Investigation procedures for recording and processing vehicle identification elements of a vehicle

Abstract

The present invention relates to a method for detecting and processing vehicle identification elements of a vehicle, wherein these are either anonymized or encrypted so strongly that reverse engineering is impossible.

Description

[0001] In the patent application DE 10 2020 102 756 A1, an investigative method for the (partial) recording and processing of a vehicle identification means of a vehicle is shown, wherein the vehicle identification means is formed from at least two or more identification marks, and wherein, by means of a recording module, only a part of the identification marks, for example only a single identification mark, are iteratively recorded and / or further processed from the image data.

[0002] It was revealed how the existing process of capturing and processing a vehicle identification identifier and subsequent processing is iteratively modified so that an identification mark is first extracted and immediately processed, allowing for a comparison with a stored database at this point, or the verification of membership in a specific property class, and, if necessary, other identification marks to be added in further steps to be subjected to a capture and processing iteration process.

[0003] The underlying principle was, and remains, to avoid the complete disclosure of all identifying marks at any given time. Therefore, the complete means of identification should never be available at any point during the processing.

[0004] For the examples described in the aforementioned document, where a genuine subset of all identification marks is processed, the method is practical and withstands all objections. However, for applications where all identification marks must be determined and processed (see there), the method is not suitable. Fig. 6), however, it can be argued that the coding may not be strong enough to withstand a “brute force” attack.

[0005] All possible combinations of identification marks are tested to reveal the internal function 100. With a maximum of 8 characters for a vehicle identification code, using a maximum of 26 letters and 10 digits as the input set, all combinations are quickly calculated. The input for this is generated as described in... Fig. 1, for example, artificially generated pixel matrices with all possible combinations of vehicle identification means, in order to obtain the coding result of all possible inputs.

[0006] The object of the invention is therefore to develop a method that takes into account the aforementioned objection and represents the vehicle identification means either anonymously or so strongly encrypted that no reverse calculation is possible, but still ensures the desired functionalities.

[0007] DE 10 2020 102 756 A1 concerns an investigative procedure for the (partial) recording and processing of a vehicle identification means of a vehicle, wherein the vehicle identification means is formed from at least two or more identification marks, wherein, by means of the recording module, only a part of the identification marks, for example only one identification mark, are iteratively recorded and / or further processed from the image data.

[0008] The link URL: https: / / en.wikipedia.org / w / index.php?title=Cryptographic_hash_function&oldid=1056074775 in Wikipedia, the free encyclopedia, last edited on 19.11.2021, shows the so-called “Cryptographic hash function”.

[0009] The link URL: https: / / en.wikipedia.org / w / index.php?title=Bruteforce_attack&oldid=105345656 in Wikipedia, the free encyclopedia, last edited on 04.11.2021, shows the so-called "brute-force attack".

[0010] The present invention relates to a method for detecting and processing vehicle identification elements of a vehicle and to a device for detecting and processing vehicle identification elements of a vehicle according to the respective preambles of claims 1 and 10.

[0011] The investigation procedure described here for capturing vehicle identification elements (VINs) comprises a first step in which at least one VIN acquisition module is activated. This VIN module can include at least one VIN, in particular an optical sensor. Furthermore, the VIN module can include an interface module for processing the pixel data captured by the VIN. In a further step, the VIN module identifies a pixel sub-area within which at least some of the VINs are located.

[0012] According to the invention, the present invention comprises a method for determining a vehicle identification means of a vehicle, wherein the vehicle identification means is formed with at least two or more vehicle identification elements.

[0013] According to the invention, in a first step at least one detection module for detecting the vehicle identification means and / or the vehicle identification elements is put into operation, wherein the detection module comprises at least one detection element, in particular an optical camera or optical sensor, and an interface module for processing the pixel matrix detected by the detection element.

[0014] In a further step according to the invention, the detection module determines at least a sub-area of ​​the pixel matrix within which at least a part of the vehicle identification means is arranged.

[0015] In a further step according to the invention, either the sub-area of ​​the pixel matrix or one or more sub-areas thereof are encoded by means of the acquisition module according to a specially generated hash value, and the hash values ​​do not correspond to the usual ASCII or ANSI values.

[0016] According to the invention, the vehicle identification means is compared with a predetermined list of vehicle identification means, using the same hash table for the comparison.

[0017] According to the invention, the hash table is regenerated each time a vehicle identification device is newly registered. This hash table is not generated just once, but anew for each vehicle check, using randomly generated hash values ​​for encoding the characters and pixel matrices. Insofar as the vehicle is registered with a pixel matrix, a new hash table is generated, and the list of authorized users is re-encoded with respect to this new hash table and subsequently permuted, the temporary table is discarded after completion of the process. This ensures that there is no possibility of decrypting the nature of the procedure, thus rendering the keys lost and the procedure anonymous. Consequently, different results are obtained with the same input, making an analytical revelation of the algorithm behind the procedure practically impossible.

[0018] According to at least one embodiment, the given list is randomly permuted after encoding according to the hash table, and the permutation key is optionally discarded to prevent a reverse calculation to the original list.

[0019] According to at least one embodiment, the hash values ​​of the hash table correspond to class “B” for all letters and to class “Z” for all digits.

[0020] According to at least one embodiment, the date, time and a location identifier or direction identifier or station type are also stored with the determined identification string.

[0021] According to at least one embodiment, there are at least two detection devices, which are either implemented on a single piece of hardware, or, if physically separate, the data from one detection element are transferred to the other for comparison.

[0022] According to at least one embodiment, the processing is combined with another cryptographic method.

[0023] According to at least one embodiment, the process is slowed down internally in such a way that a brute force attack is made impossible due to the computation time required.

[0024] According to the invention, the present invention comprises a device for detecting a vehicle identification means of a vehicle, wherein the vehicle identification means is formed with at least two or more vehicle identification elements.

[0025] According to the invention, in a first step at least one detection module for detecting the vehicle identification means and / or the vehicle identification elements is put into operation, wherein the detection module comprises at least one detection element, in particular an optical camera or optical sensor and an interface module for processing the pixel matrix (PM1) detected by the detection element.

[0026] In a further step according to the invention, the detection module determines at least a sub-area of ​​the pixel matrix within which at least a part of the vehicle identification means is arranged, and wherein in a further step, either the sub-area of ​​the pixel matrix or one or more sub-areas thereof is encoded PM3 by means of the detection module according to a hash value generated specifically for this purpose, and wherein the device is set up and provided for carrying out the detection method according to claim 1.

[0027] An optical sensor is used to capture an image of the front or rear of a vehicle ( Fig. 1), which is referred to below as the pixel matrix PM1(x,y). The variable x ranges from 1 to the number of columns, and the variable y from 1 to the number of rows. For example, the number of columns could be 1920 and the number of rows 1080, which is referred to as HD resolution. Here, we deliberately refer to a pixel matrix of an optical sensor rather than an image, because an image refers either to a display on a monitor or a file that can be displayed on a monitor. The display or storage of images of vehicles is not necessarily permitted by law.

[0028] In a further step, the acquisition module determines a pixel sub-area PM2 ( Fig. 2) within which at least one part of the vehicle identification means is contained, including at least one vehicle identification element. The detection module can divide the pixel sub-area PM2 into further sub-areas PM3 ( Fig. 3) Dissect it. You can see a single vehicle identification element as a pixel matrix.

[0029] The task of conventional Optical Character Recognition (OCR) is to encode the individual pixel sub-areas PM3 according to a generally known ASCII or ANSI table. As in Fig. As shown in figure 4, the matrix depicted there is assigned the value 67, which corresponds to the character 'C' in word processing. Using this method, one would proceed as in Fig. 5 shows a coding sequence 67 65 82 45 82 69 65 68 69 82, which corresponds to the text string ,C' ,'A' ,'R' ,'-„ ‚R' ,'E' ,'A' ,'D' ,'E' ,'R'.

[0030] In the present invention, the pixel matrices are not encoded according to known tables, but rather according to predefined table values. In cryptography, the elements of the table are called hash values, and the table is called a hash table. A pixel matrix of Fig. 3 is assigned a hash value of, for example, "j%6Fa / ". The characters used in the hash value and the number of characters per value are arbitrary. However, they must not correspond to the ASCII or ANSI character sets. The final result is a seemingly arbitrary string S such as "38uergdhuu47uv8948rvjadu,0f0f9v9iwpowoejf9jfiwviwooivi49ä 909wejfi9-ou39f 93ru9weu9t340ftg4".

[0031] In addition to the usual vehicle identification marks such as in Fig. 3. Furthermore, other vehicle identification elements such as badges, spacing between characters, trademarks, brand shapes, or other characteristics can also be integrated. Vehicle identification marks are also to be understood as vehicle identification elements, so that the elements form a superset, and, unlike previously, the term "vehicle identification elements" is generally used here.

[0032] If the aforementioned hash table is unknown, the process involves anonymization. Anonymization occurs when (personal) data is altered in such a way that it can no longer be attributed to a specific or identifiable natural person, or only with a disproportionate expenditure of time, cost, and labor. In this case, the personal data is the official vehicle registration number, which exists in the official database as an ASCII or ANSI code, and which can be used to identify a natural person, the vehicle owner. A partial objective of the invention is to identify a vehicle in such a way that at no time is it possible to trace the vehicle owner back to the natural person via the vehicle's official registration number.

[0033] The following section will show how this new technology can be used to implement the desired applications.

[0034] The topic of encryption is the domain of cryptography, which generally deals with information security, i.e., the design, definition, and construction of information systems that are resistant to manipulation and unauthorized reading. A recognized method, the asymmetric RSA algorithm, uses a key pair consisting of a private key for encryption and a public key for decryption. The private key is kept secret and cannot be calculated from the public key with reasonable effort. The method works as follows: Choose two prime numbers p and q that are as large as possible and form the product n = p*q. For the Euler function φ(n), which indicates how many natural numbers are coprime to a given natural number n, the following holds: φ(n) = φ(p-1)*φ(q-1). Euler's theorem then states: for all natural numbers k and m with m <n gilt m k*φ(n)+1 = m modulo n. If we choose another natural number e with gcd(e,φ(n))=1 (gcd = greatest common divisor), we can determine a natural number d with e*d = 1 modulo φ(n). The keys we are looking for are then the pairs d and e. To encrypt, we form m e = c modulo n and to decrypt c d = m modulo n, because c d = (m e ) d = m e*d = m k*φ(n)+1 = m modulo n.

[0035] In practice, the process can be visualized as a special lock with positions A (closed), B (open), and C (closed). The private key allows movement from position A to B and from position B to C. The public key, however, allows movement from position C to B and from position B to A. The lock can only be opened in position B. When a message is placed in the box and the public key, which everyone possesses, is used, the lock is moved to the closed position A. Only the person who possesses the private key can then move the lock back to the open position B and read the message. Thus, the public key encrypts the message, and the private key decrypts it.

[0036] The application of the procedure described here to check whether a vehicle is included in a given list of vehicles can be illustrated in a simple variant as follows: A defined set of at least two vehicle owners has consented to the processing of their vehicle registration number, for example, to enable them to enter a parking garage without further identification and have the barrier open automatically. The procedure described here first generates a list of hash values, assigning a hash value to each character (A to Z) and letter (0 to 9). The individual vehicle registration numbers are then recoded according to this list, with the index value determining the encoding. The recoded list elements are then randomly reordered using a permutation. A permutation is an arrangement of specific objects, where each element may only appear once. The position of the elements is arbitrary.The permutation key (the mapping of the objects to their new positions) is discarded after completion, so it is no longer possible to determine the original position of a vehicle license plate in the list. While all vehicle license plates are still present in the list in coded form, their position within the list is lost. The task is to determine whether a vehicle approaching the barrier is included in this list or not. The system is not concerned with which specific vehicle in the list it is, but only with the information "element of the list" or "not an element of the list." If the vehicle is an element of the list, the barrier can be opened because all elements of the list are permitted to enter.In the present invention, this is achieved as follows: the processing element generates a pixel matrix of the vehicle via the optical sensor, containing vehicle identification elements, in particular the vehicle identification markings. The pixel matrix is ​​processed until the state of... Fig. The data is reduced in 3, and then the reduced subpixel matrices are encoded according to the hash table defined above. The resulting string S can then be combined with the list of authorized vehicles, which is encoded and permuted according to the hash table and its indices. If a match occurs, the system knows that the vehicle in front of the barrier is an element of the list and can open the barrier. Only now is the system authorized to process the vehicle registration number in plain text, i.e., as an ASCII or ANSI code, because all list elements have consented to the processing. However, if the vehicle is not in the list, the process is discarded without any possibility of relating it to the official vehicle registration number.

[0037] In this respect, the invention employs a more stringent variant as described above, with the difference that the randomly generated hash values ​​for encoding the characters and pixel matrices are not created only once, but anew with each vehicle check. That is, the vehicle is recorded with a pixel matrix, a new hash table is generated, and the list of authorized users is re-encoded with respect to this new hash table and then permuted. The temporary table is discarded after completion of the process, so that there is no way to decrypt the nature of the procedure. The keys are lost, and the procedure is therefore anonymous. Furthermore, the distinctive feature of this procedure is that different results are obtained with the same input, which makes an analytical revelation of the underlying algorithm practically impossible.

[0038] The same applies to the application of the procedure described here for verifying an access violation.

[0039] Residential areas may only be accessed by authorized residents. Unauthorized access will be detected and prosecuted accordingly. To be authorized to enter, vehicle owners must consent to the recording and processing of their license plate number. This information is then collected in a list, as described above, and encoded according to a hash table. This hash table is either generated once at the start of the process or, as in the stricter version described above, regenerated after each vehicle registration. The encoded license plate list is then permuted, and the permutation key is discarded. When a vehicle enters a residential area, it is processed as described, and if it is authorized (an element in the list), it is discarded. An unauthorized vehicle, i.e., one not in the list, may then be processed further.

[0040] The procedure described here is used to deduce the origin or affiliation to a country or city from the letter and number combination of the vehicle registration plate.

[0041] As in Fig. As shown in Figure 8, a vehicle registration plate consists of a location identifier, a validation sticker or space between characters, and the identification numbers or pre-registration marks. The identification numbers or pre-registration marks are only relevant for this application insofar as they belong to the letter or number classes. Therefore, all letters from A to Z are grouped into class "B," and all numbers from zero to 9 into class "Z." The in Fig. The vehicle registration number R EN 166 shown in 8 is thus reduced to: R “B”, “B” “Z” “Z”, because E and N are letters and 1 and 6 are numbers.

[0042] This can be easily implemented using the procedure described above. All letters of the registration codes are assigned to hash table "B" and all digits to hash table "Z". Strict regulations exist for each country regarding the number of letters and digits that may be used for the registration codes, and in what order, allowing the country of origin to be determined. Together with the location identifiers, this allows for a very precise narrowing down of the origin for statistical purposes.

[0043] The method described here is also used to detect a vehicle at at least two different locations or at different times in order to deduce either the duration of stay or the average speed of the vehicle. For example, supermarket parking lots or similar areas have a maximum parking time for shopping that must not be exceeded. For this purpose, the vehicle is detected at the entrance and at the exit, and the time difference can then be used to determine whether the maximum parking time has been exceeded. Usually, the entrance and exit are spatially separated, so two different optical sensors are used. However, if the two are next to each other, or if there is only one lane used for both entry and exit, a single sensor can suffice, but this sensor must then also detect the direction of travel for differentiation purposes.The present invention provides several methods for recording vehicles in such a way that the vehicle identification number is only available in plain text to issue a warning or impose a fine on the associated owner if the maximum parking time has actually been exceeded. All other vehicles are to be recorded anonymously with the time and date, and this data is deleted after the vehicle leaves the parking area.

[0044] In the first variant, the entry detection module processes the vehicle identification elements into a sufficiently highly encoded string. Currently, values ​​of 256 bits are considered adequate. With further improvements in computing power, the encoding should be adjusted accordingly, which, as described above, is easily possible. The hash table required for the encoding can either be predefined or stored with the encoded string. Furthermore, the date and time of entry, along with any location (e.g., Entry 1 or Test Lane Entry), are also stored. To increase complexity, the procedure can be combined with another method, such as the RSA method described above. With separate entry and exit points, the encoded vehicles are transmitted either to a central location or to the exit detection module.With identical entry and exit detection modules, this is unnecessary. The situation then becomes the list comparison procedure described above. An exiting vehicle is detected and its data is compared to the list of vehicles that have already entered. Either a predefined hash table is used to encode the vehicle identification elements, or, if these tables are included in the coding string, the data is encoded and compared according to this table, which may be structured differently for each vehicle that has entered.

[0045] In another variant, the pixel matrices of entering vehicles, either in their entirety or in their reduced forms PM2 or even PM3, are transmitted to a central location—which can also be implemented using an exit detection module—along with the date and time, and ideally stored there in encrypted form. When a vehicle exits, it is recorded and processed using either a static hash table or, preferably, a randomly generated one. The stored pixel matrices of the entering vehicles are then encoded with the same hash table (static or variable) so that they can be combined with the exit process. If the duration calculated from the time difference between the entry and exit processes exceeds a certain limit, the parking facility operator has a legitimate interest in recording the vehicle in plaintext to determine the owner.The pixel matrices can then be converted into ASCII or ANSI values ​​for further processing, as is done in conventional OCR.

[0046] The application of the procedure described here for determining the average speed of a vehicle is similar to that described above. At least two checkpoints are equipped with recording systems that, in addition to vehicle identification elements, also register the location and time of passing vehicles. As described above, the recorded data pairs are combined to identify a specific vehicle, and the average speed is calculated from the distance between the locations and the timestamps. If this average speed exceeds a predetermined maximum speed, there is a legitimate interest, or rather a legal basis, to record the official vehicle registration number and process it for prosecution purposes. The procedure is described in Fig. 9 is presented as a flowchart for better understanding.

[0047] One application of the method described here is automated exit after payment in parking garages. In parking garages, the operator wants to simplify the exit process for drivers who have already paid their ticket by having the barrier open automatically. For this purpose, the vehicle's license plate is recorded at the entrance and linked to the ticket data. If the driver correctly pays the parking fee, the vehicle is unlocked for exit. This means that upon exiting, the license plate is recorded again, and if payment is successful, the barrier opens automatically. The driver no longer needs to manually validate the unlocked exit ticket; this process can be automated via the link between the ticket and license plate.Unfortunately, recording and processing vehicle license plates at the entrance is not compliant with data protection regulations because drivers who do not wish to be recorded have no way to avoid it. Therefore, as described above, a vehicle identification number must be recorded anonymously and, as also previously described, deleted during the exit process. This ensures the operator is compliant with data protection regulations.

[0048] To counter an attack by trying all possible combinations (brute force method) in the procedures described above, the procedure 100 can be used. Fig. 1. The process can be modified internally so that the calculation of the result, i.e., the anonymized identification string S, is delayed to a minimum evaluation time, independent of the hardware platform. If this is set to 100 milliseconds, for example, which does not lead to any significant limitation in real-world practice, almost 9000 years of computing time would be required for all combinations of 8 identification elements with 26+10=36 possibilities.

[0049] The invention is not limited to a description based on an exemplary embodiment. Rather, the invention encompasses any new feature, as well as any combination of features, which in particular includes any combination of features in the patent claims, even if this feature or combination itself is not explicitly specified in the patent claims or in the exemplary embodiments.

[0050] The following section explains the method described here in more detail using an exemplary embodiment and the associated figures.

[0051] The Fig. Figure 1 shows the investigation procedure (100) for capturing and processing vehicle identification elements (11) of a vehicle (10) by means of an optical sensor, in which an image of the front or rear of a vehicle is captured, which is referred to below as a pixel matrix (PM1) (x, y). The variable x ranges from 1 to the number of columns and the variable y from 1 to the number of rows. For example, the number of columns could be 1920 and the number of rows 1080, which is referred to as HD resolution. This is deliberately not an image, but a pixel matrix of an optical sensor, because an image refers either to a display on a monitor or a file that can be displayed on a monitor. The display or storage of images of vehicles is not necessarily permitted by law.

[0052] The Fig. Figure 2 shows the next step, in which the detection module (2) determines a pixel sub-area (PM2) within which at least a part of the vehicle identification means with at least one vehicle identification element is contained.

[0053] In Fig. Figure 3 shows that the acquisition module (2) can divide the pixel sub-area (PM2) into further sub-areas (PM3), where a single vehicle identification element (11) is represented as a pixel matrix. The pixel matrix shown is assigned a hash value of, for example, "j%6Fa / ". The acquisition module (2) can include at least one acquisition element (21), in particular an optical sensor. Furthermore, the acquisition module (2) can include an interface module (22) for processing the pixel data acquired by the acquisition element (21). In a further step, the acquisition module (2) determines a pixel sub-area within which at least one of the vehicle identification elements (11) is located. The characters used in the hash value and the number of characters per value are arbitrary. However, they must not correspond to the ASCII or ANSI tables.The end result is a seemingly arbitrary character string S such as “38uergdhuu47uv8948rvjadu,0f0f9v9iwp0woejf9jfiwviwooivi49ä 909wejfi9-ou39f 93ru9weu9t340ftg4” as the processing result.

[0054] In addition to the usual vehicle identification numbers from the Fig. 3. Furthermore, other vehicle identification elements (50), such as badges, spacing between characters, trademarks, brand shapes, or other characteristics, can also be integrated. Vehicle identification marks are also to be understood as vehicle identification elements, so that the elements are a superset and, in contrast to earlier, one generally speaks of vehicle identification elements.

[0055] First, the process described here generates a list of hash values, assigning a hash value to each character (A to Z) and letter (0 to 9). The individual vehicle registration numbers are then recoded according to this list, with the index value determining the encoding. The recoded list elements are then randomly reordered using a permutation. A permutation is an arrangement of specific objects, where each element may only appear once. The position of the elements is arbitrary. The permutation key (the mapping of the objects to their new positions) is discarded after completion, so that it is no longer possible to determine the position of an original vehicle registration number in the list.The processing element uses the optical sensor to generate a pixel matrix of the vehicle, which includes vehicle identification elements, in particular the vehicle identification numbers. The pixel matrix is ​​processed up to the point specified in the... Fig. The state shown in Figure 3 is reduced, and then the reduced subpixel matrices (PM3) are encoded according to the hash table defined above. This allows the resulting string S to be matched with the list of authorized vehicles, which is encoded and permuted according to the hash table and its indices. If a match occurs, it is known that the vehicle in front of the barrier is an element of the list, and the barrier can be opened. Only now is it possible to process the vehicle registration number in plain text, i.e., as an ASCII or ANSI code, because all list elements have consented to the processing. However, if the vehicle is not in the list, the process is discarded without any possibility of relating it to the official vehicle registration number.

[0056] As in Fig. As shown in Figure 4, the task of conventional Optical Character Recognition (OCR) is to encode the individual pixel sub-areas (PM3) according to a generally known ASCII or ANSI table, because the depicted matrix (50) is assigned a value 67, which corresponds to the character 'C' in text processing.

[0057] Thus, with this method, as in Fig. 5 shown, a coding sequence 67 65 82 45 82 69 65 68 69 82 was obtained, which corresponds to the text string ,C','A','R','-,,'R','E','A','D','E','R'.

[0058] For better understanding, the procedure is explained in the Fig. 6, Fig. 7 and Fig. 9 is presented as a flowchart.

[0059] The procedure is in Fig. 6. The process is disclosed, with the first step involving the detection of moving objects or triggering via a loop (200) and the second step involving the vectorization of vehicle data (300). In the next step, anonymous vector data is matched with an anonymized list (400), whereby the list of vehicles or customers who have consented to data processing is first created (600) and then the vehicle data is anonymized so that it cannot be reverse-converted (700).

[0060] If the vehicle is on the list, the barrier opens for the authorized vehicle (800). Alternatively, if the vehicle is not on the list, the data is discarded and the process is terminated (500).

[0061] Compared to Fig. 6, will be in the Fig. 7 In the event that the vehicle is not found in the list, the image of the vehicle is saved and processed further (900).

[0062] As in Fig. As shown in Figure 8, a vehicle registration plate consists of a location identifier, a validation sticker or space between characters, and the identification numbers or pre-registration marks. The identification numbers or pre-registration marks are only relevant for this application insofar as they belong to the letter or number classes. Therefore, all letters from A to Z are grouped into class "B," and all numbers from zero to 9 into class "Z." The in Fig. The vehicle registration number R EN 166 shown in the image is thus reduced to: R "B", "B" "Z" "Z" "Z", since E and N are letters and 1 and 6 are digits. This can be easily implemented using the procedure described above. All letters of the registration codes are assigned to hash table "B" and all digits to hash table "Z". Strict regulations exist for each country regarding how many letters and digits may be used for the registration codes and in what order, allowing the country of origin to be determined. Together with the location identifier, the origin can then be narrowed down very precisely for statistical purposes.

[0063] The one in Fig.The flowchart shown in Figure 9 illustrates that the first step involves capturing moving objects or triggering them via a loop (200), followed by vectorizing the vehicle data in a second step (300). The next step determines whether the process is entering (1100) or exiting (1200). For an entering process, the anonymous vector data is stored in an (presence) list (1300). Alternatively, for an exit process, the anonymous vector data is matched with the presence list, and the duration / speed is calculated (1400). If a threshold (speed / duration) is exceeded, the vehicle image is saved and further processed (1500). Reference symbol list 1 Vehicle identification device 2 Data acquisition module 10 vehicles 11 Identification element 21 Recording element 22 Interface module PM1 captured pixel matrix PM2 sub-area of ​​the pixel matrix PM3 is another sub-area of ​​the pixel matrix 50 vehicle identification elements as a pixel matrix S Anonymized identification string 100 procedures 200 Detection of moving objects or triggering via loop Vectorizing 300 vehicle data points Matching 400 anonymous vector data points with an anonymized list Discard 500 data entries, end process Create a list of 600 vehicles or customers / residents who have consented to data processing. 700 Anonymization of vehicle data so that it cannot be reversed 800 Open barrier for authorized vehicle 900 Save and process the image of the vehicle 1000 devices 1100 Entry procedure 1200 Exit procedure Store 1300 anonymous vector data points in an (attendance) list Match 1400 anonymous vector data points with an attendance list and calculate duration / speed. 1500 If a threshold (speed or duration of presence) is exceeded, save and process the image of the vehicle.

Claims

[1] Determination method (100) for recording a vehicle identification means (1) of a vehicle, wherein the vehicle identification means (1) is formed with at least two or more vehicle identification elements (11), - in a first step, at least one acquisition module (2) for acquiring the vehicle identification means (1) and / or the vehicle identification elements (11) is put into operation, wherein the acquisition module (2) - at least one detection element (21), in particular an optical camera or optical sensor - and includes an interface module (22) for processing the pixel matrix (PM1) captured by the detection element (21). - in a further step the detection module (2) determines at least one sub-area (PM2) of the pixel matrix (PM1) within which at least one part of the vehicle identification means (1) is arranged, and wherein - in a further step, using the acquisition module (2), either the sub-area of ​​the pixel matrix (PM2) or one or more sub-areas thereof (PM3) is encoded according to a specially generated hash value and - the hash values ​​do not correspond to the usual ASCII or ANSI values - the vehicle identification means is compared with a predefined list of vehicle identification means, - using the same hash table for the comparison, - the hash table is regenerated each time a vehicle identification means (1) is newly registered, the hash table with randomly generated hash values ​​for encoding the characters and pixel matrices is not only generated once, but is regenerated for each vehicle check, and insofar as the vehicle is registered with a pixel matrix and a new hash table is generated and the list of authorized users is re-encoded with respect to this new hash table and subsequently permuted each time, the temporary table is discarded after completion of the process, so that there is no possibility of decrypting the type of procedure, and thus the keys are lost, the procedure is therefore anonymous, whereby different results are obtained with the same input, which makes an analytical discovery of the algorithm behind the procedure practically impossible. [2] Investigation method (100) according to claim 1, characterized by , that - the given list is randomly permuted according to the hash table after encoding and - the permutation key is optionally discarded to prevent a recalculation to the original list. [3] Investigation method (100) according to claim 1, characterized by , that - the hash values ​​of the hash table for all letters of a class "B" and all digits of a class "Z" correspond [4] Investigation method (100) according to claim 1 characterized by , that - the date, time and a location identifier or direction identifier or station type are stored in addition to the determined identification string (S). [5] Investigation method (100) according to claim 4, characterized by , that - there are at least two detection devices that either - are implemented on a uniform hardware, or - if physically separated, the data from one acquisition element is transferred to the other for comparison. [6] Investigation procedure (100) according to at least one of the preceding claims, characterized by , that - the processing is combined with another cryptographic method. [7] Investigation procedure (100) according to at least one of the above claims, characterized by , that - the procedure (100) is slowed down internally to such an extent that a brute force attack is made impossible due to the computation time then required. [8] Device (1000) for detecting a vehicle identification means (1) of a vehicle, wherein the vehicle identification means (1) is formed with at least two or more vehicle identification elements (11), - in a first step, at least one acquisition module (2) for acquiring the vehicle identification means (1) and / or the vehicle identification elements (11) is put into operation, wherein the acquisition module (2) - at least one detection element (21), in particular an optical camera or optical sensor - and includes an interface module (22) for processing the pixel matrix (PM1) captured by the detection element (21). - in a further step the detection module (2) determines at least a sub-area of ​​the pixel matrix (PM2) within which at least a part of the vehicle identification means (1) is arranged, and wherein - in a further step, by means of the acquisition module (2), either the sub-area of ​​the pixel matrix (PM2) or one or more sub-areas thereof (PM3) is encoded according to a hash value generated specifically for this purpose, and wherein the device (1000) is set up and provided for carrying out the determination method (100) according to claim 1.

Images