Image authenticating methods

The present invention relates to methods for the authentication of images and in particular those of vehicles taken in infringement, such infringements being able for example to be related to an exceeding of the authorized speeds, to a crossing of red lights or to the circulation of an unauthorized vehicle in a lane reserved for public transport vehicles.

Indeed, until now, two methods of control are used, namely methods requiring human intervention during the observation of the offense or semi-automatic methods with image capture of the infringing vehicles.

In the first case, the speed checks require the physical intervention of law enforcement officers who in general and at first find the offense.

The report resulting from this report is then used to sanction the infringement and if necessary implement means to charge the driver in breach of the corresponding fines.

However, such a sequence of operations requires human intervention at all stages of the process.

Thus, the probability of a speed control remains relatively low and the heaviness of the subsequent processing of the minutes leads to a rather low rate of collection of fines.

This has the effect of giving some motorists the impression of a feeling of impunity, which is negative in terms of safety.

The same problem arises for offenses of different natures, for example, the crossing of a red light or traffic in a reserved lane.

In the second case, the introduction of automated procedures ranging from the finding of the offense to the recovery of the sums claimed for the fines seems likely to greatly improve the safety and the respect of the traffic rules.

Several approaches have been proposed in the past to try to automate such procedures.

For example, it has been proposed in the patent <patcit id="pcit0001" dnum="US5381155A"><text>US5,381,155</text></patcit> to use a Doppler radar to first measure the speed of the vehicles and thus be able to see if they are in violation, then trigger a camera to capture images of the vehicle or vehicles in violation.

These images are then transmitted to a computing unit to enable recognition and identification of the license plates of the vehicles in question, and then the images can then be stored in non-volatile memories.

When the said license plates have been identified, it is then possible to transmit the registration numbers of the vehicles in violation by the telecommunications systems reserved for the police and thus to allow the intervention of the latter.

The presence of law enforcement officials is then necessary for the detection of the offense.

If the infringement is contested by the driver or drivers concerned, the images recorded at the time of the offense may be extracted from the memory in which they were stored and used.

Such an approach, however, faces a major obstacle.

It is indeed easy, for example by using a graphic editing software, to modify the stored images and to replace, for example, license plate numbers by others.

Since such manipulation can be easily implemented, the legal value of transmitted images is greatly reduced.

<patcit id="pcit0002" dnum="EP0621572A"><text>EP 0 621 572</text></patcit> Al also discloses an image authentication method having means for encrypting and compressing images. In an attempt to remedy this drawback, it has been proposed, for example, in patents<patcit id="pcit0003" dnum="WO0208400A"><text>WO 02 08400 A</text></patcit> and <patcit id="pcit0004" dnum="EP0621572A"><text>EP 0 621 572 A</text></patcit> to encrypt the images to make any manipulation difficult.

Such an approach, however, has serious disadvantages.

Indeed, the dissemination of these images is then allowed only to police forces accessing the decryption algorithms and in any case to the offender in case of dispute by the latter.

Moreover, once the images have been decrypted, their alteration is again possible and the legal value of the transmitted images is again strongly reduced.

In an attempt to avoid these disadvantages, it has been proposed in the patent <patcit id="pcit0005" dnum="US5563590A"><text>US5,563,590</text></patcit> to insert in the image taken at the time of the offense, information relating to the speed of the vehicle, the time and place of the offense, etc. in the form of alphanumeric characters.

From the alphanumeric information thus collected, new alphanumeric control characters are constructed which are also inserted in the image.

The photographs corresponding to the infringing vehicles contain both the above characteristic information and the alphanumeric control characters.

Subsequently, in case of dispute of these documents, it is possible to verify that the alphanumeric characters of control are those which corresponded to the characteristic information taken during the infringement.

However, the device described in this patent has significant disadvantages of various kinds.

In particular, he uses film photography techniques that require chemical development of the films.

This entails a need for regular and costly human interventions, for example to load the film reels into the cameras and replace them when they have been used.

The mere use of digital storage media in place of silver media does not solve these problems.

In fact, the alphanumeric characters used to characterize the images and to ensure their control appear in this case in a manner directly identifiable on the images and it is relatively easy to modify them, for example using the previous graphic editing software.

In addition, this information obscures part of the image which may give rise to disputes in some cases.

In another improvement it has been proposed in the patent <patcit id="pcit0006" dnum="US6269446B"><text>US6,269,446</text></patcit> to compute a digital signature from the images, this signature being placed hidden and non-standardized, in the header of the image files.

This solution, however, has at least three serious disadvantages:<ul><li>firstly, some image file formats do not have a header, in particular most of the files representing images recorded without processing of the type of image compression, ie those providing the best definition,</li><li>secondly, the signature in question being hidden, it may be challenged by the offending motorists because it is not an integral part of the elements of the legal file of the offense,</li><li>Finally, because of the non-standard nature of these operations, this signature can be erased irreversibly during simple file backup operations.</li></ul>

The present invention is intended in particular to provide a method for authenticating images, particularly images of vehicles taken in infringement, and for this purpose, a method according to the invention comprises the following steps:<ul><li>distribution systems are arranged, arranged to allow the taking of images and the entry of the identification elements of the offenders, the means of shooting providing data representative of the images taken, hereinafter called data of pictures taken,</li><li>means are provided, hereafter referred to as informative systems, for triggering physical information relating to the offense, measuring speed, time, date, location, etc. hereinafter referred to as infringement data,</li><li>first means for storing and / or transmitting the captured image data and the infringement data are provided,</li><li>operating systems are provided for exploiting the stored and / or transmitted data,</li></ul>and is essentially characterized in that:<ul><li>the operating systems apply to the image data taken, any known processing to improve or maintain the quality of the images in question and / or to reduce the number of data necessary for a reconstruction of these images, without significant loss of their quality , in order to reduce the size of the memories necessary for storing the image data taken and / or the capacity of the means of transmission of these data, the intermediate data representative of images after these processing being called initial graphic data,</li><li>the operating systems calculate, from the infringement data and a graphical representation of the alphanumeric characters associated with the infringement data, new data representative of images, called graphic infringement data,</li><li>the operating systems merge the initial graphical data and the graphical offense data to obtain a new set of image representative data, referred to as the graphical identifying data, in which the initial graphical data and the graphic infringement data constitute accessible subsets of this new dataset,</li><li>the operating systems calculate, by application to the graphical identifying data of a non-bijective function denoted f, a set of data, hereinafter referred to as summarized data, so that knowledge of the summarized data alone does not permit go back to the graphical identifying data,</li><li>the operating systems apply to the summary data a coding method denoted c, having an associated decoding method noted c^-1, to obtain a new set of data called signature data,</li><li>the operating systems calculate, from the signature data and a graphical representation of the alphanumeric characters associated with this signature data, new data representative of images, called graphic signature data,</li><li>the operating systems merge the graphical identifying data and the graphical signature data to obtain a new set of image representative data, in which the graphical identifying data and these graphic signature data constitute accessible subsets of this new set. of data, called graphical authenticatable data,</li><li>second means are provided for storing and / or transmitting the authenticatable graphic data,</li><li>control units are provided which can respectively read and / or receive the graphical authenticatable data stored in the second storage and / or transmission means, the data actually read and / or received being called graphics received data,</li><li>the control units search among the received graphic data the subset of the graphical identifying data, hereinafter called the graphic identifying data tested,</li><li>the control units search among the received graphic data, the subset of the graphic signature data, hereinafter called the graphic signature data tested,</li><li>the control units search from the tested graphic signature data and from an alphanumeric character recognition table, a set of data representative of the signature data, called the tested signatures data,</li><li>the control units calculate by the application, to the graphical identifying data tested, of the non-bijective function f, a set of data, called summarized data tested,</li><li>the control units apply to the tested signature data, the decoding method c^-1 to obtain a set of data, called summary data received,</li></ul>and the control units compare the summarized data received and the summarized data tested, and provide an alert signal when these data are not identical and / or a confirmation signal when they are.

In preferred embodiments of the method according to the invention, use is also made of one and / or the other of the following provisions:<ul><li>the information systems comprise means for measuring the speed of vehicles,</li><li>the information systems include means for detecting the presence of an unauthorized vehicle in a reserved lane,</li><li>the information systems comprise means for detecting the crossing of a red light by a vehicle,</li><li>the camera systems provide digital images,</li><li>the coding and / or decoding methods use cryptographic techniques,</li><li>the encoding methods incorporate in the signature data, a subset of data, subset accessible and containing all the alphanumeric characters sufficient to represent the signature data,</li><li>the operating systems apply to the image data taken successively a compression method and the associated decompression method, and store and / or transfer the data obtained to storage and / or transmission means,</li><li>the first storage and / or transmission means and the second storage and / or transmission means are merged,</li><li>the character recognition table is developed by application of a character recognition program code,</li><li>the character recognition table is developed from the graphical signatures data tested.</li></ul>

The object of the invention can be obtained by a method of authentication of images and in particular images of vehicles taken in violation according to claim 1. Other features and advantages of the invention will become apparent during the description following detailed of one of its embodiments, given by way of non-limiting example with reference to the accompanying drawings.

On the drawings:<ul><li>the <figref idref="f0001">figure 1</figref> is a schematic view of a method according to the prior art where a vehicle (1) moving on the road (100) in the direction indicated by the arrow (F), is intercepted by the radar beam (200) of a informative system including in particular a speedometer (2), connected to operating systems (4), and to camera systems (3),</li><li>the <figref idref="f0001">figure 2</figref> is a schematic view of a step of the method according to the invention where a processor (400) of the operating systems merges the initial graphics data (10) of a vehicle (1) taken from the front and comprising elements of identification (101) and the graphical offense data (11) to obtain the graphical identifying data (12),</li><li>the <figref idref="f0001">figure 3</figref> is a schematic view of an intermediate step of the method according to the invention where a processor (400) of the operating systems calculates from the graphical identifying data (12) and by action of a suitable program code (40) placed in nonvolatile storage means, the summarized data (31),</li><li>the <figref idref="f0002">figure 4</figref> is a schematic view of an intermediate step of the method according to the invention in which a processor (400) of the operating systems (4) calculates from the summarized data (31) and by the action of an appropriate program code (41) placed in a nonvolatile storage means, the signature data (33),</li><li>the <figref idref="f0002">figure 5</figref> is a schematic view of an intermediate step of a method according to the invention where a processor (400) of the operating systems (4) merges the graphical identifying data (12) and the graphical signature data (13), to constitute the authenticatable graphics data (14),</li><li>the <figref idref="f0003">figure 6</figref> is a schematic view showing an example of sequence of steps of the method according to the invention, the operating systems and the control units, here placed outside not to weigh down the figure, being connected by any known means to the various elements .</li></ul>

When the speed of a vehicle (1), comprising an identification element (101), such as a license plate, exceeds the authorized speed limit, a device according to the prior art comprising a speedometer (2) and a camera system (3) is arranged to take images of the vehicle in violation, so as to allow its identification.

The speedometer may for example consist of a Doppler radar, magnetic loops buried under the roadway or a laser system.

Means are provided for providing associated physical information, such as the time and date of the offense, the location of the device, etc.

First means are provided for recording and / or transmitting the data representative of the images taken by the camera systems, hereinafter referred to as the image data taken (30), and the data representative of the physical information, hereinafter called data. infringement (20), preferentially in the form of digital data.

In the example of the <figref idref="f0003">figure 6</figref>3-point memorization and / or transmission means of the diagram have been placed, but such means can obviously be provided at any other point of this diagram at naturally determined locations in a manner known per se by a person skilled in the art.

The storage means may also be made in any known manner using, for example, semiconductor memories, magnetic memories, etc.

The data transmission means can be of various kinds: transmission by wire, by communication bus or by radio.

Operating systems (4) are provided to exploit this data.

In particular in order to reduce the transmission durations of the infringement data and the image data taken and / or to carry out the various processes implemented in the method, it is advantageous to equip the operating systems with one or more processors (400) and first volatile and nonvolatile storage means and / or transmission means (430A).

In particular embodiments, the processors can be integrated into FPGA type semiconductor components or specialized ASIC semiconductor components.

The content of these first storage means can obviously be read and written by the processors of these operating systems.

In a manner known per se, it is possible to apply to the data collected data compression methods making it possible to reduce the size of the memories used for their storage or the capacity of the means of transmission of these data.

Among the data compression methods, it is possible to use so-called lossless methods of information or so-called entropic methods, with loss of information, particularly applied to data representative of images, and thus to obtain data. compressed.

We then choose a compression factor allowing unambiguous identification of the vehicles taken in violation when viewing these images.

The choice of the compression factor can be made for example during the installation of the camera systems, the operating systems recording in the memory means the data representative of images taken for different compression factors, and transmitting the image. content of these storage means to control units (5), used exceptionally during this installation phase.

These control units have display means for checking the quality of the reconstituted images from this compressed data.

Likewise, when necessary, for example in the case of poor lighting or weather conditions, the operating systems can apply any known processing to the image data taken in order to facilitate the identification of the infringing vehicles.

For example and in known manner, it is possible to enhance the contrast of these images and recognize the alphanumeric characters on the license plate.

After the possible application of these treatments, new data representative of images of the infringing vehicles called initial graphics data (10) are obtained which can be stored and / or transmitted.

Any known method of storage can be used, and in particular non-volatile methods.

Nonvolatile storage means are used to store the graphical representation of the alphanumeric characters from the information systems, these storage means being called nonvolatile memory of the fonts (410).

In a first exemplary embodiment, the non-volatile memory of the fonts contains the graphic representation of alphanumeric characters in the form of dot matrices.

In a second exemplary embodiment, the non-volatile memory of the fonts contains the graphical representation of the alphanumeric characters in the form of barcodes.

Infringement data and the graphical representation of alphanumeric characters are used to determine new data representative of images displaying the infringement data, these images being for example: 123 km / h-01/01 / 02- 10h: 30- Paris Aima as illustrated on the <figref idref="f0001">figure 2</figref>, these data being subsequently called graphical offense data (11).

In the particular case where a compression method has not been applied to the image data taken and where it is desired to be able to simultaneously view the image of the vehicle infringed and the images displaying the infringement data, a new image is produced. set of data, called graphical identifying data (12) by the following steps:<ul><li>the relative dimensions of the images of the data representative of the images incorporating the initial graphic data and the graphic infringement data are determined, this operation having also been possible during the installation of the system,</li><li>and the initial graphical data and the graphic infringement data are stored in a memory.</li></ul>

On the <figref idref="f0003">figure 6</figref>, the dashed lines in the storage and / or transmission means (430A) are a symbolic representation of the fact that the graphic infringement data are derived from the infringement data (20) and that the initial graphic data are derived from the data of pictures taken (30).

In this way, the graphic infringing data and the initial graphic data constitute two accessible subsets of the graphical identifying data.

On the particular example shown on the <figref idref="f0001">figure 2</figref>, the images representing the crime data are placed under the images taken by the camera systems.

It goes without saying that these images representing the infringement data could equivalently be, for example, placed above the images taken by the systems of shooting, or on the sides or in any other way.

It should be noted that when a data compression method has been used, the merging of the graphical identifying data and graphic infringement data is also possible.

In this case, it is for example possible to apply in an intermediate step the decompression method associated with the two sets of data, which allows to reduce to the previous case, then to apply again the compression process to obtain graphic identifying data.

A program code (40) required for application to a set of selected data, a non-bijective calculation method known from the prior art hereinafter called function f, is stored in a non-volatile memory.

For example, it is possible to use the calculation method described in the FIPS PUB 180-1 standardization document, published by the National Technical Information Service, US Department of Commerce, Springfield CA 22161.

The implementation of this method with the graphic identifying data leads to a new set of data hereinafter called summarized data (31).

In the case above, the summarized data are then sets of 160 bits of information.

It should be noted that with images whose definition allows the identification of infringing vehicles, ie with several tens of thousands of pixels, it is obviously impossible, from the summarized data of 160 bits, to reconstruct the graphical identifying data using a reverse computation method.

The summed data are then applied to a so-called public key / private key encryption method, as described, for example, in the patent <patcit id="pcit0007" dnum="US4405829A"><text>US 4,405,829</text></patcit>, which leads to new data, hereinafter referred to as coded summary data (32).

Again, the program code (41) required for the application of the encryption method in question is stored in a non-volatile memory.

The private key above is known only to authorized staff.

The private key can be permanently stored in the operating systems or preferably in volatile memories, which makes it possible to improve the security of the process.

In the latter case, this key can be downloaded from a highly secure database.

Thus, in case of theft of a device implementing the method, the private key remains inaccessible, even in case of analysis of the constituent elements of the device.

In a particular exemplary embodiment, the operating systems merge the coded summary data with another set of data, called alphabet data (420), for example by placing all the alphanumeric characters after the coded summary data. sufficient to represent the coded summary data.

For example, when the coded summary data is represented in a hexadecimal database, the operating systems place coded abstract data, alphanumeric characters 0 to 9 and A to F which in this case constitute the alphabet data.

It should be noted that in the particular embodiment of the above method, the alphabet data is placed after the coded summary data, but that such alphabet data could be placed in an equivalent manner before these coded summary data or in any other way to reconstruct the entire subset of the alphabet data.

The coded summary data, possibly merged with the alphabet data, is called signature data (33).

It should also be noted that merging the alphabet data with the coded summary data does not change the existence of the decoding method c^-1since the coded summary data is still an accessible subset of these signature data.

Is then determined from the signature data and the graphical representation of the alphanumeric characters, new data representative of images viewing the signature data, these images being for example: 13579BDF02468ACE1357

as illustrated on the <figref idref="f0002">figure 5</figref>, these new data being subsequently called graphical signature data (13).

The graphical representations in question may consist of dot matrices or bar codes, for example.

The graphical identifying data (12) is then merged with the graphical signature data (13), for example, according to the method already used to merge the graphical offense data (11) with the initial graphical data (10) and thus obtains the authenticatable graphics data (14).

On the particular example of the <figref idref="f0002">figure 5</figref>, the images representing the signature data are placed under the images representing the graphical identifying data. It goes without saying that these images representing the signature data could be equivalently, for example, placed above the images representing the graphic identifying data, or on the sides or in any other way.

The operating systems are furthermore provided with second storage and / or transmission means (430B) in order to allow the broadcast of the graphical authenticatable data.

In a first exemplary embodiment, the operating systems comprise these second nonvolatile storage means, in removable form, for example in the form of a memory card which can be removed from the system by an operator and placed in a control unit. (5).

Such a control unit may consist of a laptop or a much smaller housing able to read the contents of the card when associated with it.

In this first exemplary embodiment, these removable nonvolatile storage means can also be used as first storage means.

In a second exemplary embodiment, the operating systems comprise telecommunication means, for example, connected to a telephone line, so as to allow the transmission of the authenticatable graphics data to a control unit.

Such a control unit may consist of a computer equipped with a modem connected to a telephone line.

In this second exemplary embodiment, these transmission means can also be used as first transmission means for telecommunication with the camera systems.

In a third exemplary embodiment, the operating systems comprise wireless telecommunication means and are arranged to transmit the graphical authenticatable data to a control unit.

The control unit may consist of a computer comprising a radio modem.

In this third embodiment, the first transmission means may be of the same nature as those above.

The data actually read and / or received by the control units is called graphical received data (50).

The control units comprise one or more processors.

The control units can record in third memory means (530A), such as hard disks, the received graphic data in the form of computer files.

These storage means can also be used to contain all the algorithms and data necessary for finding the dimensions and positions of the graphic identifying data and the graphic signature data.

The totality of the received graphical data is then separated into two subsets of data called tested graphical identification data (51) and tested graphic signature data (52), these two subsets being respectively associated with the subsets of the graphical identifying data. and graphic signature data of all the data stored and / or transmitted by the operating systems.

In a first exemplary embodiment, the control units comprise in the non-volatile memory means a recognition program code (510A) of the characters which converts the tested graphic signature data into alphanumeric characters to form a new data set called data. signatures tested.

In a second exemplary embodiment and when the coded summary data has been merged as described above with the alphabet data (420), to form the signature data, the following steps are performed at the control units:<ul><li>we search the graphical signatures data tested in the received graphics data,</li><li>searching the graphical representations of the alphabet data in the graphic signature data, called graphical graphical data tested (510B),</li><li>the tested signature data (53) is calculated by comparing the tested graphical signature data with the tested graphical alphabet data.</li></ul>

The storage means may also contain the execution program code of the non-bijective calculation used above, defined by the function f, as well as the execution program code of the decoding program.^-1, known thanks to the public key associated with the private key that was used to encode the summarized data.

The function f is respectively applied to the graphical identifier data tested (51) which leads to new data called summarized data tested (55) and the decoding program c^-1 to the tested signature data which leads to new data, called summed data received (54).

The summarized data tested and the summary data received are compared.

When these two sets of data are identical, it is assumed that the images have been authenticated.

An alphanumeric validation message can be displayed on a screen directly readable by a human operator.

In contrast, when these two sets of data are not identical, a viewable signal can be provided on a screen or by any other method which indicates that image manipulation has been detected.

By using the method just described it is thus possible to verify that there has been no falsification of an image for example by assigning to a vehicle in violation of the identification characteristics of another vehicle .

The resulting image authentication solves the unresolved issues discussed above.

It goes without saying, and as it follows from the foregoing, the invention is not limited to the particular embodiment which has just been described, but is defined by the appended claims; it embraces the variant in which the method is implemented when the offense is other than that associated with the speeding exceedances authorized by a vehicle and for example when it relates to the detection of a person in a zone of protected access for which she does not have an authorization.