METHOD AND DEVICE FOR PROVIDING OCCUPANT INFORMATION FOR A SAFETY DEVICE FOR A VEHICLE

Description

State of the art

[0001] The invention is based on a device or a method according to the class of the independent claims. The present invention also relates to a computer program.

[0002] Vehicle occupant systems such as airbags or seat belts do not, in most cases, apply any adjustment with regard to individual occupant parameters such as age, weight or gender.

[0003] US2012170817A1 discloses a vehicle control system that adapts to individual occupant parameters.

[0004] US 6 198 996 B1 discloses a method for authorizing a user, for example using a camera of a vehicle.

[0005] The paper BRAUER-BURCHARDT C ET AL: "A new algorithm to correct fish-eye and strong wide-angle lens-distortion from single images", PROCEEDINGS 2001 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, Vol. 1, October 7, 2001 (2001-10-07), pages 225-228, XP010564837, D0I: 10.1109 / ICIP.2001.958994, ISBN: 978-0-7803-6725-8 deals with an algorithm for image correction.

[0006] The paper PATEL VISHAL MET AL: "Cancelable Biometrics: A review", IEEE SIGNAL PROCESSING MAGAZINE, Vol. 32, No. 5, September 2015, pages 54-65, XP011666150, ISSN: 1053-5888, DOI: 10.11 09 / MSP.2015.2434151 deals with the use of biometric features. Disclosure of the invention

[0007] Against this background, the approach presented here presents a method for providing occupant information for a vehicle safety device, a device that uses this method, and finally a corresponding computer program according to the main claims. The measures listed in the dependent claims enable advantageous refinements and improvements of the device specified in the independent claim.

[0008] By appropriately checking the plausibility of image data from an image capture device of the vehicle, plausible occupant information can be provided regarding an occupant captured by the image capture device of the vehicle, which information can be used by a safety device or assistance system of the vehicle to adapt a safety function or assistance function to the occupant.

[0009] A method for providing occupant information for a safety device for a vehicle comprises the following steps: Reading in image data representing an occupant of the vehicle via an interface to an image capture device of the vehicle; reading in plausibility data representing a person via an interface to a mobile device; determining occupant data characterizing the occupant using the image data and the plausibility data; and providing the occupant data to an interface to the safety device for the vehicle.

[0010] The image data representing the occupant can represent an image depicting the occupant and captured by the image capture device. The image capture device can represent an image capture device that is permanently installed in the vehicle during operation of the image capture device. For example, the image capture device can be a camera. The mobile device can be a smartphone, for example. The mobile device can be carried by the person. The plausibility data can represent data stored in the mobile device or data generated by the mobile device. For example, the mobile device can be configured to generate the plausibility data using an image of the person captured by a camera of the mobile device or an image of the person stored on the mobile device.The occupant data are determined after a successful plausibility check of the image data by the plausibility data, or vice versa, using the image data and / or the plausibility data.

[0011] The plausibility check will be successful if the person matches the occupant. The occupant data can include information about the occupant's age, weight, or gender, for example. The data storage can be part of a system with data storage that may be present in the vehicle. An example of this would be a car multimedia head unit.

[0012] In the step of reading in the plausibility data, further plausibility data representing another person can be read in via the interface to the mobile device. In the determination step, the occupant data can also be determined using the further plausibility data. In this way, plausibility data of two different people can be provided from one and the same mobile device to check the plausibility of the image data provided by the vehicle's image capture. The different people can be people who share use of the mobile device or who, based on experience, are frequently in the vicinity of the mobile device. This applies, for example, to a smartphone belonging to a parent who is traveling with a child.

[0013] The method comprises a step of requesting further plausibility data representing another person if, in the determining step, the occupant data cannot be determined using the image data and the plausibility data. In this case, the method can comprise a repeated step of determining the occupant data using the further plausibility data. The fact that the occupant data cannot be determined can result, for example, from the fact that a plausibility check of the image data using the plausibility data is unsuccessful because the occupant and the person are different people. In this case, the further plausibility data can be requested from the person to whom the plausibility data is assigned.This is advantageous because, based on the plausibility data read in, it can be assumed with certainty that the person whose plausibility data was read in is in the vehicle and can therefore make a statement about the occupant recorded by the vehicle's image capture device.

[0014] The method may include a step of adapting the safety device using the occupant data. For example, a belt force of a seat belt or an internal pressure of an airbag may be adapted using the occupant data. A display of a display device relating to the use of the safety device, for example, a pictogram depicting a child or an adult, may also be adapted using the occupant data.

[0015] The method may include a step of determining the plausibility data using a recording representing the person. The recording may be stored in the mobile device or may have been recently captured using an image capture device of the mobile device. A recent capture of the recording has the advantage that it can be assumed with a high degree of probability that the recording is associated with the person currently carrying the mobile device.

[0016] In the determining step, the image representing the person can be normalized using a characteristic of the image capture device to determine the plausibility data as normalized image data. This has the advantage that the data received from the mobile device and the vehicle's image capture device can be fed into a joint image processing process.

[0017] In the determination step, parameters of the normalized recording can also be determined as the plausibility data. This eliminates the need to transmit the recording or normalized recording itself to a device implementing the method.

[0018] This method can be implemented, for example, in software or hardware or in a mixed form of software and hardware, for example in a device.

[0019] The approach presented here thus further creates a device that is designed to carry out, control or implement the steps of a variant of a method presented here in corresponding devices. This embodiment of the invention in the form of a device also allows the problem underlying the invention to be solved quickly and efficiently. In the present case, a device can be understood as an electrical device that processes sensor signals and outputs control and / or data signals as a function thereof. The device can have an interface that can be designed in hardware and / or software. In a hardware design, the interfaces can, for example, be part of a so-called system ASIC, which contains a wide variety of functions of the device.However, it is also possible for the interfaces to be separate integrated circuits or at least partially consist of discrete components. In a software-based implementation, the interfaces can be software modules that exist, for example, on a microcontroller alongside other software modules.

[0020] Also advantageous is a computer program product or computer program with program code that can be stored on a machine-readable carrier or storage medium such as a semiconductor memory, a hard disk memory or an optical memory and is used to carry out, implement and / or control the steps of the method according to one of the embodiments described above, in particular when the program product or program is executed on a computer or a device.

[0021] Embodiments of the invention are illustrated in the drawings and explained in more detail in the following description. It shows: Fig. 1 a schematic representation of a system with a device for providing occupant information according to an embodiment; Fig. 2 a flowchart of a method for providing occupant information according to an embodiment; Fig. 3 a representation of a system with a device according to an embodiment; Fig. 4 a mobile device according to an embodiment; Fig. 5 a presentation of strategies according to implementation examples Fig. 6 a flowchart of a method according to an embodiment; Fig. 7 a flowchart of a method according to an embodiment; Fig. 8 a representation of reflectance properties of human skin types according to an embodiment; Fig. 9 a representation of geometric distortions of different camera models according to an embodiment; and Fig. 10 a flowchart of a method according to an embodiment.

[0022] In the following description of advantageous embodiments of the present invention, the same or similar reference numerals are used for the elements shown in the various figures and having a similar effect, whereby a repeated description of these elements is omitted.

[0023] Fig. 1 shows a schematic representation of a system with a device 100 for providing occupant information according to one exemplary embodiment. According to this exemplary embodiment, the device 100 is integrated into a vehicle 102. The vehicle 102 further comprises an image capture device 104, for example a camera, and a safety device 106, for example a seatbelt or an airbag, for protecting an occupant 108 of the vehicle 102. A mobile device 110, for example a smartphone of the occupant 108, is also arranged in the vehicle 102. The mobile device 110 can have entered the vehicle 102 together with the occupant 108 and can be removed from the vehicle 102 again by the occupant 108 upon leaving the vehicle 102. According to one exemplary embodiment, the device 110 is a data storage device of a system present in the vehicle 102, such as a multimedia unit.

[0024] The image capture device 104 is configured to capture an image of the occupant 108 and to provide image data 105 representing the image to the device 100 via an interface. To read in the image data 105, the device 100 has a first read-in device 112.

[0025] The mobile device 110 is configured to provide plausibility data 111 representing the occupant 108 to the device 100 via an interface. For reading the plausibility data 111, the device 100 has a second reading device 114.

[0026] The device 100 further comprises a determination device 116 configured to determine occupant data 117 characterizing the occupant 108 using the image data 105 and plausibility data 111 read in by the reading devices 112, 114. According to one embodiment, the determination device 116 is configured to determine the occupant data 117 from the image data 105 and / or the plausibility data 111 if the image data 105 can be verified by the plausibility data 111. In the illustrated embodiment, such a plausibility check is possible because the plausibility data 111 and the image data 105 are associated with the same person, namely the occupant 108.The device 100 has a provision device 118 configured to provide the occupant data 117 determined by the determination device 116, which includes, for example, information about a weight, age, height, or gender of the occupant 108, to an interface to the safety device 106. The safety device 106 is configured to adapt a functionality of the safety device 106 to the occupant 108 using the occupant data 117.

[0027] If the plausibility data 111 provided by the mobile device 110 is assigned to another person, for example, another occupant 120 of the vehicle 102, the image data 105 read by the camera 104 cannot be verified by the plausibility data 111. According to one embodiment, in such a case, no occupant data 117 is provided to the safety device 106, or, for example, standard occupant data is provided.

[0028] According to one embodiment, the mobile device 110 is used by the additional occupant 120 and therefore has plausibility data 111 assigned to the occupant 120. In such a case, in which image data 105 and plausibility data 111 are assigned to different persons 108, 120, the device 100 is configured to request additional plausibility data 122 from the person 120 to whom the plausibility data 111 is assigned, which data is suitable for verifying the plausibility of the person 108 to whom the image data 105 of the image capture device 104 is assigned. In this case, the second reading device 114 can be configured to read in the additional plausibility data 122 and provide it to the determination device 116 for determining the occupant data 117.

[0029] According to one embodiment, the mobile device 110 is configured to provide two or more plausibility data 111, 122 assigned to different individuals. For example, the mobile device 110 is configured to provide plausibility data 111 assigned to the occupant 108 and further plausibility data 122 assigned to the other occupant 120. The second reading device 114 is configured to read the plausibility data 111, 122 assigned to the occupants 108, 120 and provide it to the determination device 116. The determination device 116 is designed to determine the occupant data 117 associated with the occupant 108 detected by the image capture device 104 using the image data 105 of the image capture device 104 as well as the plausibility data 111 and the further plausibility data 122.This is possible if the plausibility data 111 or the further plausibility data 122 are assigned to the occupant 108 and can thus be used to verify the plausibility of the image data 105.

[0030] The described approach enables additional functionality of a so-called Mob2Car (mobile device 110 - to - vehicle 102) system in a so-called tandem mode.

[0031] According to one embodiment, a method for logging in more than one person 108, 120 with a single mobile device 110 is implemented, such as a mother and child or a caregiver and a cared person. This approach allows Mob2Car to meet some requirements of the FMVSS 208, as will be described below with reference to Fig. 5 is shown. According to one exemplary embodiment, the approach represents an important building block of the Mob2Car approach.

[0032] Mob2Car enables passive safety to be increased through the use of the mobile communication device 110. Mob2Car pursues the following objectives and working approaches.

[0033] Occupant protection systems 106, such as airbags and seatbelts, generally do not adapt to individual occupant parameters (age, weight, gender). To better protect each individual occupant, adjustments to restriction parameters are required. Individual safety adaptation requires robust sensors and / or methods for detecting vehicle occupants 108, 122. The corresponding obvious classification task requires highly developed interior sensors 104 and entails considerable effort. The new mob2car approach uses natural data 111 of a user 108 (smartphone 110, app-based) and transmits it to the vehicle 102. There, based on an additionally transmitted image of the user 108, a plausibility check is performed, for example, to verify whether the person 108 sitting in the vehicle 102 matches the transmitted data 111.

[0034] There are two important design constraints. First, reliability is required. In any case, protection should be guaranteed at least according to current technology. Second, data protection is required. The system is designed such that there is no need to store personalized user data in the vehicle 102, for example, a rental vehicle or a fleet vehicle.

[0035] Fig. 2 shows a flowchart of a method for providing occupant information according to an embodiment. The method can be used, for example, with the aid of Fig. 1 described device.

[0036] In step 201, image data of a vehicle occupant captured by an image capture device of the vehicle is received, and in step 203, plausibility data representing a person is received via an interface to a mobile device. Steps 201, 203 can be performed in parallel or sequentially. For example, steps 201, 203 can be performed when the vehicle is started or when a person sits in the vehicle.

[0037] In a step 205, the image data and the plausibility data are used for mutual or unilateral plausibility checks, for example, by comparing the data. This allows it to be determined whether the person to whom the plausibility data is assigned is the occupant. If so, the occupant data is determined based on the image data and, additionally or alternatively, based on the plausibility data. In a step 207, the occupant data is provided to an interface to the vehicle's safety device.

[0038] Fig. 3 shows a representation of a system with a device 100 according to an embodiment. The system shows a general structure of Mob2Car. The device 100 is designed to receive image data of a first person 108 captured by an image capture device 104 via a wired interface and to receive plausibility data relating to a second person 308 from a mobile device 110 via a wireless interface. The device 100 is designed to determine, using the image data and the plausibility data, whether the persons 108, 308 are the same person, represented by the icon 310, or different persons, represented by the icon 312.

[0039] Fig. 4 shows a mobile device 110 according to an embodiment. This may be a larger view of the Fig. 3 shown mobile device, which according to this embodiment is designed as a smartphone.

[0040] Fig. 5 shows a representation of strategies according to implementation examples.

[0041] Block 501 represents that no FMVSS 208 test requirements are met. Block 502 represents the case where a seat is empty. Block 503 represents suppression, and block 504 represents presence.

[0042] Block 511 represents the FMVSS 208 test requirements for minimizing the risk of airbag-induced injury to infants, children, and other occupants. Block 512 represents the case where a baby seat is occupied by a 1-year-old dummy. Block 513 represents suppression and block 514 represents presence. Block 515 represents low-risk deployment. Block 522 represents the case of 3- and 6-year-old child dummies. Block 523 represents suppression and block 524 represents presence. Block 525 represents low-risk deployment. Block 526 represents suppression and block 527 represents an "OOP" state. Block 532 represents the case of a "5th percentile adult female dummy" in the driver position. Block 535 represents low-risk deployment. Block 536 represents suppression and block 537 represents an "OOP" state.

[0043] The described approach uses video-based interior sensing for occupant classification and misposition detection.

[0044] Tandem mode allows more than one person to log in using a single mobile device. This could be, for example, a parent and child, or a caregiver and a person being cared for.

[0045] This approach allows Mob2Car to meet certain FMVSS 208 requirements. This is illustrated using the "Parent / Child" application as an example.

[0046] A parent with a young child enters their own data (weight, age, gender) and a photo or biometric data, as well as their child's data and, optionally, a photo or biometric data (weight, age, gender, child seat type). This is an initial action. The data remains on the mobile device.

[0047] When the parent enters the vehicle, the mobile device transmits both sets of data. The plausibility check and seat assignment can be performed by the in-vehicle system, following a Mob2Car concept, by comparing the biometric data from the mobile device with available data. If the assignment and plausibility check are successful, an adaptive airbag system can be configured.

[0048] As a first step, this could be an adaptation of the hybrid suppression / deployment of a low-risk strategy, as illustrated by blocks 513, 514 and 525, 535.

[0049] Thus, Fig. 5 Different strategies permitted by FMVSS 208. Thus, in reality and taking into account real-life crash situations, occupant protection systems can be based on different protection options, such as a "static suppression option" for the 1 YOC class and LRD (low risk deployment) for the 3 YOC and 6 YOC classes on the passenger side and for the 5PFD class on the driver side. In this case, suitable occupant classification sensors are required to detect the 1 YOC class.

[0050] In the event that biometric data for the child's plausibility check is not available, the parent can be addressed via a human-machine interface (HMI) to confirm the data and the seat type (rear-facing child seat, etc.) via a keypad or by pressing a button or touchscreen, etc. Additional or alternative plausibility checks are possible.

[0051] According to one embodiment, it is advantageous to have further information, for example a clear pictogram associated with the type of child seat, or the information that the passenger safety system (passenger side) is now optimally adjusted for a child but not for adult protection, or an indication that the seat and correct use of the belt are mandatory.

[0052] Additional plausibility checks and procedures can be integrated into the approach to support the classification of a child.

[0053] For example, plausibility checks can be performed using reverse logic. If an adult is detected in a passenger seat, the system will not automatically adjust to a child seat. In this case, interaction via a human-machine interface can be provided.

[0054] Alternatively, for the example of backward logic, it can be assumed that if no adult is detected in the head box, the child hypothesis is more likely.

[0055] Plausibility can also be verified using a time-out option. If the entry in the mobile device used to provide the plausibility data or which represents the plausibility data is older than a certain period of time, which depends on the seat type (for example, three months for a baby seat, six months for a child seat), the system requires an update in the mobile device or a release confirmation. In this case, the plausibility data can include a timestamp indicating the creation time of the plausibility data and information about the seat type used by the person to whom the plausibility data is assigned. A review of the entry and / or the update or the release confirmation can, if necessary, be carried out by the mobile device or, for example, by the Fig. 1 shown device can be requested.

[0056] Fig. 6 shows a flowchart of a method according to one embodiment. According to one embodiment, the method enables image and color normalization and, additionally or alternatively, anonymous transmission of images required for a plausibility check.

[0057] A normalization process is used to normalize and anonymize image data transmitted from a mobile device, such as a smartphone.

[0058] In the Mob2Car concept outlined above, camera images of people are primarily used for plausibility checks, i.e., comparisons. These images are recorded using different camera systems.

[0059] The representation in Fig. 6 shows a first image 111 of a first device, also called device or system (A), and a second image 105 of a second device, also called device or system (B). Fig. 1 The plausibility data described may include the first recording 111. Thus, the first device may be the one described in Fig. 1 The mobile device shown is the one shown here. Fig. 1 The image data described may comprise the second recording 105. Thus, the second device may be the Fig. 1 The image capture device shown may be the image capture device. The images 105, 111 are transmitted to a block 616, in which a comparison is performed by an image processing algorithm of the second device. The first image 111 is transmitted via an interface. A result 617 of the comparison is output by block 616.

[0060] If the result 617 indicates that the images 105, 111 depict the same person, the result 617 can be distinguished from the Fig. 1 shown determination device can be used to determine occupant data regarding this person. According to one embodiment, block 616 is part of the Fig. 1 described destination device.

[0061] Depending on various camera and lighting properties, the images 105, 111 generally vary considerably. Regarding the second image 105, it should be noted that the recording situation in the vehicle is very specific. Special cameras, such as wide-angle cameras, are often used as image capture devices, as well as illumination in the non-visible range (NIR = near infrared).

[0062] A comparison algorithm performing the comparison in block 116 should therefore be very tolerant so as not to produce too many false negatives (rejections), or it should be trained for the specific recording situation. In the first case, however, this means an increased false positive rate ("FP rate"). In the second case, an algorithm that is precisely tailored to the situation is required. This eliminates the need for existing, powerful state-of-the-art algorithms. Furthermore, the second case represents a considerable application effort (image database acquisition and training).

[0063] Fig. 7 shows a flowchart of a method according to an embodiment. Shown are a first receptacle 711 of a first device and a second receptacle 105 of a second device. In contrast to Fig. 6 The first image 711 is first subjected to a transformation to obtain a normalized image 111. The normalization refers to the characteristics of the second device. The Fig. 1 The plausibility data described may include the normalized recording 111. Thus, the first device may be the one described in Fig. 1 The mobile device shown is the one shown here. Fig. 1 The image data described may comprise the second recording 105. Thus, the second device may be the Fig. 1 The normalized image 111 is transmitted from the first device and the second image 105 is transmitted from the second device to a block 616, in which a comparison is made by a

[0064] Image processing algorithm of the second device is carried out. The first image 111 is transmitted via an interface. A result 617 of the comparison is output by block 616. If the result 617 indicates that the images 105, 711 depict the same person, the result 617 can be Fig. 1 shown determination device can be used to determine occupant data regarding this person. According to one embodiment, block 616 is part of the Fig. 1 described destination device.

[0065] The Fig. 7 The approach described has the advantage of a low false positive rate and the use of conventional classification algorithms. For this purpose, the images, especially the first image 711, are normalized. The image 711 of the mobile device is subjected to a transformation (AB) that knows the characteristics of the camera of the mobile device (A) and converts them to the camera of the vehicle system (B).

[0066] The advantage of this method is that an existing powerful state-of-the-art comparison algorithm can be used without complex training and without increasing the false positive rate.

[0067] The type of parameters to be compensated for will be discussed below. Trivial differences, such as slight differences in resolution, are easily tolerated by current algorithms.

[0068] However, spectral corrections and strong distortions are important for performance: Fig. 8 shows a representation of reflectance properties of human skin types in various spectral ranges according to an exemplary embodiment. The abscissa represents the wavelength in nanometers, and the ordinate represents the reflectance. Shown are skin type 801, which is common in Asia, the light-skinned skin type 802, which is common in Northern Europe, and a particularly dark-skinned skin type 803. (Source Fig. 8 : A Generic Camera Calibration Method for Fish-Eye Lenses, Kannala, Brandt, 2004)

[0069] While the skin of a light-skinned skin type 802 appears significantly brighter (~8x) in the visible range than that of a particularly dark-skinned person (in the example), it is significantly more similar in the non-visible range (only about 2x brighter). This adaptation to, for example, a NIR system in the vehicle is ensured by the transformation (AB), through which the first image in Fig. 7 is transformed into the normalized recording.

[0070] The reflectance properties can be a parameter that can be used in the Fig. 7 described transformation can be taken into account to normalize the first image.

[0071] Fig. 9 shows a representation of geometric distortions of different camera models according to an embodiment. Shown are a perspective projection 901, a stereographic projection 902, an equidistant projection 903, an equisolid angular projection 904, and an orthogonal projection 905. (Source Fig. 9 : Human Skin Detection by Visible and Near-Infrared Imaging, Kanzawa et al, 2011).

[0072] According to one embodiment, the method based on Fig. 7 The transformation described corrects the geometric distortion of the image. The differences in camera models (for example, the mathematical camera model describes the distortions toward the edge of the image) can be very significant, even within the class of wide-angle characteristics.

[0073] The equations for curves 901, 902, 903, 904, and 905 are given below. Phi, plotted on the abscissa, is the angle of incidence, and r, plotted on the ordinate, is the distance of the image point from the principal point. f is the focal length. For the perspective projection 901: r = f tan (phi) For the stereographic projection 902: r = 2f tan(phi / 2) For the equidistant projection 903: r = f phi For the equisolid angular projection 904: r = 2f sin(phi2) For the orthogonal projection 905: r = f sin(phi)

[0074] The most common model is the decoy camera with the curve 901. For example, the imaging system (A), i.e. the one with respect to Fig. 7 the first recording device, of type 901 and the system (B), i.e. the one with respect to Fig. 7 If the second image-generating device, type 903, corresponds to a wide-angle camera that can be used in the vehicle, the transformation significantly compresses the first image towards the edges.

[0075] The input data to perform the transformation is easy to obtain. The camera models of the interior camera itself are precisely known. The characteristics of the images transmitted from the smartphone are typically less distorting than those of the interior camera. Here, one can work with typical values. However, for the characteristics of the image from the smartphone, there are also various sources of relevant data to improve the quality of the transformation: In the first case, the first image is recorded directly with the smartphone as a mobile device. Software running on the smartphone, such as an app, determines the geometric camera models of the smartphone camera (e.g., from a server database) and transmits them with the remaining data set, for example, as plausibility data, as determined using Fig. 1 According to different embodiments, the transformation can then be carried out in the smartphone itself or in the Fig. 1 shown determination device.

[0076] In the second case, the first shot was not recorded with this smartphone. Image metadata extraction can be performed. Modern cameras utilize the ability to add extensive metadata to digital images. Basic camera data is usually stored. This data is standardized in the so-called EXIF ​​format (Exchangeable Image File Format). This format allows recording parameters, focal length, resolution, etc., to be encoded.

[0077] In the third case, the first image is an image of unknown origin without a metadata file. In this case, an assumption of a typical

[0078] Geometry may be applied supported with an estimate from the image itself.

[0079] Fig. 10 shows a flowchart of a method according to an embodiment.

[0080] Shown is a first image 711 of a first device and a second image 1005 of a second device. As can be seen from Fig. 7 As described, the first image 711 is first subjected to a transformation to obtain a normalized image 1011. The normalization refers to the characteristics of the second device. In contrast to Fig. 7 the normalized first image 1011 is subjected to a feature calculation by an image processing algorithm of the first device in order to obtain first feature parameters 111 of the normalized image 1011. The Fig. 1 The plausibility data described may include the first feature parameters 111. Thus, the first device may be the device described in Fig. 1 The mobile device shown is different from Fig. 7 Furthermore, the second image 1005 is subjected to a feature calculation by an image processing algorithm of the second device in order to obtain second feature parameters 105 of the second image 1005. The Fig. 1 The image data described may include the second feature parameters 105. Thus, the second device may be the device described in Fig. 1 shown image capture device. The first feature parameters 111 are transmitted from the first device and the second feature parameters 105 are transmitted from the second device to a block 616, in which a comparison is carried out by an image processing algorithm of the second device. The first image 111 is transmitted via an interface. A result 617 of the comparison is output by block 616. If the result 617 indicates that the images 711, 1005 depict the same person, the result 617 can be Fig. 1 shown determination device can be used to determine occupant data regarding this person. According to one embodiment, block 616 is part of the Fig. 1 described destination device.

[0081] The transmission to block 617 can be encrypted. Nevertheless, a human-identifiable image of the occupant is still present somewhere in the vehicle. If this is not desired in order to protect privacy, this can be done using Fig. 10 efficient procedures described can be used.

[0082] If the normalization, which leads to the normalized image 1011, and the feature calculation, which leads to the first feature parameters 111, are combined, only the parameters 111, but not the images 711, 1011, can be transmitted via the interface to block 617. The user's face cannot be reconstructed from the parameters 111.

[0083] Advantageously, this approach ensures that no data is stored in the vehicle at any time that allows for clear conclusions about the occupant. Transmitting parameters 111 instead of images 711, 1011 also saves bandwidth.

[0084] If an embodiment includes an "and / or" link between a first feature and a second feature, this should be read as meaning that the embodiment according to one embodiment has both the first feature and the second feature and according to another embodiment has either only the first feature or only the second feature.

Claims

1. Method for providing occupant information for a safety device (106) for a vehicle (102), the method comprising the following steps: reading in (201) image data (105) representing an occupant (108) of the vehicle (102) via an interface of a determining device (116) to an image capturing device (104) of the vehicle (102), wherein the image data (105) comprise feature parameters of a recording (1005) of the occupant (108) by means of the image capturing device (104); ascertaining plausibility check data (111) using a recording (711) of a person (108) by means of a mobile device (110), wherein the recording (711) representing the person (108) is normalized to form a normalized recording (1011) using the characteristic of the image capturing device (104) by virtue of the fact that the recording (711) of the mobile device (110) is subjected to a transformation which knows a characteristic of a camera of the mobile device (110) and converts it to the image capturing device (104), wherein the plausibility check data (111) are ascertained as feature parameters of the normalized recording (1011) from which the face of the occupant (108) cannot be reconstructed, wherein the normalized recording (1011) is subjected to a feature calculation by means of an image processing algorithm of the mobile device (110) in order to obtain the feature parameters of the normalized recording (1011); reading in (203) the plausibility check data (111) representing the person (108) via an interface of the determining device (116); determining (205) occupant data (117) characterizing the occupant (108) by means of the determining device (116), after a successful plausibility check of the image data by means of the plausibility check data, wherein the plausibility check will be successful if a result (617) of a comparison of the feature parameters of the recording (1005) with the feature parameters of the plausibility check data (111) indicates that the person (108) corresponds to the occupant (108); and requesting further plausibility check data (122) representing a further person (108, 120) if, in the step of determining (205), the occupant data (117) cannot be determined using the image data (105) and the plausibility check data (111), and comprising a repeated step of determining (205) the occupant data (117) using the further plausibility check data (122); and providing (207) the occupant data (117) to an interface to the safety device (106) for the vehicle (102).

2. Method according to Claim 1, in which, in the step of reading in (203) the plausibility check data (111), further plausibility check data (122) representing the further person (120) are read in via the interface of the determining device (116) and, in the step of determining (205), the occupant data (117) are furthermore determined using the further plausibility check data (122).

3. Method according to either of the preceding claims, comprising a step of adapting the safety device (106) using the occupant data (117).

4. Apparatus (100), comprising read-in devices (112, 114), a determining device (116) and a providing device (118), for providing occupant information for a safety device (106) for a vehicle (102), which is configured to carry out the method according to any of the preceding claims.

5. Computer program which is configured to carry out the method according to any of the preceding claims.

6. Machine-readable storage medium on which the computer program according to Claim 5 is stored.