Electronic device and diagnostic method for an image sensor alarm(s), associated optronic system and computer program

Abstract

Electronic device and method for diagnosing an image sensor alarm, associated optronic system and computer program. The present invention relates to an electronic device (20) for diagnosing an image sensor alarm, mounted on board a vehicle (10) and comprising: - an acquisition module (30) – following the reception from a first sensor (18) of a first image representing an object that triggered an alarm – of a second image of said object, higher resolution than the first image, and from a second sensor (26); - a processing module (32) for applying a processing algorithm to the second image to search for a representation of said object, and if the representation is found, to zoom in towards said object, then a capture by the second sensor (26) of a third image including a representation of said object enlarged compared to that of the second image;- a module (34) for estimating the probability of a false alarm from the third image. Figure for the abbreviation: Figure 1;

Description

Title of the invention: Electronic device and method for diagnosing an image sensor alarm, associated optronic system and computer program

[0001] The present invention relates to an electronic device for diagnosing an image sensor alarm(s), intended to be installed on board a vehicle

[0002] The invention also relates to an optronic system intended to be mounted on board a vehicle, and comprising such a diagnostic device.

[0003] The invention also relates to a method for diagnosing a sensor alarm, the method being implemented by such a diagnostic device; as well as a computer program comprising software instructions which, when executed by a computer, implement such a diagnostic method.

[0004] It is known to equip a vehicle with a wide field of view sight in order to monitor an external environment around the vehicle, and where appropriate to warn a user on board the vehicle of the presence of a danger in this external environment.

[0005] When the wide-field viewfinder detects an object in its external environment, it emits an alarm associated with that object. Based on a wide-field viewfinder alarm selected by a user, a designation containing the wide-field viewfinder alarm information is typically sent by the user to a high-resolution viewfinder. The user of the high-resolution viewfinder must then determine, by reviewing the videos, whether it was a false alarm. If not, the user must then locate and identify the object that triggered the alarm.

[0006] However, all of these steps take between 10 and 20 seconds for each object selected and identified, and represent a significant cognitive load for the user.

[0007] The aim of the invention is then to propose an electronic device and a diagnostic method for an image sensor alarm(s) which make the diagnosis of such an alarm more efficient and reduce the cognitive load for the user.

[0008] To this end, the invention relates to an electronic device for diagnosing an image sensor alarm, intended to be installed on board a vehicle, the device comprising:

[0009] - an acquisition module configured to - following receipt, from a first image sensor(s), from a first image representing an object that triggered an alarm - acquire, from a viewfinder with a zoom and a second image sensor(s), a second image representing said object; the second sensor being distinct from the first sensor, the second image being more resolved than the first image;

[0010] - a processing module configured to apply a processing algorithm image to image to second image to search for a representation of said object in the second image, and if the representation of said object is found in the second image, to command a change in the viewfinder zoom to zoom in towards said object, then a capture of a third image by the second sensor; the third image including a representation of said object enlarged relative to that included in the second image;

[0011] the acquisition module being then configured to acquire the third image from the second sensor;

[0012] - an estimation module configured to estimate a probability of false alarm From the third image onwards, the probability of a false alarm is further estimated to be equal to a maximum value if the representation of said object has not been found in the second image.

[0013] With the diagnostic device according to the invention, the acquisition of the second image, which is more resolved than the first image that triggered the alarm, and then the processing applied to this second image to control the capture of a third image including a representation of the object enlarged compared to that included in the second image, the third image being in other words zoomed in compared to the second image, then makes it possible to estimate more effectively the probability of a false alarm from this third image.

[0014] By second image more resolved than first image, we mean that the second image has a higher angular resolution than that of the first image, that is to say that the elementary field of view per pixel, or IFOV (from the English Instantaneous Field Of View), is smaller on the second image than on the first image, or that the number of pixels for an object in the second image is greater than the number of pixels for the same object in the first image.

[0015] Advantageously, the viewfinder zoom is an optical zoom, and the third image has a higher angular resolution than the second image.

[0016] According to other advantageous aspects of the invention, the electronic diagnostic device comprises one or more of the following features, taken individually or in all technically possible combinations:

[0017] - the device further includes a calculation module configured to calculate a estimated distance between the vehicle and said object;

[0018] the calculation module being preferably configured to calculate said estimated distance from the first image and the third image; preferably again via triangulation from the first image and the third image;

[0019] - the device further includes an identification module configured to identify a set of object characteristics by applying a second image processing algorithm to the third image;

[0020] the set of feature(s) preferably comprising at least one dimensional feature, one shape feature and / or one object type feature;

[0021] - the device further comprises a prediction module configured to predict, in based on the probability of a false alarm, a level of danger associated with said object;

[0022] - the prediction module is configured to predict said danger level in also depending on the estimated distance between the vehicle and said object;

[0023] - the prediction module is configured to predict said danger level in function in addition to the set of characteristic(s); and

[0024] - the estimation module is configured to estimate the probability of a false alarm via an analysis of the elements present in the third image and / or a comparison of the third image with one or more reference images from a database.

[0025] The invention also relates to an optronic system intended to be mounted on board a vehicle and connected to a first image sensor(s), the optronic system comprising an electronic device for diagnosing an alarm from the first image sensor(s) and a viewfinder comprising a zoom and a second image sensor(s), the second sensor being distinct from the first sensor, the diagnostic device being as defined above, and connected to the viewfinder.

[0026] The invention also relates to a method for diagnosing an image sensor alarm(s), the method being implemented by an electronic diagnostic device intended to be installed on board a vehicle, and comprising the following steps:

[0027] - following the reception, by a first image sensor(s), of a first image representing an object that triggered an alarm - to acquire, from a viewfinder comprising a zoom and a second image sensor(s), a second image representing said object; the second sensor being distinct from the first sensor, the second image being more resolved than the first image;

[0028] - process the second image by applying an image processing algorithm to the second image to search for a representation of said object in the second image, and if the representation of said object is found in the second image:

[0029] + command a change in the viewfinder zoom to zoom in the direction of said object, then a capture of a third image by the second sensor; the third image including a representation of said object enlarged compared to that included in the second image;

[0030] + acquire the third image from the second sensor;

[0031] - to estimate a probability of false alarm from the third image, the the probability of a false alarm is further estimated to be equal to a maximum value if the representation of said object has not been found in the second image.

[0032] The invention also relates to a computer program comprising software instructions which, when executed by a computer, implement a diagnostic method as defined above.

[0033] The invention will become clearer upon reading the following description, given solely by way of non-limiting example, and made with reference to the drawings in which:

[0034] [Fig-1] [Fig.1] is a schematic representation of a first viewfinder comprising a first image sensor(s) and an optronic system according to the invention connected to the first viewfinder, the first viewfinder and the optronic system being intended to be mounted on board a vehicle, the optronic system comprising an electronic device for diagnosing an alarm from the first image sensor(s) and a second viewfinder comprising a zoom and a second image sensor(s); the first sensor being configured to take a first image representing an object, and the second sensor is then configured to take a second, or even a third image representing said object;

[0035] [Fig.2] [Fig.2] is a flowchart of a diagnostic method according to the invention. from an alarm originating from the first sensor(s), the diagnostic process being implemented by the diagnostic device of [Fig.1];

[0036] [Fig.3] [Fig.3] is a view of the first and second images according to a first example corresponding to a false alarm;

[0037] [Fig.4] [Fig.4] is a view of the first, second and third images according to a second example corresponding to a low danger; and

[0038] [Fig. 5] [Fig. 5] is a view analogous to that of [Fig. 4] according to a third example corresponding to a high level of danger.

[0039] In the following description, the expression "approximately equal to" defines a relationship of equality to plus or minus 20%, preferably to plus or minus 10%, and preferably still to plus or minus 5%.

[0040] In [Fig. 1], a vehicle 10 comprises a first sight 12 and an optronic system 15. The vehicle 10 is, for example, a maritime vehicle, such as a ship; a railway vehicle; a motor vehicle; or an aeronautical vehicle, such as an aircraft. When the vehicle 10 is a ship, the first sight 12 is advantageously installed on a mast, for example, at the top of the mast.

[0041] The first viewfinder 12 comprises a first angle sensor(s) 18. The first viewfinder 12 is typically a wide-field viewfinder. The first sensor 18 has a low angular resolution, typically an angular resolution of at most approximately 1 milliradian, and advantageously substantially between 0.1 and 1 milliradian. The first sensor 18 is for example an infrared type sensor, that is to say capable of capturing one or more infrared images, i.e. in the infrared range, of a scene towards which it is directed.

[0042] By viewfinder, we mean an electronic viewfinder or EVF (Electronic Viewfinder), that is to say, a device for viewing a scene or an image using electronic and optical components. Unlike traditional optical viewfinders, which use systems of mirrors and prisms to direct light towards the eye of an observer, electronic viewfinders capture the image using an electronic sensor, such as a CCD (Charge-Coupled Device) or CMOS (Complementary Metal-Oxide-Semiconductor) sensor, and display it on an electronic display screen, such as an LCD (Liquid Crystal Display) or OLED (Organic Light Emitting Diode) screen.

[0043] A wide-field viewfinder is defined as a viewfinder having a field of view (FOV) of at least approximately 60 degrees, advantageously between 60 and 360 degrees, and preferably between 60 and 120 degrees. A wide-field viewfinder generally exhibits low angular resolution due to this large field of view.

[0044] Angular resolution refers to the angular range covered by a pixel of the electronic sensor. The lower this angular resolution, the better the representation of an object in an image.

[0045] The optronic system 15 is mounted on board the vehicle 10 and is connected to the first sensor 18, as in the example of [Fig. 1]. The optronic system 15 includes an electronic device 20 for diagnosing an alarm from the first sensor 18 and a viewfinder, also called a second viewfinder 22, comprising a zoom 24 and at least one second image sensor(s) 26, the second sensor(s) 26 being separate from the first sensor 18 and advantageously having a higher angular resolution than the first sensor 18.

[0046] Those skilled in the art will observe that the choice to present the invention with two sensors 18, 26 is made to facilitate its understanding, and that the number of sensors is not limited to two. Alternatively, the number of sensors is strictly greater than two. The number of sensors depends, for example, on the vehicle 10 on which the invention is deployed.

[0047] The diagnostic device 20 comprises an acquisition module 30, a processing module 32 and an estimation module 34. The diagnostic device 20 is connected on one side to the first sight 12, and on the other side to the second sight 22.

[0048] As an optional addition, the diagnostic device 20 further includes a calculation module 36. As another optional addition, the diagnostic device 20 It also includes an identification module 38. As an optional addition, the diagnostic device also includes a prediction module 40.

[0049] In the example of [Fig.1], the diagnostic device 20 includes an information processing unit 50 formed for example of a memory 52 and a processor 54 associated with the memory 52.

[0050] In the example of [Fig. 1], the acquisition module 30, the processing module 32, and the estimation module 34, as well as the optional calculation module 36, identification module 38, and prediction module 40, are each implemented as software, or a software component, executable by the processor 54. The memory 52 of the diagnostic electronic device 20 is thus capable of storing acquisition software, processing software, and estimation software, as well as the optional calculation software, identification software, and prediction software. The processor 54 is then capable of executing each of the following software components: acquisition software, processing software, and estimation software, as well as the optional calculation software, identification software, and prediction software.

[0051] In an alternative not shown, the acquisition module 30, the processing module 32 and the estimation module 34, as well as optionally the calculation module 36, identification module 38 and prediction module 40, are each implemented as a programmable logic component, such as an FPGA (Field Programmable Gate Array) or as a dedicated integrated circuit, such as an ASIC (Application Specified Integrated Circuit).

[0052] When the electronic diagnostic device 20 is implemented in the form of one or more software programs, i.e., in the form of a computer program, it is also capable of being stored on a computer-readable medium (not shown). A computer-readable medium is, for example, a medium capable of storing electronic instructions and being connected to a bus of a computer system. For example, a readable medium is an optical disc, a magneto-optical disc, a ROM, a RAM, any type of non-volatile memory (e.g., EPROM, EEPROM, FLASH, NVRAM), a magnetic card, or an optical card. A computer program comprising software instructions is then stored on the readable medium.

[0053] The second viewfinder 22 is advantageously a high-resolution viewfinder. The second sensor 26 is then a sensor having a high resolution, typically an angular resolution of at most approximately 100 microradians, advantageously of at most approximately 50 microradians, and preferably of at most approximately 20 microradians, or even of at most approximately 15 microradians.

[0054] By high-resolution viewfinder, we mean a viewfinder with an angular resolution greater than that of a wide-field viewfinder.

[0055] The second sight 22 is typically a narrow-field sight, or reduced-field sight, with a smaller field of view than the wide-field sight. The second sight 22 advantageously has a field of view, or FOV, of less than 20 degrees, preferably less than 10 degrees, preferably even less than 5 degrees, and for example substantially equal to 2 degrees.

[0056] When the vehicle 10 is a ship, the second sight 22 is typically installed on the mast, for example at the top of the mast.

[0057] In the example of [Fig. 1], the second viewfinder 22 comprises a single second sensor 26. As an optional addition, not shown, the second viewfinder 22 comprises several second sensors 26, in particular two second sensors 26 of different types. According to this optional addition, the second sensors 26 are each distinct from the first sensor 18.

[0058] The 24 zoom is typically an optical zoom. Alternatively, the 24 zoom is a digital zoom.

[0059] The second sensor 26 is, for example, a visible-type sensor, that is, capable of capturing one or more images in the visible range of a scene towards which it is directed. Alternatively, the second sensor 26 is an infrared-type sensor.

[0060] When an optional complement is used, the second viewfinder 22 includes several second sensors 26, one of which is for example infrared type and the other visible type.

[0061] The acquisition module 30 is configured to – following the reception, from the first sensor 18, of a first image II representing an object that has triggered an alarm – acquire, from the second viewfinder 22, and in particular from the corresponding second sensor 26, a second image 12 representing said object. The second image 12 is higher resolution than the first image II.

[0062] The processing module 32 is configured to apply a first image processing algorithm to the second image 12 to search for a representation of said object in the second image 12.

[0063] The first image processing algorithm is for example configured to search for the representation of said object in the second image 12 by comparing the second image 12 with the first image II.

[0064] The first image processing algorithm advantageously includes an artificial intelligence model. According to this advantageous aspect, the artificial intelligence model is pre-trained, for example via supervised learning, with training data comprising, as input to the model, pairs of first image II and second image 12; and, as output, for each of said pairs, the second image 12 enriched with a frame around the representation of the object when this representation is present in the second image 12. This learning is also carried out with examples where the representation of the object is not present in the second image 12, and the second image 12 output of the model then does not have a frame.

[0065] In addition or alternatively, the first image processing algorithm is configured to search for the representation of said object via a search for points of interest in the second image 12, such as points of interest, for example related to the shape of the object.

[0066] The processing module 32 is then configured, if the representation of said object is found in the second image 12, to command a modification of the zoom 24 of the second viewfinder 22 to zoom in towards said object, then to command a capture of a third image 13 by the second sensor 26. The third image 13 includes a representation of said object enlarged compared to that included in the second image 12, by this modification of the zoom 24 in the direction of the object.

[0067] The acquisition module 30 is then configured to acquire the third image 13 from the second sensor 26.

[0068] Advantageously, when the zoom 24 of the second viewfinder 22 is an optical zoom, and the third image 13 has a higher angular resolution than the second image 12.

[0069] When the zoom 24 of the second viewfinder 22 is a digital zoom, the third image 13 has a lower resolution than the second image 12.

[0070] The estimation module 34 is configured to estimate a false alarm probability based on the third image 13. If the object representation has not been found in the second image 12, the estimation module 34 is further configured to estimate the false alarm probability to a maximum value. In other words, in the absence of a third image 13, the estimation module 34 is configured to estimate the false alarm probability to the maximum value.

[0071] Advantageously, and in the presence of the third image 13, the estimation module 34 is configured to estimate the probability of a false alarm by analyzing the elements present in the third image 13 and / or comparing the third image 13 with one or more reference images from a database, not shown. The reference images contained in the database are typically images with a resolution substantially equal to that of the third image 13, and include both representations of objects associated with false alarms and representations of objects associated with true alarms.

[0072] The calculation module 36 is configured to calculate an estimated distance between the vehicle 10 and said object.

[0073] Advantageously, the calculation module 36 is configured to calculate said estimated distance from the first image II and the third image 13; for example via triangulation from the first image II and the third image 13.

[0074] For said triangulation, a person skilled in the art will observe on the one hand that the first sight 12 and the second sight 22 are each mounted on board the vehicle 10, so that the position of the first 12 and second 22 sights is identical to the position of the vehicle 10; and on the other hand that the first sight 12 and the second sight 22 are oriented in distinct respective orientations from each other, so that the first image II and the third image 13 correspond to different directions with respect to the object.

[0075] The identification module 38 is configured to identify a set of object characteristic(s) by applying a second image processing algorithm to the third image 13.

[0076] The second image processing algorithm is configured to identify said set of feature(s) for example by comparing the third image 13 with reference images from the database

[0077] Advantageously, the feature set comprises at least one dimensional feature, one shape feature, and / or one object type feature. According to this advantageous aspect, the feature set then comprises several features associated with the object.

[0078] The at least one dimensional characteristic includes, for example, one or more dimensions of the object in different directions, such as the dimensions of the object in the three directions of a reference frame, a ratio of dimensions in two distinct directions, a volume of the object and / or an area of ​​the object in a reference plane substantially parallel to the plane of the third image 13.

[0079] The object type characteristic is, for example, chosen from a predefined list of object types. The predefined list of object types depends, for example, on vehicle 10.

[0080] As an optional complement, the identification module 38 is configured to identify the dimensional characteristic(s) of the object based also on the distance between the vehicle 10 and said object estimated by the calculation module 36.

[0081] The prediction module 40 is configured to predict, based on the probability of a false alarm, a danger level associated with said object. The predicted danger level is typically higher when the probability of a false alarm is lower. The predicted danger level is, for example, inversely proportional to the probability of a false alarm. In particular, the predicted danger level is zero, or minimal, if the estimated probability of a false alarm is equal to the maximum value.

[0082] As an optional addition, the prediction module 40 is configured to predict said level of danger based also on the estimated distance between the vehicle 10 and said object.

[0083] As an optional addition, the prediction module 40 is configured to predict said level of danger based further on the set of characteristic(s).

[0084] Advantageously, the prediction module 40 is configured to predict the level of danger associated with said object based on both the probability of false alarm, the estimated distance between the vehicle 10 and said object, and the set of characteristic(s) identified for said object.

[0085] Advantageously still, the prediction module 40 is configured to predict the level of danger associated with said object according to a set of prediction rule(s).

[0086] According to this advantageous aspect, the set of prediction rules includes, for example, a so-called distance rule according to which the smaller the estimated distance between the vehicle 10 and the object, the higher the predicted level of danger. In addition, or alternatively, the set of prediction rules includes an identification rule based on the set of characteristics identified for the object, in particular the shape of the object and / or a possible object identifier. In addition, or alternatively, the set of prediction rules includes a so-called location rule based on the position of the object relative to the vehicle 10, the predicted level of danger being higher if the object is located in an area considered dangerous.In addition, or as an alternative, the set of prediction rules includes a so-called displacement rule based on the object's movement relative to vehicle 10. The predicted danger level is higher if the object moves closer to vehicle 10, and conversely, lower if the object moves away from vehicle 10. For this displacement rule, the speed of movement is also advantageously taken into account, as the variation in the predicted danger level is greater with higher speeds. In other words, when the object moves closer to vehicle 10, the danger level will be higher the faster the object moves; and conversely, when the object moves away from vehicle 10, the danger level will be lower the faster the object moves away from vehicle 10.

[0087] According to this advantageous aspect, the set of prediction rules includes at least one rule from among the distance rule, the identification rule, the location rule, and the displacement rule. In addition, the set of prediction rules includes at least two rules, preferably at least three rules, and preferably all of the rules, from among the distance rule, the identification rule, the location rule, and the displacement rule.

[0088] The operation of the optronic system 15 according to the invention, and in particular of the diagnostic device 20, will now be explained with the help of [Fig.2] representing, a flowchart of the diagnostic process according to the invention implemented by the diagnostic device 20.

[0089] During an initial step 100, the diagnostic device 20 acquires, following the reception, from the first sensor 18, of a first respective image II representing an object which has triggered an alarm, and via its acquisition module 30, a second respective image 12 from the second sensor 26.

[0090] As previously stated, more generally, the vehicle 10 according to the invention comprises from two to N separate sensors, where N is an integer strictly greater than two, enabling the generation of two to N respective images. The acquisition module 30 is then configured to acquire each image generated by the respective sensors 18, 26.

[0091] The diagnostic device 20 then proceeds to the next step 110 in which it processes, via its processing module 32, the second image 12 acquired during the acquisition step 100, in order to search for a representation of said object in the second image 12. This search for said representation is carried out by applying the first image processing algorithm to the second image 12.

[0092] The diagnostic device 20 then proceeds to a test step 120 in which it tests, via its processing module 32, whether or not the representation of the object has been found in the second image 12.

[0093] If the test step 120 is negative, i.e. if the representation of the object was not found in the second image 12 during the processing step 110, then the diagnostic device 20 proceeds to the step 130 during which the probability of false alarm is estimated to be equal to the maximum value by the estimation module 34.

[0094] If the test step 120 is positive, i.e. if the representation of the object has been found in the second image 12 during the processing step 110, then the diagnostic device 20 proceeds to step 140 in which it first commands, via its processing module 32, a modification of the zoom 24 of the second viewfinder 22 to zoom in towards said object, then a capture of a third respective image 13 by the second sensor 26. During step 140, following this command, the diagnostic device 20 then acquires, via its acquisition module 30, the third image 13 captured by the second sensor 26.

[0095] At the end of this command and acquisition step 140, the diagnostic device 20 moves to the next step 150 during which it estimates, via its estimation module 34 and from the third image 13 acquired during step 140, the probability of false alarm associated with said object.

[0096] During this estimation step 150, the probability of false alarm is advantageously estimated via the analysis of the elements present in the third image 13 and / or the comparison of the third image 13 acquired during step 140 with the reference images from the database.

[0097] The diagnostic device 20 then proceeds to step 160, during which it calculates, via its calculation module 36, the estimated distance between the vehicle 10 and said object. During this calculation step 160, the distance is advantageously estimated from the first image II and the third image 13, typically by triangulation from the first image II and the third image 13.

[0098] The diagnostic device 20 then identifies, in the next step 170 and via its identification module 38, the set of characteristic(s) associated with said object by applying the second image processing algorithm to the third image 13 acquired during step 140.

[0099] During this identification step 170, said set of characteristic(s) is advantageously identified further according to the estimated distance between the vehicle 10 and said object calculated during the calculation step 160.

[0100] The diagnostic device 20 finally moves to step 180 during which it predicts, via its prediction module 40 and in particular according to the probability of false alarm estimated during step 150, the level of danger associated with said object.

[0101] During this prediction step 180, the danger level is optionally predicted based further on the estimated distance between the vehicle 10 and said object calculated during step 160 and / or the set of characteristic(s) identified during step 170.

[0102] During this prediction step 180, the level of danger is advantageously predicted according to one or more rules from the set of prediction rule(s).

[0103] At the end of the prediction step 180, the diagnostic device 20 returns to the initial step 100 to be ready to acquire a new second image 12 following the reception of a new first image II representing an object which has triggered a new alarm.

[0104] Figures 3 to 5 then illustrate different examples of operation of the optronic system 15 according to the invention, and in particular of the diagnostic device 20, with different objects.

[0105] Figures 3 to 5 illustrate the operation of the optronic system 15 according to the invention, in the case where the vehicle 10 is a marine vehicle, such as a ship. Those skilled in the art will understand that the operation of the optronic system 15 according to the invention is similar, with examples of operation of the same type, when the vehicle 10 is a motor vehicle, a railway vehicle, or an aircraft.

[0106] Figure 3 illustrates an example of a false alarm where only the first and second images were acquired, the second image not containing any representation of the object that triggered the alarm. In the example in Figure 3, the probability of a false alarm is therefore estimated to be equal to the maximum value, and the danger level predicted by the prediction module is a minimum level, this false alarm being due to a solar reflection on the water.

[0107] Figure 4 illustrates an example of a true alarm with successive acquisitions of the first 12, second 12 and third 13 images. In the example of Figure 4, the probability of a false alarm is then estimated to be low, and the level of danger predicted by the prediction module 40 is also low, the object concerned being a sailboat not presenting any particular danger, as is clearer in the third image 13 of Figure 4.

[0108] Figure 5 also illustrates an example of a true alarm with successive acquisitions of the first 12, second 12 and third 13 images. In the example of Figure 5, the probability of a false alarm is also estimated to be low, and the level of danger predicted by the prediction module 40 is this time a high level, the object concerned being a ferry presenting a significant risk of collision for the vehicle 10, as is more clearly shown in the third image 13 of Figure 5.

[0109] It is thus understood that the electronic diagnostic device 20 and the diagnostic method according to the invention make it possible to make the diagnosis of an alarm from the first sight 12 more efficient and to significantly reduce the cognitive load for the user when diagnosing the alarm, in particular to determine whether this alarm represents a danger or not for the vehicle 10.

Claims

Demands

1. Electronic device (20) for diagnosing an alarm from an image sensor(s) (18), intended to be mounted on board a vehicle (10), the device (20) comprising: - an acquisition module (30) configured to - following the receipt, from a first image sensor(s) (18), of a first image (II) representing an object which triggered an alarm - acquire, from a viewfinder (22) comprising a zoom (24) and a second image sensor(s) (26), a second image (12) representing said object; the second sensor (26) being distinct from the first sensor (18), the second image (12) being higher resolution than the first image (II);- a processing module (32) configured to apply an image processing algorithm to the second image (12) to search for a representation of said object in the second image (12), and if the representation of said object is found in the second image (12), to command a modification of the zoom (24) of the viewfinder (22) to zoom in towards said object, then a capture of a third image (13) by the second sensor (26); the third image (13) including a representation of said object enlarged compared to that included in the second image (12); the acquisition module (30) then being configured to acquire the third image (13) from the second sensor (26); - an estimation module (34) configured to estimate a probability of false alarm from the third image (13), the probability of false alarm being further estimated to be equal to a maximum value if the representation of said object has not been found in the second image (12).

2. Device (20) according to claim 1, wherein the device (20) further comprises a calculation module (36) configured to calculate an estimated distance between the vehicle (10) and said object; the calculation module (36) preferably being configured to calculate said estimated distance from the first image (II) and the third image (13); preferably also via triangulation from the first image (II) and the third image (13).

3. Device (20) according to any one of the preceding claims, wherein the device (20) further comprises a identification module (38) configured to identify a set of feature(s) of the object by applying a second image processing algorithm to the third image (13); the set of feature(s) preferably comprising at least one dimensional feature, one shape feature and / or one object type feature.

4. Device (20) according to any one of the preceding claims, wherein the device (20) further comprises a prediction module (40) configured to predict, based on the probability of a false alarm, a level of danger associated with said object.

5. Device (20) according to any one of the preceding claims taken together with claims 2 and 4, wherein the prediction module (40) is configured to predict said level of danger as a further function of the estimated distance between the vehicle (10) and said object.

6. Device (20) according to any one of the preceding claims taken together with claims 3 and 4, wherein the prediction module (40) is configured to predict said level of danger as a further function of the set of feature(s).

7. Device (20) according to any one of the preceding claims, wherein the estimation module (34) is configured to estimate the probability of false alarm via an analysis of the elements present in the third image (13) and / or a comparison of the third image (13) with one or more reference images from a database.

8. Optronic system (15) intended to be mounted on board a vehicle (10) and connected to a first image sensor(s) (18), the optronic system (15) comprising an electronic device (20) for diagnosing an alarm from the first sensor (18) and a viewfinder (22) comprising a zoom (24) and a second image sensor(s) (26), the second sensor (26) being distinct from the first sensor (18), characterized in that the diagnostic device (20) is according to any one of the preceding claims, and connected to the viewfinder (22).

9. A method for diagnosing an image sensor alarm(s), the method being implemented by an electronic diagnostic device (20) intended to be installed on board a vehicle (10), and comprising the following steps: - following the reception, from a first image sensor (18), of a first image (II) representing an object that triggered an alarm - acquire (100), from a viewfinder (22) comprising a zoom (24) and a second image sensor (26), a second image (12) representing said object; the second sensor (26) being distinct from the first sensor (18), the second image (12) being higher resolution than the first image (II); - process (110) the second image (12) by applying an image processing algorithm to the second image (12) to search for a representation of said object in the second image (12), and if the representation of said object is found in the second image (12): + command (140) a modification of the zoom (24) of the viewfinder (22) to zoom in towards said object, then a capture of a third image (13) by the second sensor (26);the third image (13) including a representation of said object enlarged compared to that included in the second image (12); + acquire (140) the third image (13) from the second sensor (26); - estimate (150) a probability of false alarm from the third image (13), the probability of false alarm being further estimated to be equal to a maximum value if the representation of said object was not found in the second image (12).;

10. A computer program comprising software instructions which, when executed by a computer, implement a method according to the preceding claim.

Images