n + 1 in a ramp time function within a specified period of time.

Representation of the depth of field (focus area of the film camera) in the auxiliary representation;

Automatic distance measurement at a fixed point (for example center of the auxiliary representation); Automatic measurement of the closest point in the auxiliary representation or in a defined region of the auxiliary representation;

Instead of the closest distance another criterion can be given, such as:

Averaged distance over all measuring points in the defined measuring field range;

Averaged distance according to a specific weighting and distribution method of the measuring points (eg exclusion of extremely close and / or distant measured values);

Longest distance point;

Motive tracking, visual field tracking: Automatic tracking of subjects;

Pattern recognition: Criteria of the measuring range are predefined features (such as colors, contours, etc).

All of these methods can also be used to display only a distance value. The user can also use the values as a decision aid for focus adjustment here.

Essential to the present invention is the fact that a rigid arrangement of the auxiliary camera and the auxiliary cameras and optionally the film camera is provided, wherein preferably the optical axes of all cameras are aligned parallel to each other. Naturally, the image of the auxiliary cameras differs due to the parallax depending on the distance of the recorded objects. By pattern recognition, it is now possible to identify a particular object in the auxiliary representations and to determine the optical displacement. From this it is possible to deduce the distance of the object in question. This distance signal is output to the film camera as a focusing signal.

Alternatively, the auxiliary camera may be designed as a TOF camera. Cameras of this type determine the duration of the light from or to the object, which also explains the name (time-of-flight camera, time of flight camera). For each pixel of the camera thereby a distance signal is determined and supplied.

Pattern recognition is the ability of a method to detect regularities, similarities or regularities in a set of image data.

It is state of the art that in pattern recognition three basic approaches are pursued and implemented. It is syntactic, statistical and structural pattern recognition. In syntactic pattern recognition, things are described by sequences of symbols. This can z. As the colors or certain contours. The goal of syntactic pattern recognition is to find objects of the category that have these descriptions.

Statistical pattern recognition determines the likelihood that an object belongs to one category or another and then assigns it to the category with the highest probability. Instead of evaluating features according to prefabricated rules, the measured numerical values are combined into pattern vectors. A mathematical function uniquely assigns a category to each conceivable pattern vector.

The structural pattern recognition combines different approaches of the syntactic and statistical method to a new method. An example is the here important visual field recognition, in which for different facial parts such as eye and nose different classification methods are used, each of which only state whether the sought body part is present or not. Higher-level structural procedures bring together the individual results and calculate a result - category affiliation. This makes it possible to identify one or more faces in pictures and to follow them with moving pictures.

A monitoring and control of the focusing is in particular possible in that the auxiliary representations are displayed on a monitor, and that preferably also a superposition of several auxiliary representations is made possible. The person responsible for the focus adjustment can thus monitor the system and easily identify sharp and blurred areas based on the overlay. It is also possible to display certain marked areas to be focused on.

The current focus signal is used to superimpose the auxiliary representations of at least two cameras in such a way that the area corresponding to the distance range of the film camera is exactly matched / superimposed. Areas in the auxiliary representation, which correspond to the distance range are clearly visible. Areas that do not overlap are not in the focus area and in the auxiliary display as shifted images also inaccurate to see. This allows the operator to see in the display which area has been focused.

It is particularly advantageous if, in the auxiliary representations, the area recorded by the film camera is marked, for example, by a frame, if the auxiliary representations are not brought into conformity with the recorded area by appropriate processing. For this purpose, the image area of the film camera is displayed in the auxiliary display. It is particularly efficient if the auxiliary representations are resolved into pixels or pixel groups and that a distance value is stored for each pixel or each pixel group by the stereoscopic displacement of the auxiliary representations. The recording system therefore has distance information for each part of the image, which can be used for focusing. In particular, complex focusing strategies can also be specified. Thus it can be determined in this way that to focus on the nearest object, with the exception of objects whose distance is below a certain limit, or excluding objects that lie in a particular image area. This is advantageously applicable if the person closest to the camera is to be in focus, but it is to be expected that a nearby pillar will come into view.

It is also possible to use different averaging methods, such as weighted averaging. For example, it is possible to focus on the average distance of all pixels in the specific area, but particularly close or especially distant pixels are taken into account in the averaging with only a slight weight or are not used at all to form the distance average. Similarly, it may also be specified that only pixels that have a similar distance in a contiguous area of a certain minimum size are taken into account.

The operator can select a pixel on the monitor, from which by the subordinate depth image immediately the corresponding distance value is transmitted to the lens control. This selection can be made via a mouse control, touchpad or similar.

A particularly favorable embodiment of the method according to the invention provides that an operator records an object to be recorded for the purpose of focusing. This may be, for example, an image hanging on the wall which is always in focus regardless of whether the camera is moving or panning toward the object. However, this variant of the method is particularly advantageously applicable when moving objects, such as vehicles or persons, are to be focused over a certain period of time. Through a visual field recognition, a face can be located in space, and this face can be followed. But it is also possible, for example, set in a close-up with shallow depth of field, the distance to the eyes of the person concerned.

This method variant uses known algorithms that are capable of detecting predetermined objects or patterns on images. In combination with the algorithms described above, however, the development of distance of several objects are determined simultaneously and from a derived focus value, such as an average, are calculated. Or it can be transferred between the objects in a given period of time, the distance value, so as to implement a focus ramp.

Preferably, the procedure is such that at least one auxiliary representation is output on a display device on which information about the focusing is also output. This means that a depth image calculated from the auxiliary representations is output, for example by displaying a false color representation in which the color expresses the specific distance of the respective object. For example, a certain shade of red corresponds to a distance of one meter, while a certain shade of blue means infinite.

Real image and depth image can be displayed side by side. In addition, the image of the film camera can be displayed so that the operator can also observe the real image of the film camera and see which object is in focus, or what the image of the film camera is. Furthermore, it is possible to display the image of each auxiliary camera. You can switch between these views, or these representations can be superimposed.

After each lens change, the motors should be recalibrated to control the adjustment shafts so they will not impact the mechanical stop of the lens. During the calibration process, the motor drives the mechanical stops at low speed and at the same time determines the possible adjustment path. Each lens has a more or less different adjustment. Is in the focusing a table with an assignment between Motorverdreh away stored to scale position, then the system, the scale position and thus the set focus value of the lens is known after the calibration. Normally, this allocation table is not known and not stored in the Motoransteuersystem because of the variety of different optics. It is therefore possible in the arrangement described according to the invention to deposit different lens tables in the image calculation unit and to select the correct table after a change of the objective so that the scale position and thus the focus value can be predetermined for the lens control system.

Because of the variety of different lenses, creating and storing lens tables is tedious and expensive. Therefore, a mapping table with few points after a lens change can be easily calibrated. The closest scale value and infinity are known after calibration of the engine. Intermediate points are approached on the lens and the corresponding distance value is entered. Thus, an assignment table with arbitrary interpolation points can be easily created and stored.

If easily identifiable objects (eg simple contours) are present for pattern recognition, then this calibration process of a lens can also be automated. For this purpose, the real image of the film camera and an auxiliary representation of the image calculation unit must be supplied. A shallow depth of field is set via the aperture. The focus engine automatically rotates slowly so that only objects at the appropriate distance in the image of the movie camera are focused. These images are compared with the auxiliary representation and examined for common patterns. If the pattern can be identified, the corresponding focus value of the lens of the film camera is known to the system on the basis of the existing depth information of the auxiliary representation. Thus, some or many interpolation points of an assignment table Verdrehweg servomotor can be stored to distance value.

It proves to be particularly advantageous if the current focus signal and the range of the depth of field of the image calculation unit are transmitted. The current focus signal is the distance value on which the lens is currently focused and is known to the focusing device when the allocation table Verdrehweg servomotor is stored to distance value. The depth of field results from the physics of optics and is the area where an image is sharply displayed. With these distance values, it is possible to color-highlight or even hide image areas in the auxiliary representation that are outside the depth of field. This is easily possible since the pixels or pixel groups of the auxiliary representation are stored with distance values. If the operator changes the focus, the auxiliary display also changes. As a result, the operator can easily see in which spatial area the focal plane or the focus area lies. Furthermore, the present invention relates to a device for focusing a film camera, with at least one auxiliary camera which generates an auxiliary representation, and with an image calculation device which is connected on the one hand to the auxiliary camera and on the other hand drives a focusing device of the film camera.

According to the invention, this device is characterized in that the auxiliary camera is detachably connected to the film camera.

Several auxiliary cameras can be used for the measurement. All auxiliary cameras are calibrated to each other. Several auxiliary cameras have the advantage that the shading of areas in space is minimized or the measuring range can be changed very easily.

Preferably, the auxiliary cameras are fixedly connected to one another and, more preferably, fixedly but detachably connected to the film camera, wherein the image calculating device calculates the focusing signal on the basis of a pattern recognition of the auxiliary representations.

In principle, the method according to the invention can be carried out with auxiliary cameras which are permanently connected to each other, but which are independent of the film camera. In this case, on the other hand, the distance and the different orientation of the film camera on the one hand and the auxiliary cameras on the other must be accurately detected and taken into account in the focusing in order to compensate for the parallactic errors. However, it is simple and more efficient if the auxiliary cameras are firmly but detachably connected not only to one another, but also to the film camera, for example by being arranged laterally on both sides of the film camera. Particularly preferably, the optical axes of all cameras are parallel to one another, which facilitates the implementation of the calculations accordingly.

If the image plane, which is the area in which the optics depict sharply, the film camera and the auxiliary camera shifted, can be calculated by a simple offset specification of this offset in each reading.

The calculation of the depth image is done in the image calculation device, which can form a unit with the monitor, but can also be solved by the auxiliary monitor. To achieve a high degree of operating flexibility, the auxiliary monitor can be connected to the measuring device and the lens control unit by cable or radio. Image calculation device and auxiliary monitor can also consist of one unit.

The auxiliary camera and the image calculation device can also output distance values without specification from an auxiliary monitor if calculation parameters are set on an input device. A calculation parameter can be to use the closest measuring range of the auxiliary representation. Or the measuring range of the optical center of an auxiliary camera can be output.

Preferably, the focusing device and the image calculation device can be combined in one system, wherein the servomotor for the objective is connected to the image calculation device. This ensures that the dead times are minimized in this closed system. The servo motor can react very quickly to changes in distance and it is possible to follow very fast object movements.

In order to have corresponding information available during post-processing, the auxiliary representation can be stored in a memory. In this way, the respective distance information is also available for the actual image, which information can be used, for example, for subsequent 3-D post-processing.

As such, it is possible to associate the distance information pixel by pixel with the individual pixels. In order to reduce the amount of data, however, the pixels can also be grouped by assigning a distance value to only one group of four by four pixels, for example. In this way, it is possible to significantly reduce the computational effort and thereby increase the focusing speed with a given computing capacity. This is particularly advantageous when the method is operated so that is focused on the respective nearest object.

Furthermore, the invention also relates to a device for focusing a film camera, with at least one auxiliary camera which generates an auxiliary representation, and with an image calculation device which is connected on the one hand to the auxiliary camera and on the other hand drives a focusing device of the film camera. According to the invention, this device is characterized in that the auxiliary camera is detachably connected to the film camera.

In this context, it is particularly preferred if at least two auxiliary cameras are arranged offset from one another in the use position in the vertical direction. When taking pictures that are parallaxically offset, there is naturally a shading, ie there are areas of the recorded objects that are detected only by one camera, but not by the other. If two auxiliary cameras are arranged laterally offset from each other, then in a frontally recorded from the front person, for example, one ear will be completely detected only by an auxiliary camera. It has been found that in certain situations these shadows can be problems for focusing. By a vertically stacked position of the auxiliary cameras, these problems can significantly can be reduced because shadowing on the top or bottom of recording objects are usually not disturbing.

As a result, the present invention will be explained in more detail with reference to the embodiments illustrated in FIGS.

Fig. 1 shows schematically the construction of a device according to the invention; and

FIG. 2 shows an embodiment variant of the invention in an axonometric representation.

A film camera 1 of Fig. 1 is provided with an objective 2 having an optical axis 2a. A servomotor 3 is attached to the lens 2 to perform the focus adjustment. The servo motor 3 is driven by a focusing device 4. A monitor 5 is provided in a known manner to display the image taken by the film camera 1.

In front of the film camera 1, two fixedly connected auxiliary cameras 6, 7 are arranged, which have optical axes 6a and 7a, which are parallel to one another. The auxiliary cameras 6, 7 are either firmly attached to the film camera 1 or unillustrated measuring means such as position measuring sensors are provided to determine the relative position of the film camera 1 with respect to the auxiliary cameras 6, 7. The auxiliary cameras 6, 7 feed their image into an image calculating device 8, which calculates the respective distance for the individual pixels due to the different image information of the two auxiliary cameras 6, 7 and the distance d between the optical axes 6a and 7a. The images taken by the auxiliary cameras 6, 7 are displayed in superimposed form on another monitor 9, which forms the display device. An input device 10 serves to operate the image calculation device 8.

In the image calculation device 8, the images of the auxiliary camera 6, 7 are analyzed and the pattern recognition described above is performed. At the same time, the calculations given by the operator are performed, which ultimately result in the focusing information which is passed on to the focusing device 4.

From FIG. 2 it can be seen that the two auxiliary cameras 6, 7, whose axes 6a and 7a are arranged at a distance d, are mounted vertically next to one another next to the actual film camera 1. The term vertical refers to the position of use, which is defined by the horizontal position of the longer side of the recorded image.

Furthermore, here is a single monitor 5, 9 is provided, which is switchable between different display modes. The present invention makes it possible to significantly improve the accuracy and reliability of focusing in film shooting.

Claims

PATENT APPLICATIONS

1. A method for focusing a film camera (1), wherein at least one auxiliary camera (6, 7) is provided for producing an auxiliary representation of an object to be recorded, from which the desired focus adjustment is determined and a focusing signal is output to the film camera (1), characterized in that the auxiliary camera (6, 7) is detachably connected to the film camera (1) and outputs the focusing signal to a on the film camera (1) mounted servo motor (3).

2. The method according to claim 1, characterized in that at least two auxiliary cameras (6, 7) produce at least two auxiliary representations, from which a distance value is calculated for a plurality of pixels.

3. The method according to claim 2, characterized in that the two auxiliary representations are offset in the vertical direction against each other.

4. The method according to claim 2 or 3, characterized in that the auxiliary representations are displayed on a monitor (9) as a display device, and that also a superposition of several auxiliary representations is made possible.

5. The method of claim 1 to 4, characterized in that at least one auxiliary representation on a display device (9) is output, on which also information about the focus is output.

6. The method according to claim 5, characterized in that on the display device (9) the image areas are marked within or outside the depth of field.

7. The method according to claim 1 to 6, characterized in that the auxiliary representations are resolved into pixels or pixel groups and that a distance value is stored for each pixel or each pixel group.

8. The method according to any one of claims 1 to 7, characterized in that an area is entered by an operator, from which the focus signal is calculated.

9. The method according to any one of claims 1 to 8, characterized in that an operator inputting an object to be recorded, based on the pattern recognition is performed.

10. The method according to any one of claims 1 to 9, characterized in that after the construction of film camera (1) and auxiliary cameras (6, 7) or after a change of the lens of the film camera (1), a calibration of a servo drive (3) is made who performs the focusing.

11. The method according to claim 10, characterized in that a current focus signal and optionally the range of the depth of field of the image calculation unit are transmitted.

12. The method according to claim 10 or 11, characterized in that the calibration is performed by the servo drive (3) is slowly adjusted and thereby by the film camera (1) sharply imaged sections are identified, these sections are compared with the auxiliary representation to preferably by pattern recognition to determine the respective focus value and store in an allocation table in relation to the rotational value of the servo drive (3).

13. The method according to any one of claims 1 to 12, characterized in that the focusing signal is calculated by pattern recognition.

14. The method according to any one of claims 1 to 13, characterized in that the focusing signal is determined as the distance of the nearest point.

15. An apparatus for focusing a film camera (1), with at least one auxiliary camera (6, 7) which generates an auxiliary representation, and with an image calculating device (8) which is connected on the one hand to the auxiliary camera (6, 7) and on the other hand, a focusing device (4) the film camera (1) controls, characterized in that the auxiliary camera (6, 7) is detachably connected to the film camera (1).

16. The apparatus according to claim 15, characterized in that at least two auxiliary cameras (6, 7) are connected to the film camera (1).

17. Device according to one of claims 14 or 15, characterized in that the optical axis of the auxiliary camera (6, 7) is parallel to the optical axis of the film camera (1).

18. Device according to one of claims 16 or 17, characterized in that at least two auxiliary cameras (6, 7) are arranged offset in the use position in the vertical direction against each other.

19. Device according to one of claims 15 to 18, characterized in that a display device (9) is provided, which is intended to output the auxiliary representations.

20. The device according to claim 19, characterized in that an input device (10) is provided for selecting a region which is used for calculating the focusing signal and that this region on the display device (9) can be displayed.

21. Device according to one of claims 15 to 20, characterized in that the image calculating means (8) performs a pattern recognition to calculate the focusing signal.

22. Device according to one of claims 15 to 21, characterized in that the at least one auxiliary camera is designed as a TOF camera.