DEVICE WITH A MULTI-APPERTURE IMAGING DEVICE FOR PRODUCEING A DEPTH MAPPING

Description

Zusammenfassung

[0001] The present invention relates to multi-aperture imaging devices, in particular a device comprising a multi-aperture imaging device. The device is configured to use the information contained by the multi-aperture imaging device to create a depth map and / or to accumulate image information. The present invention further relates to obtaining depth information from focus stacks using an array camera.

[0002] Multi-aperture imaging devices can image the object field using multiple partial fields of view. Concepts exist that, through the use of a beam deflection system, such as a mirror, allow the camera channels to deflect their line of sight from the device plane to another direction within the overall system, for example, approximately perpendicular to it. In the application of a mobile phone, this perpendicular direction could be towards the user's face or towards the surrounding environment, and can be achieved primarily using switchable, tilting mirrors.

[0003] DE 10 2016 204 148 A1 relates to a multi-aperture imaging device comprising at least one image sensor, an array of adjacent optical channels and a beam deflection device for deflecting a beam path of the optical channels.

[0004] WO 2018 188815 A1 relates to a device comprising an image sensor and an array of optical channels, each optical channel comprising optics for imaging a partial field of view of a total field of view onto an image sensor area of ​​the image sensor.

[0005] Devices for efficient image generation and / or devices for easy handling would be desirable.

[0006] The object of the present invention is therefore to create a device with a multi-aperture imaging device that enables efficient image creation and / or easy handling.

[0007] This problem is solved by the subject matter of independent claim 1.

[0008] One insight of the present invention is that by acquiring a sequence of images at a succession of focus positions, the depth map can be created from the image information itself, thus eliminating the need for disparity information and potentially dispensing with its use altogether. According to an embodiment of the first aspect, a device comprises a multi-aperture imaging device. The multi-aperture imaging device comprises an image sensor, an array of adjacent optical channels, each optical channel comprising optics for imaging at least a partial field of view of a total field of view onto an image sensor area of ​​the image sensor. The multi-aperture imaging device includes a beam deflection device for deflecting a beam path of the optical channels and a focusing device for setting a focus position of the multi-aperture imaging device.A control unit of the device is designed to control the focusing mechanism and to receive image information from the image sensor. The control unit is configured to guide the multi-aperture imaging device through a sequence of focus positions in order to acquire a corresponding sequence of image information of the entire field of view, and to generate a depth map for the acquired field of view based on this sequence of image information. The advantage of this is that the depth map can be generated from the sequence of focus positions, so that only a small number of optical channels are sufficient to obtain depth information.The control unit is configured to capture a corresponding number of groups of partial images in the sequence of focus positions, with each partial image assigned to a mapped partial field of view; to create the depth map from a comparison of local image sharpness information in the partial images; and to combine the partial images of a group of partial images into a complete image using the depth map. The control unit is configured to select areas of the total field of view in the preliminary depth map, based on a quality criterion, for which improvement is required, and to determine additional depth information to supplement the preliminary depth map for the selected areas, but not for unselected areas.

[0009] According to one embodiment of the first aspect, the control unit is designed to create the depth map from the sequence of image information, for example, without additional measurements or evaluation of the depth information by other methods. The advantage of this is that the depth map can be generated from the sequence of focus positions, so that a single image of the entire field of view from one viewing direction can provide sufficient information to create the depth map.

[0010] According to one embodiment of the first aspect, the optical channels are configured to capture the entire field of view at least stereoscopically. The control unit is configured to create a preliminary depth map based on disparity information obtained from the optical channels and to supplement this preliminary depth map with depth information derived from the sequence of image data to obtain the final depth map. Alternatively or additionally, the control unit is configured to create a preliminary depth map based on the sequence of image data and to supplement this preliminary depth map with depth information derived from disparity information obtained from the optical channels to obtain the final depth map. The advantage of this is that a highly accurate and high-quality depth map can be obtained, enabling very good stitching results.

[0011] According to one embodiment of the first aspect, the control unit is configured to select areas of the overall field of view in the preliminary depth map, based on a quality criterion, for which improvement is required, and to determine additional depth information to supplement the preliminary depth map for the selected areas, while not determining it for unselected areas. The advantage of this is that the additional expenditure of electrical energy and / or computational effort for supplementing the depth map by creating or determining additional depth information can be kept low, while still obtaining a high-quality depth map.

[0012] According to one embodiment of the first aspect, the control unit is designed to acquire a corresponding number of groups of partial images in the sequence of focus positions. Each partial image is assigned to a depicted partial field of view, so that each of the groups of partial images has a common focus position. The control unit is configured to perform a comparison of local image sharpness information in the partial images and to create the depth map from this comparison.This is made possible, for example, by using a known focus position, set by the control unit for capturing the group of partial images, and by identifying sharply focused objects—that is, objects located at the respective set focus position—information can be obtained indicating that the sharply focused image areas were captured at a distance from the multi-aperture imaging device that corresponds to the set focus position. By using multiple groups of partial images and thus multiple focus positions, corresponding information can be generated for different objects and therefore for the entire field of view, thus creating a depth map. This enables a large-scale or even complete mapping of the entire field of view with respect to depth information.

[0013] According to one embodiment of the first aspect, the device is designed to control the focusing mechanism such that the sequence of focus positions is distributed substantially equidistantly in the image space. This can be achieved through the most precise equidistant arrangement possible, but also by taking into account a tolerance range of up to ± 25%, ± 15%, or ± 5%, with the sequence of focus positions being distributed between a minimum and a maximum focus position. The equidistant position in the image area ensures uniform depth map precision over different distances.

[0014] According to one embodiment of the first aspect, the device is configured to generate a sequence of composite images representing the entire field of view based on the sequence of image information, with each composite image being based on a combination of partial images with the same focus position. This merging of partial images can be performed using the depth map to obtain high image quality in the assembled composite image.

[0015] According to one embodiment of the first aspect, the device is designed to modify an overall image representing the entire field of view based on the depth map. Knowing the depth map, various image manipulations can be performed, such as subsequent focusing and / or blurring of one or more image areas.

[0016] According to one embodiment of the first aspect, a first optical channel of the array is configured to image a first partial field of view of the total field of view, a second optical channel of the array is configured to image a second partial field of view of the total field of view, and a third optical channel is configured to image the entire total field of view. This allows the use of additional imaging functionalities, such as a zoom range and / or increased resolution.

[0017] According to one embodiment of the first aspect, the focusing device has at least one actuator for adjusting the focus position. The focusing device is arranged such that it is at least partially positioned between two planes spanned by the sides of a cuboid, the sides of which are aligned parallel to each other, to a line extension direction of the array, and to a portion of the beam path of the optical channels between the image sensor and the beam deflection device. The volume of the cuboid is minimal, yet designed to encompass the image sensor, the array, and the beam deflection device. This allows the multi-aperture imaging device to be designed with small dimensions along a depth direction normal to the planes.

[0018] According to an unclaimed embodiment, the multi-aperture imaging device has a thickness direction that is arranged perpendicular to the two planes. The actuator has a dimension parallel to the thickness direction. A portion of at most 50% of the actuator's dimension, extending from a region between the two planes, is arranged such that it projects beyond the two planes. By arranging the actuator to a extent of at least 50% between the planes, a thin design of the multi-aperture imaging device is achieved, which also enables a thin design of the device itself.

[0019] According to an unclaimed embodiment, the focusing device comprises an actuator for providing relative movement between an optic of at least one of the optical channels and the image sensor. This enables easy adjustment of the focus position.

[0020] According to an unclaimed embodiment, the focusing device is designed to execute the relative movement between the optics of one of the optical channels and the image sensor by simultaneously moving the beam deflection device. This allows the set optical influence to be maintained by the beam deflection device, for example, with regard to a beam-deflectoring area of ​​the beam deflection device that remains unchanged in size to deflect the beam path. This enables a smaller size for the beam deflection device, as it eliminates the need for larger areas at increased distances.

[0021] According to an unclaimed embodiment, the focusing device is arranged such that it protrudes from the area between the planes of the cuboid by a maximum of 50%. By arranging the entire focusing device, including any mechanical components and the like, in the area between the planes, a very thin design of the multi-aperture imaging device, and thus of the device as a whole, is made possible.

[0022] According to an unclaimed embodiment, the at least one actuator of the focusing device is a piezoelectric bending actuator. This enables the maintenance of a sequence of focus positions with a small time interval.

[0023] According to an unclaimed embodiment, the focusing device comprises at least one actuator configured to provide movement. The focusing device further comprises a mechanical device for transmitting the movement to the array for adjusting the focus position. The actuator is arranged on a side of the image sensor facing away from the array, and the mechanical device is arranged such that a force flow passes laterally past the image sensor. Alternatively, the actuator is arranged on a side of the beam deflection device facing away from the array, and the mechanical device is arranged such that a force flow passes laterally past the beam deflection device.This allows the multi-aperture imaging device to be designed such that the actuator is arranged laterally perpendicular to the thickness direction, thus neither blocking the beam paths of the optical channels nor simultaneously preventing the device from being detected. When using multiple actuators, exemplary embodiments provide for arranging all actuators on the side of the image sensor facing away from the array, all actuators on the side of the beam deflection device facing away from the array, or a subset of the actuators on the side of the image sensor facing away from the array and a disjoint subset thereof on the side of the beam deflection device facing away from the array.

[0024] According to an unclaimed embodiment, the relative position of the beam deflection device is switchable between a first position and a second position, such that in the first position the beam path is deflected towards a first total field of view, and in the second position the beam path is deflected towards a second total field of view. The control device is configured to control the beam deflection device to the first position in order to obtain imaging information of the first total field of view from the image sensor, and the control device is further configured to control the beam deflection device to the second position in order to obtain imaging information of the second total field of view from the image sensor.The control unit is further designed to insert a portion of the first imaging information into the second imaging information, in order to obtain accumulated image information that reproduces parts of the first total field of view and parts of the second total field of view. This allows for easy handling of the device, as it eliminates the need for time-consuming repositioning of the multi-aperture imaging device, for example, for self-photography against a background. This is particularly advantageous due to the self-generated depth map.

[0025] A second aspect described below is not part of the claimed invention. By combining image information from different total fields of view, so that the first total field of view is reproduced in some parts of an accumulated image, simple handling of the device can be achieved, since, for example, complex positioning of the user and / or the device can be avoided.

[0026] According to an embodiment of the second aspect, a device comprises a multi-aperture imaging device with an image sensor, an array of adjacent optical channels, and a beam deflection device. Each optical channel of the array includes optics for imaging at least a partial field of view of a total field of view onto an image sensor area of ​​the image sensor. The beam deflection device is configured to deflect a beam path of the optical channels, wherein the relative position of the beam deflection device is switchable between a first position and a second position, such that in the first position the beam path is deflected in the direction of a first total field of view, and in the second position the beam path is deflected in the direction of a second total field of view. The device further comprises a control device configured to control the beam deflection device to the first position.Controlled in this way, the control unit can obtain imaging information relating to the first total field of view projected onto the image sensor. The control unit is configured to move the beam deflection device to the second position to obtain imaging information for the second total field of view from the image sensor. The order in which the imaging information for the first total field of view and the imaging information for the second total field of view are obtained can be arbitrary. The control unit is configured to insert a portion of the first imaging information into the second imaging information to obtain accumulated image information that partially represents the first total field of view and partially represents the second total field of view.This allows for the combination of image content from different overall fields of view, thus eliminating the need for the complex positioning of the device and / or image objects.

[0027] According to one embodiment of the second aspect, the first overall field of view is arranged along a direction that corresponds to a user orientation of the device or to a world direction arranged oppositely to the device. This allows image content in the first overall field of view to be combined with an overall field of view that differs from it.

[0028] According to a preferred embodiment of the second aspect, the second total field of view is furthermore arranged along a direction that corresponds to the other total field of view from the user direction and the world direction, so that the two total fields of view together capture the world direction and the user direction. In the accumulated image information, content from the world direction and content from the user direction can thus be combined.

[0029] According to one embodiment of the second aspect, the control device is configured to identify and segment a person in the first image information, i.e., to separate or at least copy the image information relating to the person, and to insert the image of the person into the second image information to obtain the accumulated image information. This allows the image of the person to be inserted into an image environment that is actually arranged along a different direction of the device.

[0030] According to a preferred embodiment, the device is designed to automatically identify the person and to automatically insert the person's image into the second image information. This allows for a selfie to be taken against a different background. This eliminates the need for complex positioning of the device, the person, and / or the background. It also allows for compensation of the background's mirror inversion.

[0031] According to one embodiment of the second aspect, the device is configured to identify and / or segment the part, for example a person or at least a part thereof, using a depth map generated by the device from the first imaging information. The depth map can be created, for example, using the first aspect or by other means. This allows for simple implementation of the embodiment.

[0032] According to one embodiment of the second aspect, the device is configured to create a depth map with a plurality of depth levels for the second image information and to insert the first image information at a predetermined depth level of the second image information in order to obtain the accumulated image information. This enables the depth-correct integration of the first image information into the second image information with respect to the specified or predetermined depth level.

[0033] According to a preferred embodiment, the predetermined depth plane, within a tolerance range of 10%, is equal to the distance of the first total field of view from the device. This allows the accumulated image information to be retained in such a way that the second total field of view is displayed as if the first image information, or the section thereof, had been arranged along the opposite direction of the device.

[0034] According to a further preferred embodiment, the predetermined depth plane is based on user input associated with the placement of the first image information. This allows the depth plane to be varied between different shots and / or adapted to the user's selection via user input.

[0035] According to one embodiment of the second aspect, the device is configured to scale the first image information to obtain a scaled first image and to insert the scaled first image information into the second image information to obtain the accumulated image information. This makes it possible to insert the first image information into the second image information in such a way that a predetermined perception is preserved in the accumulated image information, particularly with regard to the size of the first image information. This is especially advantageous in combination with the adjustable depth plane into which the first image information is inserted, so that, in addition to depth-correct insertion, size-correct representation is also possible.

[0036] According to one embodiment of the second aspect, the device is configured to determine the distance of an object depicted in the first image information relative to the device, and to scale the first image information based on a comparison of the determined distance with the predetermined depth plane in the second image information. This makes it possible to automatically account for any change in the distance of the first image information due to the depth plane when inserting it into the second image information by means of scaling, i.e., adjusting the size.

[0037] According to an embodiment in accordance with the second aspect, the device is configured to acquire the first and second total fields of view within a time interval of at least 0.1 ms and at most 30 ms. The lower limit is optional. Such rapid acquisition of both total fields of view makes it possible to reduce or even avoid changes in the total fields of view caused by time.

[0038] According to an embodiment in accordance with the second aspect, the device is configured to receive the accumulated image information as a video data stream. For this purpose, the device can receive a plurality of accumulated image information for a plurality of sequential images of the first total field of view and / or the second total field of view and combine these into an image sequence as a video data stream.

[0039] Alternatively or additionally, embodiments in accordance with the second aspect provide for the accumulated image information to be made available as a single image or still image.

[0040] According to an embodiment consistent with the second aspect, the first image information comprises an image of a user, and the second image information comprises a world view of the device. The control unit is configured to segment an image of the user from the first image information, optionally based on depth map information generated by the device, and to insert the user image into the world view. This enables the easy acquisition of a self-portrait with the device.

[0041] According to one embodiment in accordance with the second aspect, the device is designed to insert the user's image into the world view with correct depth. This creates the impression that the user is standing in front of the world view without requiring the complex positioning required for this.

[0042] According to an embodiment in accordance with the second aspect, the device is configured to acquire a sequence of images of the first total field of view and / or the second total field of view at different focus positions, and to generate a depth map for the first total field of view and / or the second total field of view from the sequence of images. This enables, in particular, the combination of the second image information with the first image information in a predetermined depth plane and / or depth-correct imaging, thereby exploiting the advantages of the first aspect of the present invention. This means that the first aspect can be combined with embodiments of the second aspect and / or the second aspect can be combined with embodiments of the first aspect. In particular, the combination of the two aspects yields advantageous embodiments, which will be discussed later.

[0043] Further advantageous embodiments of the present invention are defined in dependent patent claims.

[0044] With reference to the accompanying drawings, preferred embodiments of the present invention are explained below. The drawings show: Fig. 1a a schematic perspective view of a device according to the first aspect; Fig. 1b a schematic perspective view of a device according to an embodiment of the second aspect; Fig. 1c a schematic perspective view of a device according to an embodiment combining the first aspect and the second aspect; Fig. 2a a schematic view of different focus positions according to an embodiment into which a device according to the first aspect can be controlled; Fig. 2b a schematic representation of the use of a depth map created from different focus positions according to an embodiment and its generation; Fig. 3a a schematic perspective view of a device according to an embodiment in which an image sensor spans an array of optical channels and a beam deflection device spans a cuboid in space; Fig. 3b a schematic side sectional view of the device made of Fig. 3a according to an embodiment in which the multi-aperture imaging device has a plurality of actuators; Fig. 3 is a schematic side sectional view of the multi-aperture imaging device. Fig. 3a and / or 3b, in which different overall fields of view can be detected based on different positions of the beam deflection device; Fig. 4a a schematic top view of a device according to an embodiment in which the actuator is formed as a piezoelectric bending actuator; Fig. 4b a schematic side sectional view of the device made of Fig. 4a to clarify the arrangement of the actuator between the planes of the cuboid, which is related to the Fig. 3a are described; Figs. 5a-d schematic representations of arrangements of partial fields of view in a total field of view, according to exemplary embodiments; Fig. 6 a schematic perspective view of a device according to an exemplary embodiment of the second aspect; Fig. 7a a schematic diagram to illustrate the processing of the image information that can be obtained according to an exemplary embodiment by imaging the total fields of view; Fig. 7 a schematic representation of a scaling of a part of an imaging information in the accumulated image information according to an exemplary embodiment; and Fig. 8 parts of a multi-aperture imaging device according to an exemplary embodiment, which can be used in devices of the first and / or second aspect according to the invention.

[0045] Before exemplary embodiments of the present invention are explained in detail below with reference to the drawings, it should be noted that identical, functionally equivalent or equivalent elements, objects and / or structures in the different figures are provided with the same reference numerals, so that the description of these elements shown in different exemplary embodiments is interchangeable or can be applied to one another.

[0046] Fig. 1a Figure 1 shows a schematic perspective view of a device 10 1 according to the first aspect. The device 10 1 comprises a multi-aperture imaging device, which includes an image sensor 12 and an array 14 of adjacent optical channels 16a-e. The multi-aperture imaging device further comprises a beam deflection device 18 for deflecting a beam path of the optical channels 16a-d. This allows the beam paths of the optical channels 16a-d to be deflected from a lateral path between the image sensor 12 through optics 22a-d of the array 14 towards the beam deflection device 18 to a non-lateral path. Different optical channels 16a-d are deflected in such a way that each optical channel 16a-d images a partial field of view 24a-d of a total field of view 26 onto an image sensor area 28a-d of the image sensor 12.The partial fields of view 24a-d can be distributed one-dimensionally or two-dimensionally, and, based on different focal lengths of the optics 22a-d, also three-dimensionally in space. For clarity, the total field of view 26 is described below as having a two-dimensional distribution of the partial fields of view 24a-d, whereby adjacent partial fields of view 24a-d can overlap. The combined area of ​​the partial fields of view constitutes the total field of view 26.

[0047] The multi-aperture imaging device comprises a focusing device 32 for adjusting the focus position of the multi-aperture imaging device. This can be achieved by changing a relative position between the image sensor 12 and the array 14, wherein the focusing device 32 can be configured to change a position of the image sensor 12 and / or a position of the array 14 in order to obtain a variable relative position between the image sensor 12 and the array 14 in order to adjust the focus position of the multi-aperture imaging device.

[0048] The relative position can be adjusted individually for each channel, for groups of optical channels, or globally for each channel. For example, a single optic 22a-d, a group of optics 22a-d, or all optics 22a-d together can be moved. The same applies to the image sensor 12.

[0049] The device comprises a control unit 34 configured to control the focusing device 32. Furthermore, the control unit 34 is configured to receive image information 36 from the image sensor 12. This information may, for example, be the partial fields of view 24a-d mapped onto the image sensor areas 28a-d, or information or data corresponding to the images. This does not preclude intermediate processing of the image information 36, for example, with regard to filtering, smoothing, or the like.

[0050] The control unit 34 is configured to control the multi-aperture imaging device through a sequence of focus positions in order to acquire a corresponding sequence of image information of the total field of view 26. The control unit 34 is configured to generate a depth map 38 for the total field of view 26 from the sequence of image information. The depth map 38 can be provided via a corresponding signal. Based on the different focus positions obtained by different relative positions between the image sensor 12 and the array 14, the control unit 34 can acquire different images of the same field of view 26, or differently focused partial images thereof, in accordance with the segmentation by the partial fields of view 24a-d.

[0051] Depth maps can be used for various purposes, such as image processing and image stitching. For example, the control unit 34 can be configured to combine individual images acquired from the image sensor areas 28a to 28d using the depth map 38 to obtain image information 42 that represents the entire field of view 26, i.e., a complete image. The use of a depth map is particularly advantageous for such methods of stitching partial images.

[0052] Using the depth map, the control unit can be configured to assemble the partial images of a group of partial images into a complete image. This means that the depth map used for stitching can be generated from the partial images to be stitched themselves. For example, based on the sequence of image information, a sequence of complete images representing the entire field of view can be generated. Each complete image can be based on a combination of partial images with the same focus position. Alternatively or additionally, at least two, several, or all of the complete images from the sequence can be combined to obtain a complete image with enhanced information, such as to create a bokeh effect. Alternatively or additionally, the image can also be displayed in such a way that the entire image is artificially sharp, meaning that a greater number of areas than in the individual images of partial areas are in focus, for example, the entire image.

[0053] The device 10 1 is configured to create the image of the entire field of view as a mono image and to generate the depth map 38 from the sequence of mono images. Although multiple scans of the entire field of view 26 are also possible, the device 10 can generate the depth map from a single mono image, thus saving the need for additional images from different viewing directions, for example, by using multiple images with the same device or by redundantly arranging additional optical channels.

[0054] Fig. 1b Figure 1 shows a schematic perspective view of a device 10 2 according to an embodiment of the second aspect. Compared with device 10 1, device 10 2 has a control unit 44 instead of the control unit 34, which is configured to control the beam deflection device to different positions 18 1 and 18 2. In the different positions 18 1 and 18 2, the beam deflection device 18 has a different relative position, so that imaging information from different total fields of view 26 1 and 26 2 is obtained in the different positions or orientations, since the beam paths of the optical channels 16a-d are directed in different directions influenced by the different positions 18 1 and 18 2.Alternatively or additionally to the control unit 34, the device 10 2 has a control unit 44 configured to control the beam deflection device to the first position 18 1 in order to obtain imaging information of the first total field of view 26 1 from the image sensor 12. Prior to or following this, the control unit 44 is configured to control the beam deflection device 18 to the second position 18 2 in order to obtain imaging information of the second total field of view 26 2 from the image sensor 12. The control unit 44 is configured to insert a portion of the first imaging information 46 1 into the second imaging information 46 2 in order to obtain a combined or accumulated image information 48. The accumulated image information 48 can reproduce the first total field of view 26 1 in some places and the second total field of view 26 2 in others, including image manipulation or image processing steps.This means that the accumulated image information 48 is based in some places on an image of the total visual field 26 1 and in other places on an image of the total visual field 26 2 .

[0055] The control unit 44 can be configured to provide a signal 52 containing or reproducing the accumulated image information 48. Optionally, the image information 461 and / or 462 can also be output by the signal 52.

[0056] Fig. 1c shows a schematic view of a device 10 3 according to an exemplary embodiment, which replaces the control device 34 made of Fig. 1a and instead of the control unit 44 from Fig. 1b a control unit 54 which combines the functionality of the control unit 34 and the control unit 44 and is configured to create the depth map 38 based on a variable focus position of the device 10 3 and to extract the accumulated image information 48 from Fig. 1b to provide.

[0057] Fig. 2a Figure 1 shows a schematic view of different focus positions 56 1 to 56 5, into which a device according to the first aspect, for example the device 10 1 and the device 10 2, can be controlled. The different focus positions 56 1 to 56 5 can be understood as positions or distances 58 1 to 58 5 at which objects in the recorded total field of view are sharply imaged onto the image sensor 12. The number of focus positions 56 can be arbitrary and may be greater than 1.

[0058] Distances 62 1 to 62 4 between adjacent focus positions can refer to distances in the image space, although it is also possible to extend or adapt the explanation to distances in the object space. However, considering the image space is advantageous because it takes into account the properties of the imaging multi-aperture device, particularly with regard to minimum and maximum object distances. The control unit 34 and / or 54 can be configured to control the multi-aperture device so that it has two or more focus positions 56 1 to 56 5. In each focus position, individual images 64 1 and 64 2 can be acquired according to the number of captured partial fields of view 24.Based on the knowledge of which focus positions 561 to 565 were set to obtain the respective partial images 461 and 462, the control unit can determine, by analyzing the image information regarding which parts of the image are in focus, the distance at which these in-focus objects are arranged relative to the device. This distance information can be used for the depth map 38. This means that the control unit can be configured to acquire a corresponding number of groups of partial images in the sequence of focus positions 561 to 565, with each partial image being assigned to a depicted partial field of view. The group of partial images can thus correspond to those partial images that depict the entire field of view in the set focus position.

[0059] The control unit is designed to create a depth map by comparing local image sharpness information in the partial images. Local sharpness information can indicate in which areas of the image objects are sharply rendered or are rendered sharply within a predefined tolerance range. For example, by determining the edge blur function and detecting the distances over which the edges extend, it can be determined whether a corresponding image area, object, or part thereof is rendered sharply or blurred on the image sensor. Furthermore, the point image or line blur function can be used as a quality criterion for the sharpness of an image element. Alternatively or additionally, any known optical sharpness metric, as well as the well-known Modulation Transfer Function (MTF), can be used.Alternatively or additionally, the sharpness of the same objects in neighboring images of the stack, the mapping of the focus actuator position to the object distance via a calibrated lookup table, and / or the direction of the through-focus scan can be used to partially recursively derive depth information from neighboring images of the stack and avoid ambiguities. Knowing the set focus position, which is uniquely correlated with an object distance that is rendered sharply, it is thus possible to deduce the distance of the area of ​​the image, the object, or a part thereof from the knowledge that the object is rendered sharply, at least within the previously defined tolerance range. This can form the basis for the depth map 38.

[0060] Using the depth map, the control unit is designed to assemble the partial images of a group of partial images into a complete image. This means that the depth map used for stitching can be generated from the partial images to be stitched themselves.

[0061] The device can be configured to control the focusing unit 32 such that the sequence of focus positions 56 1 to 56 5 is distributed equidistantly in the image space between a minimum focus position and a maximum focus position within a tolerance range of ± 25%, ± 15%, or ± 5%, preferably as close as possible to 0%. To save time when setting a focus position, it is advantageous, but not necessary, to sequentially control the focus positions 56 1 to 56 5 at increasing or decreasing intervals. Rather, the sequence of the set focus positions 56 1 to 56 5 is arbitrary.

[0062] Fig. 2b Figure 1 shows a schematic representation of the use and generation of the depth map 38. The partial images 641 and 642 can each be used to obtain partial information 381 to 385 of the depth map 38 from the respective focus positions 561 to 565, since the objects sharply displayed in the individual images 641 and 642 can be precisely determined with respect to their distance. However, interpolation methods can also be applied between the focus positions 561 and 565, so that even with slightly blurred objects, sufficiently accurate information for the depth map 38 can still be obtained. The distance information contained in the partial information 381 to 385 can be combined by the control unit to form the depth map 38.The depth map 38 can be used to combine the individual images 64 1 and 64 2 from the different focus positions 56 1 to 56 5 into a corresponding number of total images 42 1 to 42 5.

[0063] Fig. 3a Figure 1 shows a schematic perspective view of a device 30 according to an exemplary embodiment. The image sensor 12, the array 14, and the beam deflection device 18 can span a cuboid in space. The cuboid can also be understood as a virtual cuboid and can, for example, have a minimum volume and, in particular, a minimum vertical extent along a direction parallel to a thickness direction y, which is parallel to a line extension direction 66. The line extension direction 66 runs, for example, along a z-direction and perpendicular to an x-direction, which is arranged parallel to a path of the beams between the image sensor 12 and the array 14. The directions x, y, and z can span a Cartesian coordinate system. The minimum volume of the virtual cuboid, orThe minimum vertical extent of the virtual cuboid can be such that it still encompasses the image sensor 12, the array 14, and the beam deflection device 18. The minimum volume can also be understood as describing a cuboid spanned by the arrangement and / or operational movement of the image sensor 12, the array 14, and / or the beam deflection device 18. The line extension direction 66 can be arranged such that the optical channels 16a and 16b are positioned side by side, optionally parallel to each other, along the line extension direction 66. The line extension direction 66 can be fixed in space.

[0064] The virtual cuboid can have two sides that are opposite each other and parallel to the line extension direction 66 of the array 14, as well as parallel to a portion of the beam path of the optical channels 16a and 16b between the image sensor 12 and the beam deflection device 18. In simplified terms, but without any limiting effect, these can be, for example, a top and a bottom of the virtual cuboid. The two sides can span a first plane 68a and a second plane 68b. That is, each of the two sides of the cuboid can be part of plane 68a or 68b, respectively. Further components of the multi-aperture imaging device can be arranged completely, or at least partially, within the area between planes 68a and 68b, so that the space requirement of the multi-aperture imaging device along the y-direction, which is parallel to a surface normal of planes 68a and / or 68b, can be small, which is advantageous.The volume of the multi-aperture imaging device can have a small or minimal space between planes 68a and 68b. Along the lateral sides or extension directions of planes 68a and / or 68b, the space of the multi-aperture imaging device can be large or arbitrarily large. The volume of the virtual cuboid is influenced, for example, by an arrangement of the image sensor 12, the array 14, and the beam deflection device 18, wherein the arrangement of these components, according to the embodiments described herein, can be such that the space of these components along the direction perpendicular to the planes, and thus the distance between planes 68a and 68b, becomes small or minimal. Compared to other arrangements of the components, the volume and / or the distance on other sides of the virtual cuboid can be increased.

[0065] The device 30 comprises an actuator 72 for generating a relative movement between the image sensor 12, the single-row array 14, and the beam deflection device 18. This can, for example, be a positioning movement of the beam deflection device 18 for switching between the functions associated with the Fig. 1b The described positions include. Alternatively or additionally, actuator 72 can be configured to perform the functions associated with the Fig. 1a The relative movement described above is used to change the relative position between the image sensor 12 and the array 14. The actuator 72 is arranged at least partially between the planes 68a and 68b. The actuator 72 can be configured to move at least one of the image sensor 12, the single-row array 14, and the beam deflection device 18, which may involve rotational and / or translational movements along one or more directions. Examples include a channel-specific change in the relative position between image sensor areas 28 of a respective optical channel 16, the optics 22 of the respective optical channel 16, and the beam deflection device 18 or the corresponding segment or facet, and / or a channel-specific change in an optical property of the segment / facet relating to the deflection of the beam path of the respective optical channel.Alternatively or additionally, the Aktor 72 can at least partially implement autofocus and / or optical image stabilization.

[0066] The actuator 72 can be part of the focusing device 32 and configured to provide relative movement between at least one optic, at least one of the optical channels 16a and 16b, and the image sensor 12. The relative movement between the optic 22a and / or 22b and the image sensor 12 can be controlled by the focusing device 32 such that the beam deflection device 18 performs a simultaneous movement. If the distance between the optic 22a and / or 22b and the image sensor is reduced, the distance between the beam deflection device 18 and the image sensor 12 can be reduced accordingly, so that the relative distance between the array 14 or the optic 22a and / or 22b and the beam deflection device 18 remains essentially constant.This allows the beam deflection device 18 to be designed with small beam deflection areas, since a beam cone that grows due to an increasing distance between the array 14 and the beam deflection device 18 can be compensated for by maintaining the distance to the beam deflection device 18.

[0067] The focusing device 32 and / or the actuator 72 are arranged such that they project no more than 50% beyond the area between planes 68a and 68b. The actuator 72 can have a dimension or extent 74 parallel to the thickness direction y. A proportion of no more than 50%, no more than 30%, or no more than 10% of the dimension 74 can project beyond plane 68a and / or 68b from an area between planes 68a and 68b and thus protrude from the virtual cuboid. This means that the actuator 72 projects only minimally beyond plane 68a and / or 68b. According to exemplary embodiments, the actuator 72 does not project beyond planes 68a and 68b. The advantage of this is that the extension of the multi-aperture imaging device along the thickness direction y is not increased by the actuator 72.

[0068] Although the beam deflection device 18 is shown to be rotatably mounted about a rotational axis 76, the actuator 72 can alternatively or additionally generate a translational movement along one or more spatial directions. The actuator 72 can comprise one or more individual actuators, possibly to generate different individual movements that can be controlled separately. The actuator 72, or at least one individual actuator thereof, can, for example, be used as a component in connection with Fig. 4 The piezoelectric actuator, described in more detail, may be implemented or include a piezoelectric bending actuator. A piezoelectric bender enables rapid and reproducible position changes. This property advantageously allows for the acquisition of focus stacks in the sense of multiple or many images in a short time. Piezoelectric benders, as actuators extending along one dimension or direction, are particularly advantageous in the described architecture because they have a favorable form factor, i.e., an extension primarily in one direction.

[0069] The array 14 can comprise a substrate 78 to which the optics 22a and 22b are attached or arranged. The substrate 78 can be at least partially transparent for the beam paths of the optical channels 16a and 16b by means of cutouts or suitable material selection, although this does not preclude manipulations in the optical channels, for example by arranging filter structures or the like.

[0070] Several requirements for the actuator 72, including rapid adjustability for quickly setting the different focus positions 56, high force with low installation space requirements and the like, can be obtained by using piezoelectric actuators.

[0071] Fig. 3b Figure 1 shows a schematic side-section view of the device 30 according to an exemplary embodiment. The multi-aperture imaging device of the device 30 can, for example, have a plurality of actuators, e.g., more than one, more than two, or another number >0. For example, actuators 721 to 725 can be arranged, which can be used for different purposes, such as adjusting the focus position and / or changing the position of the beam deflection device 18, adjusting the viewing direction of the multi-aperture imaging device, and / or providing optical image stabilization by rotary movement of the beam deflection device 18 and / or translational movement of the array 14.

[0072] The actuators 721 to 725 can be arranged such that they are at least partially located between the two planes 68a and 68b spanned by sides 69a and 69b of the virtual cuboid 69. Sides 69a and 69b of the cuboid 69 can be aligned parallel to each other and parallel to the line extension direction of the array and a portion of the beam path of the optical channels between the image sensor 12 and the beam deflection device 18. The volume of the cuboid 69 is minimal, yet it encompasses the image sensor 12, the array 14, and the beam deflection device 18, as well as their operational movements. The optical channels of the array 14 have optics 22, which can be identical or different for each optical channel.

[0073] The volume of the multi-aperture imaging device can have a small or minimal installation space between planes 68a and 68b. Along the lateral sides or extension directions of planes 68a and / or 68b, the installation space of the multi-aperture imaging device can be large or arbitrarily large. The volume of the virtual cuboid is, for example, influenced by an arrangement of the image sensor 12, the single-row array 14, and the beam deflection device, wherein the arrangement of these components according to the embodiments described herein can be such that the installation space of these components along the direction perpendicular to the planes, and thus the distance between planes 68a and 68b, becomes small or minimal. Compared to other arrangements of the components, the volume and / or the distance between other sides of the virtual cuboid can be increased.

[0074] The virtual cuboid 69 is represented by dotted lines. The planes 68a and 68b can encompass or span two sides of the virtual cuboid 69. A thickness direction y of the multi-aperture imaging device can be arranged normal to the planes 68a and / or 68b and / or parallel to the y-direction.

[0075] The image sensor 12, the array 14, and the beam deflection device 18 can be arranged such that the perpendicular distance between the planes 68a and 68b along the thickness direction y, which can be simplified and referred to as the height of the cuboid without any limiting effect, is minimal, thus avoiding the need to minimize the volume, i.e., the other dimensions of the cuboid. The extent of the cuboid 69 along the y direction can be minimal and essentially determined by the extent of the optical components of the imaging channels, i.e., the array 14, the image sensor 12, and the beam deflection device 18, along the y direction.

[0076] The volume of the multi-aperture imaging device can have a small or minimal installation space between planes 68a and 68b. Along the lateral sides or extension directions of planes 68a and / or 68b, the installation space of the multi-aperture imaging device can be large or arbitrarily large. The volume of the virtual cuboid is, for example, influenced by an arrangement of the image sensor 12, the single-row array 14, and the beam deflection device, wherein the arrangement of these components according to the embodiments described herein can be such that the installation space of these components along the direction perpendicular to the planes, and thus the distance between planes 68a and 68b, becomes small or minimal. Compared to other arrangements of the components, the volume and / or the distance between other sides of the virtual cuboid can be increased.

[0077] The actuators 721 to 725 can each have a dimension or extent parallel to the direction y. A proportion of at most 50%, at most 30%, or at most 10% of the dimension of the respective actuator 721 to 725 can project beyond or out of the area between the two planes 68a and 68b, starting from a region between them. This means that the actuators 721 to 725 project at most insignificantly beyond the planes 68a and / or 68b. According to exemplary embodiments, the actuators do not project beyond the planes 68a and 68b. The advantage of this is that the extent of the multi-aperture imaging device along the thickness direction or direction y is not increased by the actuators.

[0078] Although terms such as above, below, left, right, front, or back are used here for clarity, they are not intended to have any limiting effect. It is understood that these terms are interchangeable based on a rotation or tilt in space. For example, the x-direction from the image sensor 12 towards the beam deflection device 18 can be understood as front or forward. A positive y-direction can, for example, be understood as above. An area along the positive or negative z-direction away from or spaced apart from the image sensor 12, the array 14, and / or the beam deflection device 18 can be understood as adjacent to the respective component. In simplified terms, an image stabilizer can be at least one of the actuators 721 to 725. This at least one actuator can be arranged in a plane 71 or between planes 68a and 68b.

[0079] In other words, the actuators 721 to 725 can be arranged in front of, behind, or beside the image sensor 12, the array 14, and / or the beam deflection device 18. According to exemplary embodiments, the actuators 36 and 42 are arranged with a maximum extent of 50%, 30%, or 10% outside the area between the planes 68a and 68b.

[0080] Fig. 3c Figure 1 shows a schematic side-section view of the multi-aperture imaging device, where different total fields of view 261 and 262 can be acquired based on different positions of the beam deflection device 18, since the multi-aperture imaging device then has different viewing directions. The multi-aperture imaging device can be configured to change the tilt of the beam deflection device by an angle α, so that alternately different main sides of the beam deflection device 18 are oriented towards the array 14. The multi-aperture imaging device can include an actuator configured to tilt the beam deflection device 18 about the axis of rotation 76. For example, the actuator can be configured to move the beam deflection device 18 into a first position in which the beam deflection device 18 deflects the beam path 26 of the optical channels of the array 14 in the positive y-direction.For this purpose, the beam deflection device 18 can, for example, have an angle α of > 0° and < 90°, of at least 10° and at most 80°, or of at least 30° and at most 50°, for example 45°, in the first position. The actuator can be configured to deflect the beam deflection device in a second position about the axis of rotation 76 such that the beam deflection device 18 deflects the beam path of the optical channels of the array 14 towards the negative y-direction, as shown by the viewing direction towards the overall field of view 26 2 and the dashed representation of the beam deflection device 18. For example, the beam deflection device 18 can be designed to be reflective on both sides, so that in the first position the viewing direction points towards the overall field of view 26 1.

[0081] Fig. 4a Figure 1 shows a schematic top view of a device 40 according to an exemplary embodiment, in which the actuator 72 is designed as a piezoelectric bending actuator. The actuator 72 is configured to perform a bend in the x / z plane, as indicated by the dashed lines. The actuator 72 is connected to the array 14 via a mechanical deflection device 82, so that bending of the actuator 72 allows a lateral displacement of the array 14 along the x-direction, thus changing the focus position. For example, the actuator 72 can be connected to the substrate 78. Alternatively, the actuator 72 can also be arranged on a housing that contains at least some of the optics 22a to 22d in order to move the housing. Other variants are also possible.

[0082] Optionally, the device can have 40 additional actuators 84 1 and 84 2 configured to generate movement on the array 14 and / or the beam deflection device 18, for example to position the beam deflection device 18 into different positions or orientations and / or for optical image stabilization by translational displacement of the array 14 along the z-direction and / or by generating a rotational movement of the beam deflection device 18 about the rotational axis 76.

[0083] Unlike the preceding figures, the beam deflection device 18 can have several spaced-apart but common movable facets 86a to 86d, with each optical channel being assigned to one facet 86a to 86d. The facets 86a to 86d can also be arranged directly adjacent to each other, i.e., with little or no distance between them. Alternatively, a plane mirror can also be arranged.

[0084] By actuating the actuator 72, a distance 88 1 between at least one of the optics 22a-d and the image sensor 12 can be changed from a first value 88 1 to a second value 88 2, for example increased or decreased.

[0085] Fig. 4b Figure 1 shows a schematic side-sectional view of the device 40 to illustrate the arrangement of the actuator 72 between levels 68a and 68b, which is related to the Fig. 3a are described. The actuator 72, for example, is arranged entirely between the planes 68a and 68b, as is the mechanical deflection device 82, which may have several force-transmitting elements, for example connecting webs, wires, ropes or the like, and mechanical bearings or deflection elements.

[0086] The mechanical deflection device or mechanical device for transmitting the movement to the array 14 can be arranged on a side of the image sensor 12 facing away from the array 14, i.e., originating from the array 14 behind the image sensor 12. The mechanical device 82 can be arranged such that a force flow runs laterally past the image sensor 12. Alternatively or additionally, the actuator 72 or another actuator can be arranged on a side of the beam deflection device 18 facing away from the array 14, i.e., originating from the array 14 behind the beam deflection device 18. The mechanical device 82 can be arranged such that a force flow runs laterally past the beam deflection device 18.

[0087] Although only one actuator 72 is shown, a higher number of actuators can also be arranged and / or more than one side of the actuator 72 can be connected to a mechanical deflection device 82. For example, a centrally mounted or supported actuator 72 can be connected to a mechanical deflection device 82 on both sides and, for example, act on both sides of the array 14 to enable homogeneous movement.

[0088] Fig. 5a Figure 1 shows a schematic representation of an arrangement of partial fields of view 24a and 24b within a total field of view 26, which can be detected, for example, by a multi-aperture imaging device described herein, such as the multi-aperture imaging device 101, 102, 103, 30, and / or 40, and which can correspond, for example, to the total field of view 261 and / or 262. For example, the total field of view 26 can be imaged onto the image sensor area 28b by optical channel 16b. For example, optical channel 16a can be configured to detect partial field of view 24a and image it onto the image sensor area 28a. Another optical channel, such as optical channel 16c, can be configured to detect partial field of view 24b and image it onto the image sensor area 28c. This means that a group of optical channels can be configured to capture exactly two partial fields of view 24a and 24b.This allows for the simultaneous recording of the total visual field and the partial visual fields, which together again represent the total visual field 26.

[0089] Although depicted with different dimensions for better differentiation, the partial visual fields 24a and 24b can have the same or comparable dimensions along at least one image direction B1 or B2, for example, along image direction B2. The dimensions of the partial visual fields 24a and 24b can be identical to the dimensions of the total visual field 26 along image direction B2. This means that the partial visual fields 24a and 24b can completely capture or encompass the total visual field 26 along image direction B2 and only partially capture or encompass the total visual field along another image direction B1 arranged perpendicular to it, and can be arranged offset from each other so that, combinatorially, a complete capture of the total visual field 26 is also achieved along the second direction.The partial fields of view 24a and 24b can be disjoint from each other or overlap at most incompletely in an overlap area 25, which may extend completely along the image direction B2 within the total field of view 26. A group of optical channels comprising optical channels 16a and 16c can be configured to completely image the total field of view 26, for example, by a single image combined with partial images that together image the total field of view. The image direction B1 can, for example, be a horizontal axis of an image to be provided. In simplified terms, the image directions B1 and B2 represent two different image directions arbitrarily positioned in space.

[0090] Fig. 5b Figure 1 shows a schematic representation of an arrangement of partial visual fields 24a and 24b, which are offset from each other along a different image direction, image direction B2, and overlap each other. Partial visual fields 24a and 24b can each completely capture the total visual field 26 along image direction B1 and incompletely along image direction B2. The overlap area 25, for example, is completely located within the total visual field 26 along image direction B1.

[0091] Fig. 5c Figure 1 shows a schematic representation of four partial visual fields 24a to 24b, which incompletely cover the total visual field 26 in both directions B1 and B2. Two adjacent partial visual fields 24a and 24b overlap in an overlap area 25b. Two overlapping partial visual fields 24b and 24c overlap in an overlap area 25c. Similarly, partial visual fields 24c and 24d overlap in an overlap area 25d, and partial visual field 24d overlaps with partial visual field 24a in an overlap area 25a. All four partial visual fields 24a to 24d can overlap in an overlap area 25e of the total visual field 26.

[0092] To capture the total visual field 26 and the partial visual fields 24a-d, a multi-aperture imaging device similar to that used in connection with Fig. 1a-c described, wherein the array 14 can, for example, have five optics, four for capturing partial visual fields 24a-d and one optic for capturing the entire visual field 26. Accordingly, the array can be configured in conjunction with the Fig. 5a-b It must be designed with three optical channels.

[0093] In the overlap areas 25a to 25e, a large amount of image information is available. For example, the overlap area 25b is captured via the total visual field 26, the partial visual field 24a, and the partial visual field 24b. An image format of the total visual field can correspond to a redundancy-free combination of the depicted partial visual fields, for example, partial visual fields 24a-d in Fig. 5c , whereby the overlapping areas 25a-e are each counted only once. In connection with the Fig. 5a und 5b This applies to the redundancy-free combination of the partial visual fields 24a and 24b.

[0094] An overlap in the overlap areas 25 and / or 25a-e can, for example, comprise a maximum of 50%, a maximum of 35% or a maximum of 20% of the respective sub-images.

[0095] In other words, the first aspect allows for a reduction in the number of optical channels, resulting in cost savings and a reduction in lateral installation space requirements. The first aspect also enables an alternative form of depth information acquisition to stereoscopic scanning, one that does not require additional sensors such as Time-of-Flight (TAF), Structured or Coded Light (SCL), or similar technologies. This avoids the need for low-resolution TAF sensors and high-energy-consumption Structured Light sensors. Both approaches still have limitations in strong ambient light, particularly sunlight. Exemplary embodiments demonstrate that the device operates without such sensors.According to one embodiment, a piezoelectric bender serves as an extremely fast, low-power focusing factor. However, other embodiments combine the acquisition of depth information from a sequence of focus positions with the acquisition of depth information from disparity-based depth information. Preferably, a disparity-based depth map is created first, and if this map has deficiencies, it is supplemented, corrected, or improved by the additional depth information obtained from the sequence of focus positions. The described architecture of the multi-aperture imaging device enables the use of such piezoelectric benders, since the otherwise cubic form factor of the camera module makes the use of long piezoelectric benders difficult or even impossible. With short exposure times, this allows the acquisition of focus stacks, i.e.,Numerous images are taken in rapid succession, each with a slightly different focus on the scene. Examples of this technique involve scanning the entire depth of the scene, from macro (the closest possible shot) to infinity (the furthest possible distance). The intervals can be equidistant in object space, but preferably in image space. Alternatively, another suitable interval can be chosen. The number of focus positions can be, for example, at least two, at least three, at least five, at least ten, at least 20, or any other arbitrary number.

[0096] The user can be shown multiple images. Alternatively or additionally, embodiments provide for combining the individual image information so that the user can be provided with an image that contains combined image information. For example, an image with depth information, which offers the possibility of digital refocusing. The displayed image can offer a so-called bokeh effect, a blurring of the background. Alternatively, the image can also be displayed in such a way that the entire image is artificially sharp, meaning that a larger distance range than in the individual images of partial areas is in focus, for example, the entire image. With a small f-number of the lenses used, the sharpness measured in the individual images can result in a different image.Blur and other information, such as the sharpness of the same objects in neighboring images of the stack, an assignment of the focus actuator position to an object distance, for example using a calibrated lookup table, a direction of the focus position sequence (through-focus scan) both on its own and recursively from other images to avoid ambiguities, the object distance of the individual elements of the scene can be reconstructed and a depth map in image resolution can be created from this.

[0097] The first aspect achieves the elimination of channel duplication for stereo imaging while still enabling the creation of a depth map. This depth map allows for the stitching of the different partial images from the multi-aperture imaging device. For example, halving the number of optical channels significantly reduces the lateral dimensions, such as along the numerical extension direction, thus reducing costs. Image processing can deliver images of at least equal quality through other steps. Alternatively or additionally, the need for additional Time-of-Flight or Structured Light sensors is eliminated. This advantage remains even if the aforementioned duplication is ultimately performed, which can also offer benefits.

[0098] For example, the entire visual field 26 can be captured at least stereoscopically, as described, for example, in DE 10 2013 222 780 A1, in order to obtain depth information from multiple, i.e., at least two, captures of the entire visual field or the partial visual fields 24. The at least stereoscopic capture, which is described, for example, in Fig. 5d As illustrated, depth information can be obtained by viewing the same partial field of view 24a or 24b through two optical channels 16a and 16c or 16c and 16d spaced apart by a base distance BA. The number of partial fields of view and their arrangement are freely selectable; see, for example, the differences in the Fig. 5a-c . Advantageous with regard to the avoidance of occlusions is the in Fig. 5d arrangement of the partial fields of vision shown according to Fig. 5b .

[0099] In Fig. 5d Only a portion of a multi-aperture imaging device 50 is shown. Elements such as the beam deflection device 18 or actuators are not depicted. The multi-aperture imaging device 50 now has two sources of depth information that can be used to create a depth map. These are, firstly, the control of the multi-aperture imaging device through a sequence of focus positions, and secondly, the disparity between optical channels for capturing a corresponding image content.

[0100] This can offer advantages insofar as every information source combines advantages with disadvantages. For example, disparity-based depth information can be incomplete or of low quality in places due to occlusions or occlusions, but in contrast to the sequence of focus positions, it is fast and requires little electrical energy for actuators and / or computing power (which requires both corresponding computing resources and electrical energy).

[0101] Therefore, exemplary embodiments provide for combining the depth information from both information sources. For example, a preliminary depth map can be created from the disparity-based depth information and supplemented or improved by a fully or partially generated additional depth map from the sequence of focus positions. In this context, "preliminary depth map" does not necessarily describe a temporal relationship, as the order in which the two depth maps to be combined are generated can be arbitrary. According to less preferred embodiments, the preliminary depth map can be generated from the sequence of focus positions and improved or enhanced by disparity-based depth information.

[0102] The control unit 34 or another instance of the multi-aperture imaging device 50 can be configured to verify the quality or reliability of the depth information or depth map, for example, by checking the resolution of the depth information and / or monitoring the occurrence of occlusions or other effects. For affected areas of the depth map, additional depth information can be generated from the sequence of focus positions to improve or correct the preliminary depth map. For example, this supplementary depth information can be obtained in an energy- and / or computationally efficient manner by generating the sequence of focus positions only for the areas of the preliminary depth map to be supplemented, i.e., only in a sub-region of the depth map. By improving the depth map and thus the depth information, high-quality stitching results can be obtained.

[0103] Exemplary implementations provide that the control unit specifies a local area and / or a range of depth planes for determining depth information based on the sequence of focus positions, i.e., a value range between minimum and maximum focus positions based on the points to be supplemented or corrected in the preliminary depth map, whereby multiple areas can also be set. This means that at least some of the possible focus positions, or those set for creating the depth map solely from the sequence of image information or focus positions, can be omitted, which can save computational effort, time, and electrical energy.Alternatively or in addition to limiting the number and / or position of the focus positions, the control device can be designed to recalculate the depth information only for those areas of the total field of view 26 where an improvement, optimization or correction of the preliminary depth map is required, which can also save energy and time.

[0104] The control unit can be configured to select areas of the overall field of view in the preliminary depth map, based on a quality criterion, that require improvement, and to supplement the preliminary depth map in the selected areas but not in unselected areas. For example, the additional depth information for supplementing the preliminary depth map can be determined only for the selected areas and not for unselected areas. For example, the control unit can be configured to determine at least one area and / or the focus positions by performing a comparison that indicates whether the quality or reliability of the depth map in the specified area meets at least a threshold value.This can be achieved by evaluating a reference parameter to determine whether at least a minimum quality is reached, but also by checking whether a negative quality criterion (such as the number of errors or the like) is not exceeded, i.e., whether a threshold is not undercut or at least not exceeded.

[0105] The aforementioned preliminary depth map can be a depth map created using the available depth information. Alternatively, the preliminary depth map can be understood as a collection of depth information (without a specific map format).

[0106] In other words, in addition to generating the depth map solely through focus stacks, without redundant field-of-view acquisition (for generating the parallax for the depth map), this approach can be applied to systems with a minimal number of channels (ideally only two). This concept can be modified to support such an architecture (e.g., 2x field of view above, 2x field of view below as in Fig. 5d (represented, or implemented differently) by means of a depth map generated by means other than disparity, to improve the latter, which can ultimately also contribute to improved stitching results. Here, implementation examples based on the aspect of focus stacks for multimodal support are preferred, avoiding other mechanisms such as depth mapping via Time of Flight (ToF) or coded light.

[0107] These embodiments relate to the aspect of combining depth maps derived from disparity and focus stacks. They assume the described architecture with more than two channels, which primarily derives a depth map from the naturally occurring disparity / parallax of the channels. Depending on the available computing and / or power budget, either a further depth map can always be generated from focus stacks and combined with the first depth map to improve it (essentially by filling in holes at overlaps) and to improve the stitching results, or preferably only after obvious deficiencies in the depth map derived from disparity have been identified. Another sequence is less preferred, as generating the focus stacks can involve additional energy expenditure and potentially significant time loss, resulting in considerably more complex exposure conditions for the sequential images.Advantages that may arise from the additional use of a depth map generated from FocusStacks, images acquired extremely quickly in succession, include: fewer holes in the depth map due to occlusions, possibly additional depth planes, especially for larger object distances, possibly improved lateral resolution of the depth map and overall improved signal-to-noise ratio in the depth map due to additional information acquisition, and thus fewer ambiguities or even multiple ambiguities that would otherwise lead to artifacts in the stitched images of the entire field of view.

[0108] Fig. 6 Figure 1 shows a schematic perspective view of a device 60 according to an exemplary embodiment with regard to the second aspect. The described embodiments also apply to devices 101, 103, 30 and / or 40. By controlling the beam deflection device to different positions, the device 60 or the multi-aperture imaging device of the device 60 can capture two spaced-apart entire fields of view 261 and 262.

[0109] The device 60 is, for example, designed as a portable or mobile device, in particular a tablet computer or a mobile phone, especially a smartphone (smartphone).

[0110] One of the fields of view 26 1 and 26 2 can, for example, be arranged along a user direction of the device 60, as is common, for example, in the context of self-portraits (selfies) for photos and / or videos.

[0111] The other total field of view can, for example, be arranged along an opposite direction and / or a world direction of the device 60 and, for example, be arranged along the direction along which the user looks when looking at the device 60 from the total field of view along the user direction.

[0112] For example, the beam deflection device 18 can be in Fig. 1b be formed reflectively on both sides and, for example, deflect the beam path of the optical channels 16a-d in different positions with different main sides, so that, starting from the device 60, the total fields of view 26 1 and 26 2 are arranged opposite each other and / or at an angle of 180°.

[0113] Fig. 7a Figure 1 shows a schematic diagram illustrating the processing of the image information 461 and 462, which can be obtained by mapping the total visual fields 261 and 262. The control unit is configured to separate a part 92 of the mapping information 461 of the visual field 261, for example, by cutting it out, isolating it, or copying only part 92. The control unit is further configured to combine the separated or segmented part 92 with the mapping information 462, i.e., to insert part 92 into the mapping information 462 to obtain the accumulated image information 48. This accumulated image information 48 contains the total visual field 262 in some areas and the image information 461 in others, namely where part 92 was inserted.It is pointed out that the preservation of the accumulated image information 48 is not limited to the insertion of a single part 92, but that any number of parts 92 can be segmented from the image information 46 1 and one, several or all of these parts can be inserted into the image information 46 2.

[0114] A location or position where part 92 is inserted into the second imaging information 46 2 can be automatically determined by the control device, for example by projecting part 92 through the device 60 into the second field of view 26 2, but can alternatively or additionally also be selected by a user.

[0115] According to one embodiment, the control unit is configured to identify and segment a person in the first image information 46 1, for example, by pattern matching and / or edge detection, but especially based on the depth map generated by the device itself. The control unit can be configured to insert the image of the person into the second image information 46 2 to obtain the accumulated image information 48. This means that part 92 can be a person, such as a user of the device 60. Embodiments provide that the device is configured to automatically identify the person and to automatically insert the image of the person, i.e., part 92, into the second image information 46 2.This makes it possible to automatically take a self-portrait or self-shot in front of or within the second total field of vision 26 2, without having to laboriously position the device 60 and / or without having to laboriously position the user.

[0116] Exemplary embodiments provide that the control device uses a depth map, such as the depth map 38, to position the part 92 in the second imaging information 46 2. The depth map 38 can have a plurality or multiple depth planes, for example, according to the number of focus positions considered, or a reduced number obtained therefrom, or a higher number interpolated therefrom. The control device can be configured to insert the part 92 in the predetermined depth plane of the second imaging information 46 2 in order to obtain the accumulated image information 48. The predetermined depth plane can essentially correspond, i.e., within a tolerance range of ± 10%, ± 5%, or ± 2%, to a distance of the first total field of view 26 2 from the device 60 or to the distance of the segmented part 92 from the device 60.This can also be described as correctly inserting part 92 into the second figure information 46 2.

[0117] Fig. 7b Figure 1 shows a schematic representation of the scaling of part 92 in the accumulated image information 48. Alternatively, a different depth level can be selected, for which various embodiments are provided. For example, the predetermined depth level can be influenced by or determined by the placement of part 92 in the second image information 46 2. The placement can be automatic or user-initiated. If, for example, the user selects a specific location or position within the second image information 46 2 for inserting part 92, the control device can be configured to determine the distance of the area in which part 92 is to be inserted within the second image information 46 2.Knowing the distance of part 92 to the device and the objects in the second image information 46 2, for example using depth maps, a virtual change in the distance of part 92 caused by user input can be compensated for by scaling part 92.

[0118] Thus, a one-dimensional, two-dimensional, or three-dimensional size 94 of part 92 can be changed to a size 96, for example, reduced if the distance of part 92 from the first image information 46 1 to the second image information 46 2 is increased, or increased if the distance from the first image information 46 1 to the second image information 46 2 is decreased. Independently of this, but also in combination with a placement of part 92 in the first image information 46 1 based on an associated user input, the device can be configured to scale the image information 46 1 to obtain scaled image information. The scaled image information can be inserted by the control unit into the image information 46 2 to obtain the accumulated image information 48.The device can be configured to determine the distance of an object representing part 92, depicted in the first image information 46 1, relative to the device 60. The device can scale image information 46 1, or part 92 thereof, based on a comparison of the determined distance with the predetermined depth plane in the second image information 46 2. It is advantageous if the two image information pieces 46 1 and 46 2 are acquired at short intervals. It is advantageous if this interval is approximately 0.1 ms, within a time interval of at most 30 ms, at most 10 ms, at most 5 ms, or at most 1 ms. This time can, for example, be used for switching or repositioning the beam deflection device and can be at least partially determined by the duration of this process.

[0119] The accumulated image information 48 can be obtained as a single image, or alternatively or additionally as a video data stream, for example as a large number of individual images.

[0120] According to one embodiment, a device is configured in accordance with the second aspect such that the first image information 46 1 comprises an image of a user and the second image information 46 2 comprises a world view of the device. The control unit is configured to segment an image of the user from the first image information 46 1 and insert it into the world view. For example, the device may be configured to insert the image of the user into the world view with correct depth.

[0121] In other words, in connection with the second aspect, taking a selfie image or selfie video can involve a depth-based combination of quasi-simultaneous image captures using the front-facing camera (front-facing camera / view) and the rear-facing camera (main camera / view) of a device, particularly a mobile phone. The foreground of the selfie image, i.e., the self-portrait, can be superimposed onto the foreground of the main camera image. A very rapid switch between front-facing and rear-facing images, achieved by changing the position of the beam deflection device, enables the aforementioned quasi-simultaneous capture of the world-facing and user-facing camera images using the same image sensor.Although a single-channel imaging device can also be used according to the second aspect, the second aspect offers advantages, particularly with regard to multi-aperture imaging devices, since these can already create or use a depth map to stitch the individual images together. This depth map can also be used to determine depth information for synthesizing the accumulated imaging information 48. A process is enabled that can be described as follows: 1. Use the depth map of the selfie to segment the foreground, i.e., the person(s) taking the selfie, from the background; 2. Use the depth map of the world-side image to determine a foreground and a background from it, i.e., to separate depth information; and 3. Add the foreground, i.e., the person(s) taking the selfie, from the selfie image into the world-side image, specifically its foreground.

[0122] The advantage of this is that the selfie can be combined with the world-viewed image as a background without having to rotate the phone 180°, as would otherwise be necessary, to include oneself in the scene. Alternatively or additionally, it avoids the problem of shooting backwards, which requires always thinking in reverse regarding the phone's orientation relative to the scene. Furthermore, the depth map can be generated in the same way as described in the first aspect, thus eliminating the need for additional Time-of-Flight or Structured Light sensors.

[0123] The following section refers to some advantageous embodiments of the multi-aperture imaging device to explain the advantages of the invention.

[0124] Fig. 8 Figure 80 shows parts of a multi-aperture imaging device 80 which can be used in devices of the first and / or second aspect according to the invention, wherein a possible focusing device and / or actuator for implementing optical image stabilization is not shown, but can be implemented without difficulty.

[0125] The multi-aperture imaging device 80 from Fig. 8 comprises a multi-row or preferably single-row array 14 of adjacent optical channels 16a-d. Each optical channel 16a-d comprises an optic 22a-d for imaging a respective partial field of view 24a-d of a total field of view 26, optionally also of a total field of view, as described in connection with Fig. 5 The imaged field of view of the multi-aperture imaging device 80 is mapped onto a respective assigned image sensor area 28a-d of an image sensor 12.

[0126] The image sensor areas 28a-d can, for example, each be formed from a chip comprising a corresponding pixel array, with the chips arranged as in the Fig. 8 As indicated, the image sensor areas 28a-d can be mounted on a common substrate or circuit board 98. Alternatively, it would also be possible for each image sensor area 28a-d to be formed from a portion of a common pixel array that extends continuously or intermittently across the image sensor areas 28a-d, with the common pixel array being formed, for example, on a single chip. In this case, for instance, only the pixel values ​​of the common pixel array in the image sensor areas 28a-d are read out. Various combinations of these alternatives are also possible, such as the presence of one chip for two or more channels and another chip for yet other channels, or the like. In the case of multiple image sensor chips 12, these can, for example, be mounted on one or more circuit boards, such as all together, in groups, or the like.

[0127] In the exemplary embodiment of Fig. 8 Four optical channels 16a-d are arranged side-by-side in a single row along the row direction of the array 14, but the number four is merely exemplary and could also assume any other number greater than one; that is, N optical channels with N>1 can be arranged. Furthermore, the array 14 can also have additional rows extending along the row direction. The array 14 of optical channels 16a-d is understood to be a combination of the optical channels or a spatial grouping thereof. The optics 22a-d can each comprise a lens, a lens assembly or stack, or a combination of an imaging optic with other optical elements, including filters, apertures, reflective or diffractive elements, or the like. The array 14 can be configured such that the optics 22a-d are arranged, attached, or mounted on the substrate 78 individually, in groups, or globally (i.e., all channels together).This means that a single substrate 78, several parts thereof, or no substrate 78 can be arranged, for example if the optics 22a-d are held elsewhere.

[0128] Optical axes, or the beam paths 102a-d of the optical channels 16a-d, can, according to an example, run parallel to each other between the image sensor areas 28a-d and the optics 22a-d. For this purpose, the image sensor areas 28a-d are arranged, for example, in a common plane, as are the optical centers of the optics 22a-d. Both planes are parallel to each other, i.e., parallel to the common plane of the image sensor areas 28a-d. Furthermore, when projected perpendicularly onto the plane of the image sensor areas 28a-d, the optical centers of the optics 22a-d coincide with the centers of the image sensor areas 28a-d. In other words, in these parallel planes, the optics 22a-d and the image sensor areas 28a-d are arranged with the same repetition spacing in the line extension direction.

[0129] The image-side distance between image sensor areas 28a-d and the associated optics 22a-d is set such that the images onto the image sensor areas 28a-d are focused at a desired object distance. This distance is, for example, equal to or greater than the focal length of the optics 22a-d, or between one and two times the focal length of the optics 22a-d, including both. The image-side distance along the optical axis 102a-d between image sensor area 28a-d and optics 22a-d can also be adjustable, for example, manually by a user and / or automatically via a focusing device or autofocus control.

[0130] Without additional measures, the partial fields of view 24a-d of the optical channels 16a-d overlapped essentially completely due to the parallelism of the beam paths or optical axes 102a-d. To cover a larger overall field of view 26 and to ensure that the partial fields of view 24a-d only partially overlap spatially, the beam deflection device 18 is provided. The beam deflection device 18 deflects the beam paths 102a-d or optical axes, for example, with a channel-specific deviation, into an overall field of view direction 104. The overall field of view direction 104 runs, for example, parallel to a plane that is perpendicular to the line extension direction of the array 14 and parallel to the course of the optical axes 102a-d before or without beam deflection.For example, the overall field of view direction 104 is derived from the optical axes 102a-d by rotation about the line extension direction by an angle that is > 0° and < 180° and, for example, lies between 80° and 100° and can be, for example, 90°. The overall field of view 26 of the multi-aperture imaging device 80, which corresponds to the total coverage of the partial fields of view 24a-d, is therefore not located in the direction of an extension of the cascaded arrangement of the image sensor 12 and the array 14 in the direction of the optical axes 102a-d, but, due to the beam deflection, the overall field of view is located laterally to the image sensor 12 and array 14 in a direction in which the height of the multi-aperture imaging device 80 is measured, i.e., the lateral direction perpendicular to the line extension direction.

[0131] In addition, the beam deflection device 18 deflects, for example, each beam path or the beam path of each optical channel 16a-d with a channel-specific deviation from the deflection leading to direction 104 mentioned above. For this purpose, the beam deflection device 18 comprises, for example, an individually configured element for each channel 16a-d, such as a reflective facet 86-d and / or a reflective surface. These are slightly inclined relative to each other. The relative tilt of the facets 86a-d is selected such that, when the beam is deflected by the beam deflection device 18, the partial fields of view 24a-d are provided with a slight divergence, such that the partial fields of view 24a-d only partially overlap. As can be seen, for example, in Fig. 8As indicated, the individual deflection can also be designed in such a way that the partial visual fields 24a-d cover the total visual field 26 two-dimensionally, i.e., are arranged two-dimensionally distributed in the total visual field 26.

[0132] According to another example, the optics 22a-d of an optical channel can be configured to generate the divergence in the beam paths 102a-d completely or partially, thus making it possible to dispense with the inclination between individual facets 86a-d completely or partially. If the divergence is provided entirely by the optics 22a-d, for example, the beam deflection device can also be formed as a plane mirror.

[0133] It should be noted that many of the details described so far regarding the multi-aperture imaging device 80 were chosen only as examples. This already applied, for instance, to the previously mentioned number of optical channels. The beam deflection device 18 can also be configured differently than described so far. For example, the beam deflection device 18 does not necessarily have to be reflective. It can therefore be designed differently than in the form of a faceted mirror, such as transparent prism wedges. In this case, for example, the mean beam deflection could be 0°, meaning that the direction 104 could, for example, be parallel to the beam paths 102a-d before or without beam deflection. In other words, the multi-aperture imaging device 80 could still "look straight ahead" despite the beam deflection device 18.The channel-specific deflection by the beam deflection device 18 would again result in the partial fields of view 24a-d only slightly overlapping each other, e.g. in pairs with an overlap < 10 % with respect to the solid angle areas of the partial fields of view 24a-d.

[0134] The beam paths 102a-d, or optical axes, could also deviate from the described parallelism, yet the parallelism of the beam paths of the optical channels could still be so pronounced that the partial fields of view covered by the individual channels 16a-N, or mapped onto the respective image sensor areas 28a-d, would largely overlap without further measures, such as beam deflection. Therefore, in order to cover a larger overall field of view with the multi-aperture imaging device 80, the beam deflection device 18 provides the beam paths with an additional divergence, such that the partial fields of view of N optical channels 16a-N overlap less. For example, the beam deflection device 18 ensures that the overall field of view has an opening angle greater than 1.5 times the opening angle of the individual partial fields of view of the optical channels 16a-N.With a kind of preliminary divergence of the beam paths 102a-d, it would also be possible that, for example, not all facet inclinations differ, but that some groups of channels possess facets with the same inclination. The latter can then be formed in one piece or continuously merging into one another, as if as a single facet assigned to this group of channels adjacent in the line extension direction.

[0135] The divergence of the optical axes 102a-d of these channels 16a-d could then originate from the divergence of these optical axes 102a-d as achieved by lateral offset between the optical centers of the optics 22a-d and the image sensor areas 28a-d of the channels 16a-d, or prism structures, or decentered lens sections. The pre-divergence could, for example, be limited to one plane. The optical axes 102a-d could, for example, run in a common plane before or without beam deflection 18, but divergent within this plane, and the facets 86a-d merely cause an additional divergence in the other transverse plane, i.e., they are all parallel to the line extension direction and only inclined differently to each other with respect to the aforementioned common plane of the optical axes 102a-d, whereby several facets 86a-d can again have the same inclination.could be jointly assigned to a group of channels whose optical axes already differ pairwise before and without beam deflection, for example, in the aforementioned common plane of optical axes.

[0136] By omitting the beam deflection device 18 or designing the beam deflection device 18 as a planar mirror or the like, the entire divergence could also be achieved by the lateral offset between optical centers of the optics 22a-d on the one hand and centers of the image sensor areas 28a-d on the other hand, or by prism structures or decentered lens sections.

[0137] The aforementioned possible pre-divergence can be achieved, for example, by having the optical centers of the optics 22a-d lie on a straight line along the line extension direction, while the centers of the image sensor areas 28a-d are arranged differently from the projection of the optical centers along the normal of the plane of the image sensor areas 28a-d onto points on a straight line in the image sensor plane, such as at points that deviate from the points on the aforementioned straight line in the image sensor plane channel-specifically along the line extension direction and / or along the direction perpendicular to both the line extension direction and the image sensor normal.Alternatively, pre-divergence can be achieved by having the centers of the image sensors 28a-d lie on a straight line along the line extension direction, while the centers of the optics 22a-d are arranged differently from the projection of the optical centers of the image sensors along the normal of the plane of the optical centers of the optics 22a-d onto points on a straight line in the optics center plane, such as at points that differ from the points on the aforementioned straight line in the optics center plane channel-specifically along the line extension direction and / or along the direction perpendicular to both the line extension direction and the normal of the optics center plane.

[0138] It is preferred if the aforementioned channel-specific deviation from the respective projection occurs solely in the line extension direction, meaning that the optical axes 102a-d lie only in a common plane and are provided with a pre-divergence. Both optical centers and image sensor area centers then lie on a straight line parallel to the line extension direction, but with different distances between them. In contrast, a lateral offset between lenses and image sensors in a perpendicular lateral direction to the line extension direction leads to an increase in the overall height. A purely in-plane offset in the line extension direction does not change the overall height, but may result in fewer facets and / or the facets exhibiting only a tilt in one angular orientation, which simplifies the design.

[0139] Although some aspects have been described in connection with a device, it is understood that these aspects also constitute a description of the corresponding process, so that a block or component of a device is also to be understood as a corresponding process step or as a feature of a process step. Similarly, aspects described in connection with or as a process step also constitute a description of a corresponding block, detail, or feature of a corresponding device.

[0140] The embodiments described above merely illustrate the principles of the present invention. It is understood that modifications and variations of the arrangements and details described herein will be obvious to other people skilled in the art. Therefore, it is intended that the invention be limited only by the scope of protection set forth in the following claims and not by the specific details presented herein by way of description and explanation of the embodiments.

Claims

1. A device comprising: a multi-aperture imaging device comprising: an image sensor (12); an array (14) of adjacently arranged optical channels (16a-d), each optical channel (16a-d) including optics (22a-d) for projecting at least a partial field of view (24a-d) of a total field of view (261, 262) on an image sensor area (28a-d) of the image sensor (12), a beam deflector (18) for deflecting a beam path (104) of the optical channels (16a-d), focusing means (32) for adjusting a focal position of the multi-aperture imaging device; the device further comprising: control means (34; 54) configured to control the focusing means (32) and to receive image information (36) from the image sensor (12); wherein the control means (34; 54) is configured to control the multi-aperture imaging device to move to a sequence of focal positions (56) so as to detect a corresponding sequence of image information (36) of the total field of view (261, 262), and to produce a depth map (38) for the detected total field of view on the basis of the sequence of image information (36); wherein the control means (34; 54) is configured to detect, in the sequence of focal positions (56), a corresponding number of groups of partial images (64), each partial image (64) being associated with a projected partial field of view (24a-d); to produce the depth map (38) from a comparison of local image sharpness information in the partial images (64); and to assemble, while using the depth map (38), the partial images (64) of a group of partial images into a total image (42); to select areas of the total field of view in the preliminary depth map on the basis of a quality criterion for which an improvement is required, and to determine additional depth information for supplementing the preliminary depth map for the selected areas and not to determine them for non-selected areas.

2. The device as claimed in claim 1, wherein the quality criterion indicates whether at least a minimum quality is achieved or whether a negative quality criterion is not exceeded.

3. The device as claimed in claim 1 or 2, wherein each partial image is associated with a projected partial field of view and each of the groups of partial images comprises a common focal position, wherein the control means (34; 54) is configured to perform the comparison of local image sharpness information in the partial images and to produce the depth maps therefrom.

4. The device as claimed in any one of the preceding claims, wherein the control means is configured to obtain, knowing the focal position that has been adjusted for detecting the group of partial images and after determining sharply projected objects located in the respectively adjusted focal position, information that the respective sharply projected image areas have been taken in a distance from the multi-aperture imaging device corresponding to the adjusted focal position.

5. The device as claimed in any one of the preceding claims, wherein the control means is configured to obtain, using several groups of partial images and several focal positions, information that the respective sharply projected image areas have been taken in a distance from the multi-aperture imaging device, to obtain for different objects and to produce for the entire field of view in order to achieve the depth map.

6. The device as claimed in any one of the preceding claims, wherein the control means is configured to determine, on the basis of the knowledge of which one of the focal positions has been adjusted to receive the partial image and by analyzing the image information for which parts of the image are sharply projected, in which distance said sharply projected objects are arranged with regard to the device and to use this information with regard to the distance for the depth map.

7. The device as claimed in any one of the preceding claims, wherein the control means is configured to use the sharpness of the same objects in neighboring images of the stack, the allocation of the focus-actuator position to the object distance via a calibrated look-up table and / or the orientation of the through focus scan in order to gain the depth information partially recursively from neighboring images of the stack and to avoid ambiguities.

8. The device as claimed in any one of the preceding claims, wherein the control means is configured to produce the depth map (38) from the sequence of image information (36).

9. The device as claimed in any one of the preceding claims, wherein the optical channels are configured to detect the total field of view in an at least stereoscopic manner; wherein the control means is configured to produce a preliminary depth map on the basis of disparity information obtained from the optical channels; and to supplement the preliminary depth map on the basis of depth information based on the sequence of image information in order to obtain the depth map (38); or wherein the control means is configured to produce a preliminary depth map on the basis of the sequence of image information; and to supplement the preliminary depth map on the basis of depth information based on disparity information obtained from the optical channels in order to obtain the depth map (38).

10. The device as claimed in any of the preceding claims, configured to control the focusing means (32) so that the sequence of focal positions (56) is equidistantly distributed in a tolerance range of 25% in the image space between a minimum focal position and a maximum focal position.

11. The device as claimed in any of the preceding claims, configured to generate a sequence of total images (42) representing the total field of view (261, 262) on the basis of the sequence of image information (36), wherein each total image (42) is based on a combination of partial images (64) of the same focal position (56).

12. The device as claimed in any of the preceding claims, configured to change a total image (42), which represents the total field of view (261, 262), on the basis of the depth map (38) by subsequent focusing and / or defocusing of one or more image areas.

13. The device as claimed in any of the preceding claims, configured to produce an image of the total field of view (261, 262) as a mono image and to produce the depth map (38) from the sequence of mono images.

14. The device as claimed in any of the preceding claims, wherein a first optical channel (16a) of the array (14) is configured to map a first partial field of view (24a) of the total field of view (261; 262), wherein a second optical channel (16b) of the array (14) is configured to image a second partial field of view (24b) of the total field of view (261; 262), and wherein a third optical channel (24c) is configured to completely image the total field of view (261; 262).

15. The device as claimed in any of the preceding claims, wherein the focusing means (32) comprises at least one actuator (74) for adjusting the focal position (56), wherein the focusing means (32) is arranged so that the same is arranged at least partially between two planes (68a, 68b) spanned by sides (69a, 69b) of a cuboid (69), wherein the sides (69a, 69b) of the cuboid (69) are aligned in parallel to each other and to a line extension direction (2) of the array (14) and part of the beam path (104) of the optical channels (16a-d) between the image sensor (12) and the beam deflector (18) and the volume of the same is minimal, but still includes the image sensor (12), the array (14) and the beam deflector (18).