Device and method for online georeferencing of camera captures of map material with mobile terminals

Abstract

The invention relates to a device for georeferencing according to one embodiment. The device comprises a map image provision unit (110) for providing a map image captured by a camera. The device also comprises a match determining unit (120) for determining one or more matches between the captured map image and a reference map, in which, in order to determine a match among the one or more matches, a point on the captured map image is assigned a point on the reference map, and / or, in order to determine the match, a region of the captured map image is assigned a region of the reference map.

Description

[0001] Device and method for online georeferencing of camera images of map material using mobile devices

[0002] Description

[0003] The application relates to a device and a method for online georeferencing of camera images of map material using mobile devices.

[0004] Often, maps on display boards or printed maps provide additional information, such as hiking trails, specific locations and information, which provides added value for the user. It would be desirable if such maps could be used on a mobile device.

[0005] In general, only pre-georeferenced maps can be used on mobile devices. Offline georeferencing of analogue maps or other map material is possible using several methods, which usually involve determining control points and deriving a transformation from them.

[0006] There are many ways to register 2D map images. These include feature-based methods, such as SIFT (scale invariant feature transform), SURF (speeded-up robust features), and intensity-based methods. However, these methods have the disadvantage of a lack of robustness (feature-based methods), a small convergence radius, and the presence of local minima (intensity-based methods such as mutual information).

[0007] Another possibility is to use path information as a source of correspondence. This could be done using a graph-based approach to directly map the paths to each other. However, this requires robust extraction of the path from the photographed map, which in turn represents a very difficult image processing problem.

[0008] In the context of georeferencing maps, various prior art is known. WO 2001 / 073690 A2 and US 7167187 B2 disclose a system and method for georeferencing digital raster maps by mapping specific points on the raster map to corresponding points on a georeferenced vector map or another georeferenced raster map. For example, a user is simultaneously presented with a raster map and a corresponding vector map. The user then locates a common geographic point or feature on each map and marks each of them as a unique point pair. Once the user has marked at least two point pairs, the system calculates a georeferencing function based on the pixel coordinates of the points marked on the raster map and the corresponding geographic coordinates of the points on the vector map.The geographic coordinates of each point on the map can then be easily calculated. One example proposes that, as more point pairs are identified, the georeferencing function is modified for increased accuracy.

[0009] US 7038681 B2 and US 2005 / 0073532 A1 show that, after paper flood maps have been scanned into a computer system, they are each displayed side by side with the relevant part of a stored vector map covering the geographical area covered by all maps. During display, corresponding points or pairs of points on each map are marked using a keyboard or mouse. The corresponding points can be, for example, roads, riverbeds, railway tracks, intersections, buildings, mountain peaks, etc., which are identifiable on each displayed map. When two pairs of points are marked, a stored algorithm calculates georeferencing functions that enable the x and y coordinates of the raster map to be translated into the latitude and longitude coordinates of the vector map, and vice versa. As more pairs of points are marked, the georeferencing functions are refined.Another algorithm removes seemingly bad points from the georeferencing calculations; the points can be reused later if they appear to be not bad.

[0010] US 2006 / 0041375 A1 discloses a system and method for automatically georeferencing digitized images that select and analyze image landmarks from different sources. A composite alignment of raster images and geographic target vectors is automatically generated using metadata sources for use in geographic display and editing environments. Cartographic alignment of raster images to geographic vector datasets is performed to co-position image data with geographic vector datasets. Meaningful overlay of the datasets and improved comparison of vector data with the images based on the physical relationships between the vector data and the underlying image features is enabled.By integrating landmark metadata, users can query target vectors to view attributes of associated landmarks and create a highly accurate parcel map to validate imagery in areas where data was collected under suboptimal conditions.

[0011] It would be desirable if improved concepts were provided that would enable the use of printed maps on mobile devices.

[0012] An apparatus according to claim 1, a method according to claim 24 and a computer program according to claim 25 are provided.

[0013] A device for georeferencing according to one embodiment is provided. The device comprises a map image provision unit for providing a map image recorded by a camera. Furthermore, the device comprises a correspondence determination unit for determining one or more correspondences between the recorded map image and a reference map by assigning a point of the recorded map image to a point of the reference map to determine a correspondence of the one or more correspondences, and / or by assigning an area of ​​the reference map to an area of ​​the recorded map image to determine the correspondence.

[0014] Furthermore, a method for georeferencing is provided according to one embodiment. The method comprises:

[0015] Providing a map image captured by a camera.

[0016] Determining one or more correspondences between the recorded map image and a reference map by assigning a point of the recorded map image to a point of the reference map to determine a correspondence of the one or more correspondences, and / or by assigning an area of ​​the reference map to an area of ​​the recorded map image to determine the correspondence. Furthermore, a computer program with program code for carrying out the method described above is provided according to one embodiment.

[0017] According to one embodiment, maps or a photo of the map are used with a mobile device by georeferencing the photo. For example, in one embodiment, the location of the mobile device can be displayed live on the photo of the map.

[0018] According to one embodiment, for example, to achieve this georeferencing, georeferenced points are determined on the photo of the map image. This determination can be performed, for example, using a GPS receiver in the terminal device and manual interaction with the captured map, or using an automatic process for image registration of the photographed map image with an already georeferenced map. In one embodiment, the georeferencing of another map is transferred to the photo.

[0019] Some embodiments solve a difficult image registration problem, since the photo of the map image typically contains completely different content and map styles than the reference maps. In one embodiment, modern learning-based methods are used that enable robust correspondence detection, thus enabling this application to be implemented on mobile devices.

[0020] In particular, embodiments include online use of photographed maps with mobile devices.

[0021] Preferred embodiments of the invention are described below with reference to the drawings.

[0022] The drawings show:

[0023] Fig. 1 shows a device for georeferencing according to a

[0024] Embodiment.

[0025] Fig. 2 shows an example of a recorded hiking map with marked trail information (left) and a georeferenced map (right) of the relevant area, which is used as a reference map. Fig. 3 shows an example of the correspondences found between a camera image and a reference map.

[0026] Fig. 4 shows an example of the automatic registration of the map image with a reference map.

[0027] Fig. 5 shows examples of automatic georeferencing (left) of a

[0028] Camera image of a map and the map section (right).

[0029] Fig. 6 shows another example of correspondences found for a

[0030] Camera image with a reference card.

[0031] Fig. 7 shows another example of found correspondences of a camera image with a reference map.

[0032] Fig. 8 shows another example of automatic georeferencing of a photographed map image with a reference map.

[0033] Fig. 1 shows a device for georeferencing according to an embodiment.

[0034] The device comprises a map image providing unit 110 for providing a map image captured by a camera.

[0035] Furthermore, the device comprises a correspondence determination unit 120 for determining one or more correspondences between the recorded map image and a reference map by assigning a point of the reference map to a point of the recorded map image in order to determine a correspondence of the one or more correspondences, and / or by assigning an area of ​​the reference map to an area of ​​the recorded map image in order to determine the correspondence.

[0036] According to one embodiment, the correspondence determination unit 120 can be designed to determine a position of the device in the recorded map depending on a localization of a position of the device in the reference map. And / or, wherein the correspondence determination unit 120 is designed to determine a position of the other device in the recorded map depending on a localization of a position of another device in the reference map. In one embodiment, the device can, for example, have a localization unit which can, for example, be designed to determine a position of the device in the reference map. In this case, the correspondence determination unit 120 can, for example, be designed to determine the position of the device in the recorded map depending on the localization of the position of the device in the reference map.

[0037] According to one embodiment, the localization unit can be designed, for example, to determine the position of the device in the reference map by means of GPS.

[0038] In one embodiment, the localization unit can be designed, for example, to determine the position of the device in the reference map by means of text recognition of one or more place names and / or one or more street names on the recorded map.

[0039] According to one embodiment, the localization unit can be designed, for example, to determine map material containing the one or more place names and / or the one or more street names recognized by means of text recognition, depending on the one or more place names and / or one or more street names.

[0040] In one embodiment, the device may be, for example, a mobile device.

[0041] According to one embodiment, the mobile device may, for example, have an Internet connection configured to receive the reference card.

[0042] In one embodiment, the mobile device may be, for example, a mobile radio device.

[0043] According to one embodiment, the map image providing unit 110 may, for example, comprise the camera, which may, for example, be configured to photograph or otherwise record a map in order to obtain the recorded map image.

[0044] In one embodiment, the correspondence determination unit 120 can be configured, for example, to use a learning-based method trained on map images to determine the one or more correspondences between the recorded map image and the reference map. According to one embodiment, the correspondence determination unit 120 can be configured, for example, to subdivide the recorded map image into a plurality of regions of the recorded map image. In this case, the correspondence determination unit 120 can be configured, for example, to determine the one or more correspondences depending on the subdivision into the plurality of regions.

[0045] In one embodiment, the correspondence determination unit 120 may, for example, be configured to determine, depending on the one or more correspondences, a transformation that maps the reference map or one or more regions of the reference map to the recorded map image or to one or more regions of the recorded map image.

[0046] According to one embodiment, the correspondence determination unit 120 can, for example, be designed to determine as the one or more correspondences a plurality of correspondences between the recorded map image and one of two or more reference maps that comprise said reference map by assigning a point of the recorded map image to a point of one of the two or more reference maps to determine a correspondence of the plurality of correspondences, and / or by assigning an area of ​​one of the two or more reference maps to determine the correspondence, wherein the plurality of correspondences comprise at least two correspondences that relate to at least two different ones of the two or more reference maps.

[0047] In one embodiment, the correspondence determination unit 120 may, for example, be configured to determine a plurality of correspondences.

[0048] According to one embodiment, the correspondence determination unit 120 can, for example, be designed to determine the plurality of correspondences in such a way that a first correspondence of the plurality of correspondences is determined by assigning a point of the reference map to a point of the recorded map image or by assigning an area of ​​the reference map to an area of ​​the recorded map image, and that a further correspondence of the plurality of correspondences is determined by assigning a point of a further reference map to a further point of the recorded map image or by assigning an area of ​​the further reference map to a further area of ​​the recorded map image. In one embodiment, the correspondence determination unit 120 can, for example, be designed to determine a plurality of correspondences whose number corresponds to at least one threshold value.In this case, the correspondence determination unit 120 can be designed, for example, if the reference map is not sufficient to determine the number of correspondences specified by the threshold value, to use another reference map to determine one or more additional correspondences by determining an additional correspondence of the one or more additional correspondences, by assigning a point of the other reference map to a point of the recorded map image, or by assigning an area of ​​the other reference map to an area of ​​the recorded map image.

[0049] According to one embodiment, the correspondence determination unit 120 may, for example, be configured to determine an order in which a plurality of reference cards are used to determine the plurality of correspondences depending on similarities that depend on the plurality of reference cards and the acquired card.

[0050] In one embodiment, the correspondence determination unit 120 may, for example, be configured to determine the order in which a plurality of reference maps are used to determine the plurality of correspondences depending on a similarity function of map styles, which was found on the basis of already registered maps and a number of found, correct correspondences by means of a metric learning approach.

[0051] According to one embodiment, the plurality of reference cards may comprise, for example, at least one of the following types of cards:

[0052] Maps with elevation profiles,

[0053] - already georeferenced hiking or road maps,

[0054] Maps obtained based on a conditional GAN ​​network from already registered reference and query maps.

[0055] In one embodiment, the device may, for example, comprise a user interface, wherein, when the device is in a manual mode, the user interface may, for example, enable user input by which a user may indicate a location of the device in the captured map image.

[0056] According to one embodiment, the device can be designed, for example, to determine its current location by means of GPS and to display it to a user.

[0057] In one embodiment, when the device is in a manual mode, the user interface may, for example, enable user input at various positions where the device is located, allowing a user to specify a location of the device at the various positions in the captured map image.

[0058] Specific embodiments of the invention are described below.

[0059] Embodiments are based on a robust and sufficient determination of 2D / 2D correspondences in order to achieve a transformation of a pixel coordinate system into a geocoordinate system.

[0060] According to embodiments, an algorithm for automatically georeferencing a photo of a map image (e.g., a hiking map on a board) may comprise a multi-step process, which may be implemented, for example, using two separate procedures. The procedures are mutually optional, using an automatic procedure and a manual fallback mode:

[0061] In some embodiments, only one of the two processes is implemented, either the automatic process or the manual process.

[0062] Other embodiments provide both methods, i.e., both the automatic method and the manual method. In particular embodiments, the processes are mutually optional: The automatic method is provided, along with a manual fallback mode.

[0063] The procedure for automatic georeferencing is described below:

[0064] An (automatic) process without manual interaction can be implemented, for example, as follows: First, an initial rough localization is established. This can be done, for example, using a GPS position of the mobile device; and / or text recognition of place names on the displayed map is performed and a search for corresponding map material for automatic image registration is performed. For example, the found map material can be used as the reference map.

[0065] Based on the initial coarse localization, maps of the area are used or generated as reference maps for image registration (see Fig. 2 to Fig. 5). The maps can be used, for example, using a cascading approach. If a sufficient number of correspondences are not found with one map (the determination can be made, for example, using a statistically determined threshold), an alternative map is used for registration. The order of the maps can, for example, result from a similarity function of map styles, which was determined using a metric learning approach based on previously registered maps and the number of correct correspondences found.The possible reference maps that can be selected here can include, for example, maps with elevation profiles, as well as already georeferenced hiking and road maps and / or maps that were obtained on the basis of a conditional GAN ​​(Generative Adversarial Networks) from already registered reference and query maps.

[0066] Then, correspondences (e.g., 2D / 2D) between the geo-registered reference maps and the captured map image are determined. The correspondences can be determined, for example, using a learning-based method trained on corresponding map images. The images can be divided into cells, for example, to limit the storage requirements on mobile devices (see Fig. 3, Fig. 6, and Fig. 7).

[0067] Finally, a 2D transformation is determined based on the 2D / 2D correspondences found in 3.2c. This transformation can, for example, have four degrees of freedom, e.g., translation (x, y), rotation, and scaling, in the case of perspective rectification. Homography in the absence of perspective rectification can, for example, be realized. A parametric transformation field can, for example, be realized if geometric distortions are to be compensated. Using a robust inlier method (e.g., RANSAC), the geometrically consistent correspondences can be selected and accumulated (see Fig. 4), which enables both robust registration of strongly deviating map styles and high-dimensional transformation determinations (e.g., transformation fields) (see Fig. 4 and Fig. 8).

[0068] Now the manual procedure for georeferencing is described:

[0069] A process with manual interaction (e.g. as a fallback mode - if automatic georeferencing via image registration cannot be performed) can be implemented as follows:

[0070] The captured image of the map is perspectively corrected (parallel lines run parallel in the image; e.g. achievable by detecting the edges of the image

[0071] Position 1 is determined with a built-in GPS receiver and

[0072] In the captured image of the map, the user annotates the current location of position 1 .

[0073] This localization and annotation of one or more positions takes place at at least one or optionally one or more additional locations (i.e., the position is determined using the built-in GPS receiver and the current location is annotated by the user). This can be used to calculate a compensating transformation (e.g., using least-squares optimization), which can then be used for georegistration.

[0074] Embodiments provide robust 2D / 2D correspondence finding of widely differing map styles for determining georeferencing across reference maps.

[0075] For this purpose, for example, a learning-based method can be trained on a large number of corresponding map images. The training data is available in large quantities thanks to the large number of map providers and can be used on a global scale without great effort to train a highly robust matcher for this application. Since the maps provided are usually georeferenced, one embodiment can use supervised learning, for example. For example, so-called dense correspondence networks (e.g., based on a vision transformer in combination with a cost volume) are used as architectures, which exhibit a much higher correspondence performance than feature-based methods (including learning-based methods, e.g., sparse feature matching).

[0076] In some embodiments, the method can operate on a dense cost volume, which is recursively traversed and analyzed for geometric consistency. Geometric consistency can be analyzed, for example, in an aggregation phase and recursively accumulated, enriched, and evaluated with consistency values.

[0077] Exemplary application scenarios of embodiments are described below with reference to Figs. 2 to 8.

[0078] Fig. 2 shows an example of a recorded hiking map with marked trail information (left) and a georeferenced map (right) of the relevant area, which is used as a reference map. As expected, the reference map exhibits significant differences in appearance, which must be compensated for using the robust registration procedure.

[0079] Fig. 3 shows an example of correspondences found between a camera image and a reference map. Fig. 3 shows all the correspondences found without geometric filtering. The robustness of the learned method is clearly superior to conventional methods. In this example, no geometric filtering of the correspondences was performed.

[0080] Fig. 4 shows an example of automatic registration of the map image with a reference map. The figure shows overlays of the acquired map with an already georeferenced map of the same area (left with edge overlay, right with alternating checkerboard overlay).

[0081] Fig. 5 shows examples of automatic georeferencing (left) of a camera image of a map and the map section (right). To enable the photographed map (left) and the map section (right) to be used with a navigation application on a mobile device, this image is georeferenced.

[0082] Fig. 6 shows an example of correspondences found between a camera image and a reference map. Fig. 6 shows all correspondences found without geometric filtering. Fig. 7 shows an example of correspondences found between a camera image and a reference map. Fig. 7 shows all correspondences found after geometric filtering.

[0083] Fig. 8 shows an example of automatic georeferencing of a photographed map image with a reference map. The figure shows an overlay of the captured map with an already georeferenced map of the same area.

[0084] In some embodiments, robust correspondence detection allows very different card styles to be registered with each other. This enables automatic, robust functionality of the application on mobile devices.

[0085] According to embodiments, with a sufficient number of correspondences, imaging errors, e.g. strongly differing map contents (additional symbols and deviating temporal contents in the maps) and geometric deformations (e.g. waves in books, etc.) can be compensated.

[0086] Some embodiments implement an automatic or manual procedure based on control points.

[0087] Embodiments can be used, for example, in the context of mobile devices. Embodiments provide a mobile solution for using maps or photos of maps directly with a mobile device.

[0088] Further embodiments can be used in the digitization and georeferencing of analog map images.

[0089] Some embodiments can be implemented as an app. In special embodiments, such an app can be integrated into another app, for example, in the area of ​​navigation for pedestrians / hiking.

[0090] In some specific embodiments, images taken can, for example, be stored and used for one or more various applications and services.

[0091] Although some aspects have been described in the context of a device, it should be understood that these aspects also represent a description of the corresponding method, so that a block or component of a device can also be understood as a corresponding method step or as a feature of a method step. Analogously, aspects described in the context of or as a method step also represent a description of a corresponding block, detail, or feature of a corresponding device. Some or all of the method steps may be performed by (or using) a hardware apparatus, such as a microprocessor, a programmable computer, or an electronic circuit. In some embodiments, some or more of the key method steps may be performed by such an apparatus.

[0092] Depending on specific implementation requirements, embodiments of the invention may be implemented in hardware or in software, or at least partially in hardware or at least partially in software. The implementation may be carried out using a digital storage medium, for example a floppy disk, a DVD, a Blu-ray disc, a CD, a ROM, a PROM, an EPROM, an EEPROM, or a FLASH memory, a hard disk, or other magnetic or optical storage device on which electronically readable control signals are stored that can interact or interact with a programmable computer system such that the respective method is carried out. Therefore, the digital storage medium may be computer-readable.

[0093] Some embodiments according to the invention thus comprise a data carrier having electronically readable control signals capable of interacting with a programmable computer system such that one of the methods described herein is carried out.

[0094] In general, embodiments of the present invention may be implemented as a computer program product having a program code, wherein the program code is effective to perform one of the methods when the computer program product is run on a computer.

[0095] The program code can, for example, also be stored on a machine-readable medium.

[0096] Other embodiments include the computer program for performing one of the methods described herein, wherein the computer program is stored on a machine-readable medium. In other words, one embodiment of the method according to the invention is thus a computer program that has program code for performing one of the methods described herein when the computer program is executed on a computer.

[0097] A further embodiment of the method according to the invention is thus a data carrier (or a digital storage medium or a computer-readable medium) on which the computer program for performing one of the methods described herein is recorded. The data carrier or the digital storage medium or the computer-readable medium is typically tangible and / or non-transitory.

[0098] A further embodiment of the method according to the invention is thus a data stream or a sequence of signals that represents the computer program for carrying out one of the methods described herein. The data stream or the sequence of signals can be configured, for example, to be transferred via a data communication connection, for example, via the Internet.

[0099] A further embodiment comprises a processing device, for example a computer or a programmable logic device, which is configured or adapted to carry out one of the methods described herein.

[0100] A further embodiment comprises a computer on which the computer program for performing one of the methods described herein is installed.

[0101] A further embodiment according to the invention comprises a device or system designed to transmit a computer program for performing at least one of the methods described herein to a recipient. The transmission can be electronic or optical, for example. The recipient can be, for example, a computer, a mobile device, a storage device, or a similar device. The device or system can, for example, comprise a file server for transmitting the computer program to the recipient.

[0102] In some embodiments, a programmable logic device (e.g., a field-programmable gate array, an FPGA) may be used to perform some or all of the functionalities of the methods described herein. In some embodiments, a field-programmable gate array may cooperate with a microprocessor to perform any of the methods described herein. In general, in some embodiments, the methods are performed by any hardware device. This may be general-purpose hardware such as a computer processor (CPU) or method-specific hardware such as an ASIC.

[0103] The above-described embodiments are merely illustrative of the principles of the present invention. It is understood that modifications and variations of the arrangements and details described herein will be apparent to others skilled in the art. Therefore, it is intended that the invention be limited only by the scope of the following claims and not by the specific

[0104] Details presented in the description and explanation of the embodiments herein are limited.

Claims

Patent claims 1 . A device for georeferencing, comprising: a map image providing unit (110) for providing a map image recorded by a camera, a correspondence determining unit (120) for determining one or more correspondences between the recorded map image and a reference map by assigning a point of the reference map to a point of the recorded map image to determine a correspondence of the one or more correspondences, and / or by assigning an area of ​​the reference map to an area of ​​the recorded map image to determine the correspondence.

2. Device according to claim 1, wherein the correspondence determination unit (120) is designed to determine a position of the device in the recorded map depending on a localization of a position of the device in the reference map, and / or wherein the correspondence determination unit (120) is designed to determine a position of the other device in the recorded map depending on a localization of a position of another device in the reference map.

3. Device according to claim 2, wherein the device comprises a localization unit configured to determine a position of the device in the reference map, wherein the correspondence determination unit (120) is configured to determine the position of the device in the recorded map depending on the localization of the position of the device in the reference map.

4. Device according to claim 3, wherein the localization unit is configured to determine the position of the device in the reference map by means of GPS.

5. Device according to claim 3 or 4, wherein the localization unit is designed to determine the position of the device in the reference map by means of text recognition of one or more place names and / or one or more street names on the recorded map.

6. Device according to claim 5, wherein the localization unit is designed to determine map material containing the one or more place names and / or the one or more street names depending on the one or more place names and / or one or more street names recognized by means of text recognition.

7. Device according to one of the preceding claims, wherein the device is a mobile device.

8. The apparatus of claim 7, wherein the mobile device has an internet connection configured to receive the reference card.

9. The device according to claim 7 or 8, wherein the mobile device is a mobile radio device.

10. The device according to any one of the preceding claims, wherein the map image providing unit (110) comprises the camera configured to photograph or otherwise record a map to obtain the recorded map image. 11 . Device according to one of the preceding claims, wherein the correspondence determination unit (120) for determining the one or more correspondences between the recorded map image and the reference map is trained to use a learning-based method that was trained on map images.

12. Device according to one of the preceding claims, wherein the correspondence determination unit (120) is designed to subdivide the recorded map image into a plurality of regions of the recorded map image, and wherein the correspondence determination unit (120) is designed to perform the one or more correspondences depending on the subdivision into the plurality of regions.

13. Device according to one of the preceding claims, wherein the correspondence determination unit (120) is designed to determine, depending on the one or more correspondences, a transformation which maps the reference map or one or more regions of the reference map onto the recorded map image or onto one or more regions of the recorded map image.

14. Device according to one of the preceding claims, wherein the correspondence determination unit (120) is designed to determine, for determining as the one or more correspondences, a plurality of correspondences between the recorded map image and one of two or more reference maps comprising said reference map by assigning a point of one of the two or more reference maps to a point of the recorded map image to determine a correspondence of the plurality of correspondences, and / or by assigning an area of ​​one of the two or more reference maps to determine the correspondence, wherein the plurality of correspondences comprise at least two correspondences that relate to at least two different ones of the two or more reference maps.

15. Device according to one of the preceding claims, wherein the correspondence determination unit (120) is designed to determine a plurality of correspondences.

16. The device according to claim 15, wherein the correspondence determination unit (120) is configured to determine the plurality of correspondences in such a way that a first correspondence of the plurality of correspondences is determined by assigning a point of the reference map to a point of the recorded map image or by assigning an area of ​​the reference map to an area of ​​the recorded map image, and that a further correspondence of the plurality of correspondences is determined by assigning a point of a further reference map to a further point of the recorded map image or by assigning an area of ​​the further reference map to a further area of ​​the recorded map image.

17. The device according to claim 15 or 16, wherein the correspondence determination unit (120) is designed to determine a plurality of correspondences whose number corresponds to at least one threshold value, wherein the correspondence determination unit (120) is designed, if the reference map is not sufficient to determine the number of correspondences predetermined by the threshold value, to use another reference map to determine one or more additional correspondences by determining an additional correspondence of the one or more additional correspondences, by assigning a point of the other reference map to a point of the recorded map image, or by assigning an area of ​​the other reference map to an area of ​​the recorded map image.

18. The apparatus according to claim 17, wherein the correspondence determination unit (120) is designed to determine an order in which a plurality of reference cards are used to determine the majority of correspondences are used to determine similarities depending on the majority of reference maps and the captured map.

19. The apparatus according to claim 17 or 18, wherein the correspondence determination unit (120) is configured to determine the order in which a plurality of reference maps are used to determine the plurality of correspondences depending on a similarity function of map styles, which was found on the basis of already registered maps and a number of found, correct correspondences by means of a metric learning approach.

20. The apparatus of claim 18 or 19, wherein the plurality of reference cards comprises at least one of the following types of cards: Maps with elevation profiles, - already georeferenced hiking or road maps, Maps obtained based on a conditional GAN ​​network from already registered reference and query maps.

21. Apparatus according to any preceding claim, wherein the apparatus comprises a user interface, wherein, when the apparatus is in a manual mode, the user interface enables user input by which a user can indicate a location of the apparatus in the captured map image.

22. Device according to claim 21, wherein the device is designed to determine its current location by means of GPS and to display it to a user.

23. The device of claim 21 or 22, wherein, when the device is in a manual mode, the user interface enables user input at various positions where the device is located, allowing a user to specify a location of the device at the various positions in the captured map image.

24. A georeferencing method comprising: Providing a map image captured by a camera, Determining one or more correspondences between the recorded map image and a reference map by assigning a point of the reference map to a point of the recorded map image to determine a correspondence of the one or more correspondences, and / or by assigning an area of ​​the reference map to an area of ​​the recorded map image to determine the correspondence.

25. A computer program having a program code for carrying out the method according to claim 24.

Images