Decision support method for the approach phase of an aircraft landing on a runway

Abstract

Decision support method for approach phase for landing an aircraft on a runway. The present invention relates to a decision support method (40) for the approach phase for landing an aircraft on a runway, the method (40) comprising the following steps: - obtaining (42) characteristic elements detected via an image capture device, - determining (44) an approach scene recognition score associated with said detected characteristic elements, based on a reference approach scene; - comparing (46) said score to a predetermined decision threshold value; - decision support (48) by providing the comparison result to a pilot, or to an autopilot system, capable of continuing the approach phase if the score is higher than said predetermined decision threshold value, or of interrupting the approach phase, by going around, if the score is lower.Figure for the abbreviation: Figure 4.

Description

Title of the invention: Decision support method for the approach phase of an aircraft landing on a runway

[0001] The present invention relates to a decision support method in the approach phase for landing an aircraft on a runway.

[0002] The present invention also relates to an electronic decision support device for the approach phase for landing an aircraft on a runway, as well as an aircraft comprising such a device.

[0003] In the field of aeronautics, and more specifically in aeronautical systems engineering combined with logic and statistical techniques, it is known to equip aircraft with approach and landing assistance devices to assist a human pilot or even an autopilot. Such devices are designed, in particular, to determine the aircraft's position relative to the runway.

[0004] On approach for landing of an aircraft, when an aircraft pilot (human or even automatic) reaches a "decision height" known via the aeronautical acronym DH (from the English Decision Height) or DA (from the English Decision Altitude), he must decide whether or not to continue his approach with a go-around in the second case (i.e., interrupt the approach).

[0005] The approach scene is defined in particular in the document of the International Civil Aviation Organization ICAO, Annex 14, Volume I, in paragraphs 3.1, 4.1, 5.2 and 5.3.

[0006] Normative texts, such as:

[0007] - the electronic code of federal regulations e-CFR (from the English electronic Code of Federal Regulations) and in particular paragraph 91.176,

[0008] - document DO-341, paragraphs 3.1.3.1 of the Radio Technical Commission for Aeronautics RTCA (from the English Radio Technical Commission for Aeronautics),

[0009] - the Federal Aviation Administration document AC90-106A (from English federal aviation administration) and in particular Table 1,

[0010] - the AMC1 CAT.OP.MPA.305(c) document of the civil aviation authority of The United Kingdom sets minimum conditions to be met to allow an approach to proceed, these conditions being intended for a human pilot.

[0011] However, and particularly in low visibility conditions, even with a visual aid such as that provided by an enhanced flight vision system EFVS (of (English Enhanced Flight Vision System), it is not always easy to appreciate the conditions to be met as indicated in particular within the specific training and qualifications authorizing such approaches.

[0012] There is therefore a need for a decision support process in the approach phase for landing of an aircraft on a runway which makes it possible to reduce the mental load of human pilots and / or to open a new path of autopilot in the approach and landing phase, and / or to reduce the number of pilots at the controls of an aircraft, in particular for special operations SPO (from the English Specialised Operations).

[0013] To this end, the invention relates to a decision-support method for the approach phase for landing of an aircraft on a runway, the runway having a longitudinal axis equidistant from the longitudinal edges of the runway, called the runway axis, the runway comprising a set of characteristic elements, the aircraft comprising at least one image capture device configured to detect, during the approach for landing, within at least one image, at least one characteristic element of the runway, each detection being associated with coordinates,

[0014] the method being implemented by an electronic decision support device during the approach phase for landing, and comprising the following steps:

[0015] - obtaining said characteristic elements detected via said at least one device image capture,

[0016] - determination of an approach scene recognition score associated with said characteristic elements detected, based on a reference approach scene associated with said landing runway;

[0017] - comparison of said approach scene recognition score to a value of predetermined decision threshold corresponding to the minimum value to be respected to authorize the continuation of the approach for landing;

[0018] - decision support by providing the result of the comparison to a pilot, or to a autopilot system, capable of continuing the approach phase in case of a score greater than said predetermined decision threshold value, or of interrupting the approach phase, by going around, in case of a score less than said predetermined decision threshold value.

[0019] Thus, the present invention proposes to assist the piloting task by determining and providing an approach scene recognition score whose value is compared to a decision value, so that the result of this comparison serves the pilot (human or automatic) to decide whether or not to interrupt the approach phase by going around.

[0020] In other words, the present invention proposes an automatic recognition processing (without intervention of the human pilot or the autopilot) of an approach scene by providing via said score a quantified information for the recognition of an approach scene for landing which makes it possible in parallel to lighten the mental load of a human pilot or to open a new path of autopilot.

[0021] According to other advantageous aspects of the invention, the decision support method for the approach phase for landing an aircraft on a runway comprises one or more of the following features, taken individually or in all technically possible combinations:

[0022] - said determination of an approach scene recognition score includes:

[0023] - based on said detected characteristic elements, a classification of information associated with the aforementioned detected characteristic elements, within a set of distinct predetermined elementary categories;

[0024] - for each elementary category, a determination of an elementary score associated with the information classified in said category;

[0025] - a combination, according to a predetermined logic, of said elementary scores to form said approach scene recognition score;

[0026] - said set of distinct predetermined elementary categories comprises at least three distinct subsets of distinct elementary categories relating respectively to:

[0027] - elements relating to approach lights and runway lights;

[0028] - runway-related elements, distinct from approach light-related elements and runway lights;

[0029] - other specific elements of the approach scene and distinct from the elements relating to approach lights, runway lights and relating to the runway;

[0030] - each elementary score is:

[0031] - standardized according to the information corresponding to said elementary category and associated with said reference approach scene, and / or

[0032] - multiplied by one when said information is associated with said elements Detected characteristics satisfy a predetermined condition and are zero otherwise; and / or

[0033] - weighted according to a predetermined weighting;

[0034] - said predetermined combination logic is capable of varying according to the coordinates associated with each detection, and / or depending on the presence of runway lights, approach lights on said landing runway;

[0035] - said predetermined logic is representative of at least one pattern of elements reference characteristics of said runway;

[0036] - said decision threshold value is predetermined according to of a degree of resemblance of said reference approach scene to a set of elements surrounding said landing runway.

[0037] The invention also relates to a computer program comprising software instructions which, when executed by a computer, implement a decision support method in the approach phase for landing of an aircraft on a runway as defined above.

[0038] The invention also relates to an electronic decision-support device for the approach phase of an aircraft landing on a runway, the runway having a longitudinal axis equidistant from the longitudinal edges of the runway, called the runway axis, the runway comprising a set of characteristic elements, the aircraft comprising at least one image capture device configured to detect, during the approach for landing, within at least one image, at least one characteristic element of the runway, each detection being associated with coordinates,

[0039] the electronic decision support device for the approach phase for landing comprising:

[0040] - a retrieval module configured to obtain said characteristic elements detected via said at least one image capture device,

[0041] - a determination module configured to determine a recognition score of approach scene associated with said detected characteristic elements, according to a reference approach scene associated with said landing runway;

[0042] - a comparison module configured to compare said recognition score of approach scene to a predetermined decision threshold value corresponding to the minimum value to be respected to authorize the continuation of the approach for landing;

[0043] - a decision support module configured to assist in decision-making by feedback of the result of the comparison to a pilot, or to an autopilot system, capable of continuing the approach phase in the event of a score higher than said predetermined decision threshold value, or of interrupting the approach phase by going around in the event of a score lower than said predetermined decision threshold value.

[0044] The invention also relates to an aircraft comprising an electronic decision support device for the approach phase for landing an aircraft on a runway as described above.

[0045] Other features and advantages of the invention will become apparent from the following description of embodiments of the invention, given by way of example only, and with reference to the drawings which are:

[0046] - [Fig.1] [Fig.1], a schematic representation of an aircraft approaching a runway landing,

[0047] - [Fig.2] [Fig.2], a schematic top-view representation of an example of runway and the horizontal projection of an aircraft's radar in the horizontal plane of the runway,

[0048] - [Fig.3] [Fig.3], a schematic representation of an example device Electronic decision support system for the approach phase of an aircraft landing on a runway,

[0049] - [Fig.4] [Fig.4], a flowchart of an example of a decision support process in the approach phase for landing of an aircraft on a runway.

[0050] A runway 10 and an aircraft 12 in flight are illustrated in [Fig. 1]. As can be seen in this [Fig. 1], the aircraft 12 is approaching the runway 10 in preparation for landing on this runway 10.

[0051] Aircraft 12 is, for example, an airplane or a helicopter.

[0052] As seen in [Fig.1], the aircraft 12 includes an electronic device 14 for decision support during the approach phase for landing of an aircraft on a runway, and at least one image capture device 15 configured to detect, during the approach for landing, within at least one image, at least one characteristic element of the runway 10, each detection being associated with coordinates.

[0053] Said at least one image capture device 15 is an element belonging to the group comprising:

[0054] - a radar, and / or

[0055] - a camera measuring visible or infrared radiation.

[0056] Preferably, said at least one image capture device 15 is a radar whose coordinate system is represented in [Fig. 1] by a Cartesian coordinate system with center OR, abscissa XR, ordinate YR, and elevation ZR. The ordinate axis YR is the longitudinal axis of detection and is called the radar axis. When the radar 15 is correctly positioned on the aircraft 12, the radar axis YR coincides with the longitudinal axis of the aircraft 12.

[0057] The radar 15 is advantageously a millimeter wave radar. Preferably, the radar's range resolution is on the order of a few meters, and the radar's angular resolution is fine, that is to say, on the order of a few tenths of a degree.

[0058] According to an unshown variant, the electronic decision support device 14 for the approach phase of an aircraft landing on a runway itself comprises at least one image capture module 15. In other words, according to this unshown variant, the electronic decision support device 14 for the approach phase of landing is also configured to implement itself- even image capture by integrating (within its casing) said image capture module 15.

[0059] The runway 10 is a rectangular surface intended for the landing and takeoff of aircraft. The runway 10 comprises longitudinal edges (two) and transverse ends (two) delimiting the runway 10.

[0060] The length of runway 10 is, for example, between 1200 and 2400 m. The width of runway 10 is, for example, between 30 and 45 m. According to another example, the length of runway 10 is between 800 and 5400 m, and its width between 18 and 60 m.

[0061] As can be seen in [Fig.2], the runway 10 has a longitudinal axis in the longitudinal direction of the runway 10 and equidistant from the longitudinal edges of the runway 10. This longitudinal axis is called the axis of the runway Y. An axis perpendicular to the axis of the runway Y is also represented in [Fig.2], by the reference "X".

[0062] The landing strip 10 comprises a set of characteristic elements 16.

[0063] For the sake of clarity, only certain characteristic elements 16 are numbered on [Fig.2],

[0064] Mandatory characteristic elements, not shown, define first of all the track, as such, in particular via the contrast of the bitumen or concrete which composes it with the surroundings of the track, these surroundings being generally composed of grass, earth, sand, etc., and allow in particular to determine the parallelism and degree of alignment of the edges of the track, the homogeneity of the interior of the track, whose shape is a rectangle, its length, its width its symmetry, etc.

[0065] These characteristic elements defining the track, as such, are standardized.

[0066] The characteristic elements 16 are, for example, lamps, also called light beacons. Alternatively, the characteristic elements 16 are radar reflectors (of the trihedral or Luneberg lens type). Alternatively still, the characteristic elements 16 are other existing elements on the periphery of a runway or lights flush with the runway centerline.

[0067] Following the example of [Fig.2], the characteristic elements 16 are distributed on the landing strip 10 in at least two longitudinal rows 18A, 18B and at least one transverse row 20.

[0068] The longitudinal rows 18A, 18B are substantially parallel to the axis of runway Y. Two of the longitudinal rows 18A, 18B are arranged each along a distinct longitudinal edge of runway 10. By the term "arranged along", it is understood that the rows are arranged within three meters of the corresponding longitudinal edge.

[0069] The or each transverse row 20 is substantially perpendicular to the runway axis Y. In the example illustrated by [Fig. 1], the runway 10 comprises three transverse rows 20A, 20B, 20C of characteristic elements 16 and a longitudinal row 20D upstream and in line with the runway axis, said longitudinal row 20D forming a cross with each of the transverse rows 20B and 20C. The last transverse row 20A in the direction of aircraft landing on the runway 10 is called the runway threshold (row 20A in the example of [Fig. 2]).

[0070] Elements 18A, 18B and 20A form what are commonly called runway lights. The aforementioned characteristic elements, not shown, define the runway as such, that is, independently of the runway lights.

[0071] The characteristic elements corresponding to the other transverse rows 20B and 20C as well as other rows not shown, transverse or longitudinal and distinct from the aforementioned rows, are suitable for forming at least one cross-shaped pattern arranged on the ground and located in front of the runway, and form what are commonly called approach lights, in particular approach lights for a runway of type "Category I" as described in particular in the aforementioned document of the International Civil Aviation Organization ICAO, Annex 14, Volume I, in paragraphs 3.1, 4.1, 5.2 and 5.3.

[0072] Runway lights, like approach lights, are standardized optional characteristic elements.

[0073] Other characteristic elements not shown are also likely to be detected such as reflectors specific to the aerodrome in question, these other optional characteristic elements not being mandatory nor all standardized.

[0074] Alternatively, the landing runway 10 comprises at least three longitudinal rows: the two longitudinal rows 18A, 18B and a third longitudinal row 18C (not shown) arranged along the axis of the runway Y and below the runway threshold.

[0075] Fig. 3 schematically illustrates an example of the architecture of an electronic device 14 for decision support during the approach phase for landing of an aircraft on a runway.

[0076] According to the architectural example of [Fig.3], the electronic decision support device 14 for the approach phase for landing includes firstly a module 22 for obtaining the said characteristic elements detected via said at least one image capture device.

[0077] The electronic decision support device 14 for the approach phase for landing also includes a determination module 24 configured to determine an approach scene recognition score associated with said elements characteristics detected, based on a reference approach scene associated with said landing runway.

[0078] The electronic decision support device 14 for the approach to landing further includes a comparison module 26 configured to compare said approach scene recognition score to a predetermined decision threshold value corresponding to the minimum value to be respected to allow the continuation of the approach to landing.

[0079] The electronic decision support device 14 for approach to landing finally includes a decision support module 28 configured to assist in the decision by providing the result of the comparison to a pilot, or to an autopilot device, capable of continuing the approach phase in the event of a score greater than said predetermined decision threshold value, or of interrupting the approach phase by going around in the event of a score less than said predetermined decision threshold value.

[0080] In the example illustrated by [Fig.3], the electronic decision support device 14 for the approach phase for landing includes an information processing unit 30 formed for example of at least one memory 32 and a processor associated with the memory 34 and, optionally, not shown, an information support reader, or an optional human-machine interface including a keyboard and a display.

[0081] In the example of [Fig. 3], the acquisition module 22, the determination module 24, the comparison module, and the decision support module are each implemented as a software program, or a software component, executable by the processor. The memory of the electronic decision support device during the approach phase for landing is then capable of storing acquisition software, determination software, comparison software, and decision support software. The processor is then capable of executing each of the following software programs: acquisition software, determination software, and comparison software, as well as, optionally, the decision support software.

[0082] In an alternative not shown, the acquisition module, the determination module and the comparison module, as well as, as an optional complement, the decision support module, are each implemented in the form of a programmable logic component, such as an FPGA (Field Programmable Gate Array), or an integrated circuit, such as an ASIC (Application Specified Integrated Circuit).

[0083] When the electronic decision support device in the approach phase for landing is implemented in the form of one or more software programs, i.e. in the form of a computer program, also called a computer program product, it is also capable of being recorded on a medium, not shown, readable by computer. A computer-readable medium is, for example, a medium capable of storing electronic instructions and being connected to a computer system bus. Examples of such media include optical discs, magneto-optical discs, ROMs, RAM, any type of non-volatile memory (such as FLASH or NVRAM), or magnetic cards. A computer program containing software instructions is then stored on this readable medium.

[0084] The operation of the decision support device 14 in the approach phase for landing will now be described with reference to [Fig.4] which schematically illustrates an example of the implementation of a decision support method 40 in the approach phase for landing.

[0085] The decision support method 40 for approach to landing is implemented during an approach of the aircraft 12 to the runway 10 for landing.

[0086] The decision support method 40 for the approach phase for landing is advantageously implemented in real time, that is to say at every instant.

[0087] The main steps of the process according to the present invention are described first below.

[0088] The method 40 for decision support during the approach phase for landing of an aircraft on a runway includes a step 42 of obtaining OBT of said characteristic elements detected via said at least one image capture device 15 configured to detect, during the approach for landing, within at least one image, at least one characteristic element 16 of the runway, each detection being associated with coordinates.

[0089] The method 40 further includes a step 44 of determining DET an approach scene recognition score associated with said detected characteristic elements, based on a reference approach scene associated with said landing runway.

[0090] The method 40 also includes a step 46 of comparing said CoMP approach scene recognition score to a predetermined decision threshold value corresponding to the minimum value to be met to allow the continuation of the approach for landing.

[0091] Finally, the method includes a decision aid step 48 A_D by providing the result of the comparison to a pilot, or to an autopilot device, capable of continuing the approach phase in the event of a score greater than said predetermined decision threshold value, or of interrupting the approach phase, by going around, in the event of a score less than said predetermined decision threshold value.

[0092] More specifically, "rendering" means the display (via a screen or part of a screen) of the result of the comparison, the audible reproduction (via a loudspeaker) of the result of the comparison, or the transmission, by radio or otherwise, of a message containing the result of the comparison. Alternatively, the comparison result is suitable for being provided to a pilot after being transformed into a literal form of advice or instruction to be applied, namely in the form of a message inviting (i.e. authorizing) the continuation of the approach phase if the approach scene recognition score is higher than the predetermined decision threshold value, or to interrupt the approach phase and go around if the score is lower than the predetermined decision threshold value.

[0093] Each of the aforementioned steps and associated examples or implementation variants are described in more detail below.

[0094] During the first step 42, the electronic device 14 for decision support during the approach phase for landing of an aircraft on a runway obtains (i.e. collects) the characteristic elements detected via said at least one image capture device 15 (or image capture module when, according to a variant, the image capture functionality is integrated within the device 14 for decision support during the approach phase for landing according to the present invention.

[0095] According to a preferred aspect, the image capture device 15 is a radar. The advantage of basing the detection of characteristic elements of an approach scene on radar-type measurements is to allow the exploitation of wavelengths that are not very sensitive to atmospheric conditions (fog, snow, etc.), of low-cost production technologies, such as frequency-modulated continuous wave (FMCW) radars, and of distance and longitudinal spacing information (sometimes difficult to acquire in the field of optics).

[0096] For example, the characteristic elements obtained during step 42 are previously detected (i.e. identified) using known techniques, including identification techniques from radar data, such as those described in patents or patent applications FR 1912940, FR2309055, or using a Hough transform (i.e. or any other pattern recognition technique) which make it possible to identify points of interest (POI) and lines of interest of an approach scene, for example lights, the threshold and runway edges, as well as the recognition of patterns associated with the runway, associated with approach lights, the threshold, the runway edge, etc.

[0097] In other words, the output provided by such techniques is an input of the present invention which relates to a specific post-processing of the characteristic elements of approach scene suitable for being identified via said known techniques.

[0098] Following this obtaining 42, the process 40 includes the aforementioned determination step 44.

[0099] As an optional supplement, as illustrated by [Fig. 2], said determination 44 of an approach scene recognition score comprises, a first sub-step 50 implementation based on said detected characteristic elements, to provide a CLASS classification of information associated with said detected characteristic elements, within a set of distinct predetermined elementary categories.

[0100] Note that this classification 50 is specific to the needs of the electronic decision support device 14 for the approach phase of an aircraft landing on a runway. It is possible that, beforehand, another classification of information may be obtained jointly with the information on the detected characteristic elements by said at least one image capture device 15 provided as input to said electronic decision support device 14 for the approach phase, but it is likely that this prior classification will not correspond to its needs.

[0101] According to a variant of this optional complement, said set of predetermined distinct elementary categories comprises at least three distinct subsets of distinct elementary categories relating respectively to:

[0102] - elements relating to approach lights and runway lights;

[0103] - runway-related elements, distinct from approach light-related elements and runway lights;

[0104] - other specific elements of the approach scene and distinct from the elements relating to approach lights, runway lights and relating to the runway.

[0105] For example, the subset of elements (i.e., information) relating to lights, namely runway lights and approach lights, is itself organized into classes, in particular six classes, with a first class of information Cl relating to the number of lights identified by type of light such as longitudinal approach lights, transverse approach lights, threshold lights, runway edge lights, runway centerline lights, etc. Each type of this class Cl is also optionally associated respectively with an identifier such as CIA for longitudinal approach lights, C1B for transverse approach lights, CIC for threshold lights, C1D for runway edge lights, CIE for runway centerline lights.

[0106] Another class of information C2 relates to the degree of alignment of the longitudinal lights identified by type of light such as longitudinal approach lights, longitudinal runway edge lights, longitudinal runway center lights, etc. Each type of this class C2 is also optionally associated with an identifier such as C2A for longitudinal approach lights, C2B for longitudinal runway edge lights, C2C for longitudinal runway center lights, etc.

[0107] Another class of information C3 relates to the number C3A, the width C3B and the detected symmetry C3C of the transverse rows of lights.

[0108] Another class of information C4 relates to the degree of perpendicularity of the transverse lights to the longitudinal lights, and / or to the perpendicularity of the runway threshold (and of the runway threshold lights if installed on the ground) to the longitudinal axis of the runway (determined by the longitudinal approach lights (if installed) and / or by the runway edges and / or by the runway edge lights (if installed).

[0109] Another class of information C5 relates to the regularity of the spacing between the longitudinal approach lights C5A, edge lights C5B, and runway centerline lights C5C, considering that by regularity we mean "equally spaced between them", this spacing being in accordance with the regulations (specific to each category of lights and each category of approach); this regularity information being suitable for taking into account possible detection gaps.

[0110] Another class of information C6 relates to the contrast of the lights, in level (i.e. intensity of reflections) or in homogeneity, and this type by type: C6A for longitudinal approach lights, C6B for transverse approach lights, C6C for threshold lights, C6D for runway edge lights, C6E for runway centerline lights, etc.

[0111] Similarly, the subset of elements (i.e. information) relating to the runway, as such, (and distinct from the previous elements relating to the lights including runway lights) are organized into classes, in particular five classes with a class C7 relating to the parallelism and degree of alignment of the runway edges (i.e. on its symmetry in terms of C7A runway contrast and in terms of C7B runway edge lights), a class C8 relating to the homogeneity of the interior of the runway rectangle (excluding runway center lights), a class C9 relating to the detected length of the sides of the apparent runway rectangle, a class C10 relating to the detected width of the runway when available, a class Cl1 relating to the symmetry of the runway with respect to the runway centerline lights.

[0112] As an optional supplement, the subset relating to other specific elements of the approach scene is organised into class(es) and includes in particular class C12 relating to reflectors specific to the runway in question.

[0113] According to this optional supplement, said determination 44 of an approach scene recognition score includes, a second substep 52, implemented for each elementary category, of determination D_S of an elementary score associated with the information classified in said category, and a third substep 54 of combination CoMB, according to a predetermined logic, of said elementary scores to form said approach scene recognition score.

[0114] As an optional supplement, each elementary score from substep 52 is:

[0115] - normalized according to the information corresponding to said elementary category and associated with said reference approach scene, and / or

[0116] - multiplied by one when the said information associated with the said elements Detected characteristics satisfy a predetermined condition and are zero otherwise; and / or

[0117] - weighted according to a predetermined weighting.

[0118] Subsequently, each predetermined condition specific to each landing runway and to each class (or even to each subclass as explained below by way of example) is for example stored in a database by being bijectively associated with the current runway in question (i.e. of the flight in progress), said database instantiating, where appropriate, a predetermined standard such as a standard of the International Civil Aviation Organization ICAO (in particular the aforementioned document of the International Civil Aviation Organization ICAO, Annex 14, Volume I, in paragraphs 3.1, 4.1, 5.2 and 5.3).

[0119] For example, for the aforementioned Class Cl concerning the number of identified lights, and for the runway in question, the predetermined condition PreCl that must be verified by the detected information relating to the number of lights is that there must be at least four longitudinal approach lights, four longitudinal runway edge lights, and at least three transverse approach lights and three transverse threshold lights. If this is not the case, the elementary score of Class Cl (i.e., Elementary Category Cl) is zero.

[0120] If on the contrary this predetermined condition for class Cl is verified, the elementary score of class Cl is obtained by normalizing type by type the number of fires detected.

[0121] For example, for CIA longitudinal approach lights, five lights were detected via the image capture device, whereas according to the aforementioned database, the total number of longitudinal approach lights is equal to 10, so that the score element SCia=5 / 10=0.5.

[0122] Similarly for the transverse approach lights Cl B, an element score SCib=7 / 10=0.7 is obtained when seven out of ten approach lights normally present according to the database have been detected within the image.

[0123] Similarly, for CIC threshold lights, an element of score SCic =3 / 5=0.6 is obtained when three out of five threshold lights normally present according to the database have been detected within the image.

[0124] Similarly, for C1D runway edge lights, an element with a score of SCid=4 / 16=0.25 is obtained when four out of sixteen runway edge lights normally present according to the database were detected within the image.

[0125] Similarly, for CIE runway centerline lights, an element with a score of SCie=4 / 8=0.5 is obtained when four out of eight runway centerline lights normally present according to the database have been detected within the image.

[0126] The elementary score SCi of class Cl is then obtained, for example, by summing, with or without weighting, (or, according to another example, by multiplying, with or without weighting) each of the score elements Scia, Scib, Scie, SCid and Scie and normalizing by the number of fire types, namely, according to this example, five types, such as:

[0127] SCi=(0.5+0.7+0.6+0.25+0.5) / 5=0.51.

[0128] For the aforementioned class C2, which relates to the degree of alignment of longitudinal lights identified by light type, the predetermined condition PreC2, which must be verified by the detected information relating to this class C2, is that none of the angular deviations of the lines formed by contiguous lights in pairs must be greater than or equal to 0.2°. If this is not the case (i.e., at least one angular deviation greater than or equal to 0.2°), the elementary score SC2 of class C2 (i.e., elementary category C2) is zero. Conversely, if the predetermined condition PreC2 is verified, the elementary score SC2 of class C2 is equal to the one's complement of the mean, expressed in degrees, of the angular deviations of the lines formed by contiguous lights in pairs (with respect to the linear regression of the position of the lights of the considered type C2A, C2B, C2C).

[0129] For the aforementioned class C3 relating to the number C3A, the width C3B and the detected symmetry C3C of the transverse rows of lights, the predetermined condition PreC3 which must be verified by the detected information relating to this class C3 is defined in relation to at least one standardized text, for example a standard of the International Civil Aviation Organization ICAO, the number C3A detected having to be strictly greater than the number defined in the standard reduced by 20% (which amounts to being greater than this number according to the standard multiplied by 0.8), the width C3B having to be strictly greater than the width defined in the standard reduced by 10% (which amounts to being greater than this width according to the standard multiplied by 0.9), the symmetry C3C having to be greater than 90%.

[0130] If PreC3 is not verified (i.e., if the C3A number is less than or equal to 0.8 times the standard number, and / or if the C3B width is less than or equal to 0.9 times the standard width, or if the symmetry is less than or equal to 90%), the elementary score of class C3 is zero; otherwise, the elementary score of class C3 is obtained from the three score elements SC3a, SC3b, SC3c such that the score element SC3a is equal to the detected number normalized using (in the denominator) the total number of transverse rows normally present on the runway in question (defined according to a standard or as stored in the database accessible by said decision support device during the approach phase for landing 14); the score element SC3b relating to the width of the transverse rows is equal to the detected width normalized by the total width of the transverse rows defined for the runway considered, according to a standard or as stored in the database accessible by said decision support device during the approach phase for landing 14), the SC3c score element relating to the symmetry of the transverse rows is defined by the following equation: ç _ M <ix-Min oq |y|ax cj mjn représentent les valeurs minimale et maximale des distances des extrémités des détections de feux d’approche transversaux par rapport à l’axe des feux longitudinaux (déterminé par régression linéaire par exemple).

[0131] The elementary score SC3 of class C3 is then obtained for example by summing, with or without weighting, (or according to another example by multiplying, with or without weighting) each of the score elements SC3a, SC3b, SC3c and normalizing by the number of score elements.

[0132] For the aforementioned Class C4 concerning the degree of perpendicularity of transverse approach lights to longitudinal approach lights, the predetermined condition PreC4 that must be verified by the detected information relating to this Class C4 is that the lines formed by linear regression of the detections of longitudinal approach lights and the detections of transverse approach lights must be perpendicular to within 0.2° (this value being given by way of example). If this is not the case (i.e., the perpendicularity error is greater than or equal to 0.2°), the elementary score SC4 of Class C4 (i.e., elementary category C4) is zero.Conversely, if the predetermined condition PreC4 is verified, the elementary score SC4 of class C4, applied to perpendicularity, is equal to the complement to one of the mean, expressed in degrees, of the angular deviations of the lines formed by the lights contiguous two by two, (with respect to the linear regression of the position of the approach, edge, mid-runway type lights).

[0133] For the aforementioned Class C5, which relates to the regularity of the spacing between the longitudinal approach lights C5A, the wingtip lights C5B, and the runway centerline lights C5C, the predetermined condition PreC5, which must be verified by the detected information relating to this Class C5, is that the deviation from a specified predetermined regularity (defined according to a standard or as stored in the database accessible by said decision support device during the approach phase for landing 14) must be at most 10%. If this is not the case, the elementary score So of the Class C5 (i.e., elementary category C5) is zero. Conversely, if the predetermined condition PreC5 is verified, the elementary score SCs of class C5 is equal to the complement to one of the average of the deviations from the theoretical spacing divided by the value of the theoretical spacing (the theoretical spacing being notably equal to 30 or 60m according to ICAO).

[0134] Note that lights can "appear" and then "disappear" in instantaneous measurements, and optionally another class relating to the regularity of spacing between rows of transverse lights could be used (in the case of an appropriate transverse light pattern) by applying a predetermined condition and an elementary score calculation similar to those indicated in relation to class C5 above.

[0135] For the aforementioned class C6 relating to the contrast of the lights, in level (i.e. intensity of reflections) or in homogeneity, the predetermined condition PreC6 which must be verified by the detected information relating to this class C6 is that the ratio of the minimum detected signal-to-noise ratio (SNR) to the theoretical or maximum signal-to-noise ratio must be strictly greater than 80%.

[0136] To do this, type by type of lights: C6A for longitudinal approach lights, C6B for transverse approach lights, C6C for threshold lights, C6D for runway edge lights, C6E for runway centerline lights, etc., an SNR is detected, and it is the minimum SNR that is used to check the predetermined condition PreC6.

[0137] If this is not the case, the SC6 elementary score of class C6 (i.e., elementary category C6) is zero. Conversely, if the predetermined condition PreC6 is met, the SC6 elementary score of class C6 is equal to the average of the SNRs obtained type by type of fire out of the theoretical or maximum SNR (expected elementary score strictly greater than 90%).

[0138] For the aforementioned class C7, which relates to the parallelism and degree of alignment of the runway edges (i.e., its symmetry), the predetermined condition PreC7, which must be verified by the detected information relating to this class C7, is that the maximum angular deviation, in alignment and parallelism, of the lines formed by the runway edges (obtained via contrast) case C7A, and by the lines formed by runway lights contiguous in pairs case C7B, must be strictly less than 0.2°. If this is not the case (i.e., at least one angular deviation greater than or equal to 0.2°), the elementary score SC7 of the class C7 (i.e., elementary category C7) is zero.Conversely, if the predetermined condition PreC7 is verified, the elementary score Sc? of class C7, is obtained from the score element SC7A equal to the one's complement of the angular deviation between the linear regressions forming the two runway edges (expressed in degrees), and from the score element SC7b equal to the one's complement of the mean, expressed in degrees, of the angular deviations of the lines formed by the runway edge lights contiguous two by two.

[0139] The elementary SC7 score of class C7 is then obtained, for example, by summing, with or without weighting, (or, according to another example, by multiplying, with or without weighting) each of the score elements SC7A and SC3b and normalizing by the number of score elements.

[0140] For the aforementioned Class C8, which relates to the homogeneity of the interior of the runway rectangle (excluding center lights), the predetermined condition PreC8, which must be verified by the detected information for this Class C8, is that there are no significant reflectors on the runway (i.e., that no detected SNR exceeds a predetermined threshold). If this is not the case, the elementary SCs score of Class C8 (i.e., elementary category C8) is zero. Conversely, if the predetermined condition PreC8 is verified, the elementary SCs score of Class C8 is equal to the one's complement of the ratio: (Max SNR - Min SNR) / Min SNR (in dB).

[0141] For the aforementioned class C9 relating to the detected length of the sides of the apparent rectangle of the runway, the predetermined condition PreC9 which must be verified by the detected information relating to this class C9 is that the detected length is greater than or equal to 600 m (typically: the distance from the runway threshold to the end of the ground contact zone TDZ (from the English Touch Down Zone).

[0142] If this is not the case, the elementary score Sc 9 of class C9 (i.e., elementary category C9) is zero. Conversely, if the predetermined condition PréC9 is met, the elementary score Sc 9 of class C9 is equal to one (value saturated to one).

[0143] For the aforementioned CIO class relating to the detected runway width when available, the predetermined condition PreCIO, which must be verified by the detected information relating to this CIO class, is that the detected runway width be between -2m and +10m relative to the theoretical runway width. If this is not the case, the elementary SCio score of the CIO class (i.e., elementary CIO category) is zero. Conversely, if the predetermined condition PreCIO is verified, the elementary SCio score of the CIO class is equal to the ratio between the measured (i.e., detected) width and the theoretical width (defined according to a standard or as stored in the database accessible by said decision support device during the approach for landing phase 14).

[0144] For the aforementioned Class Cil concerning the symmetry of the runway with respect to the runway centerline lights, the predetermined condition PréCl 1, which must be verified by the detected information relating to this Class Cl 1, is that the runway must exhibit 90% symmetry. If this is not the case, the elementary score SCn of Class Cl 1 (i.e., elementary category Cl 1) is zero. Conversely, if the predetermined condition PréCl 1 is verified, the elementary score SCn of Class Cil is equal to ç_MaxTMin2, where 'cil “Min2 Max2 and Min2 represent the minimum and maximum distances of the runway edge points from the runway centerline as determined by linear regression of the longitudinal approach lights.

[0145] For the aforementioned class C12 relating to runway-specific reflectors, the predetermined condition PréC12, which must be verified by the detected information relating to this class C12, is that more than 70% of said runway-specific reflectors be detected, or at least three. If this is not the case, the elementary score SCi2 of the class C12 (i.e., elementary category Cl2) is zero. Conversely, if the predetermined condition PréC12 is verified, the elementary score Scn of the class C12 is equal to the ratio of the number of specific reflectors detected to the theoretical number of specific reflectors. Note that such specific reflectors are defined, in particular, by the operational procedures "CAT.OP.MPA."310 and approach publications, such as standardized reflectors corresponding to Precision Approach Path Indicators (PAPI) or Instrument Landing System (ILS) indicators, namely the runway centerline (called "localizer" - LOC) and the glide slope (called "glide slope" - GLIDE), or non-standardized reflectors such as small technical buildings, masts, taxiways and maneuvering areas at the end of the runway, etc., these reflectors associated with the runway in question being defined / stored in the database accessible by said decision support device in the approach phase for landing 14.

[0146] Optionally, and preferably, each elementary score will be provided with its predetermined confidence intervals, taking into account the level of integrity specific to each stage of the approach. For example, relating to class C8, concerning the homogeneity of the interior of the runway rectangle (excluding runway center lights): the intensity of a radar detection is typically expressed as follows: 30 dB plus or minus 2 dB (at 95%).

[0147] According to another example relating to classes C2, C4 or C5..., the position of a detection is estimated by radar to be more or less 2 meters (at 95%).

[0148] Furthermore, according to a preferred variant not shown (as in the case of human piloting, i.e., without recourse to the present invention, where the human pilot determines whether or not they recognize the approach scene), the determination of each elementary score is designed to take into account the memory of past observations (i.e., the concept of an "integrated score"). This avoids being limited to the set of instantaneous measurements collected, by considering a set whose time horizon is, for example, defined by consensus with the pilots and the competent authorities. In other words, according to this preferred variant, it is proposed to access elementary scores previously calculated for similar missions (i.e., the same runway) and stored in the memory of said decision support device during the approach phase for landing 14.

[0149] It should be noted that the database to be used according to the present invention is also capable of containing a significant supplement from "pilot inputs" (i.e. from the human experience of pilots and their interpretation of normative texts and resulting from their intellectual approach, particularly in the event of low visibility) such as DAL-B criticality level information (specified by standard DO 178) (with DAL from the English Design Assurance Level) relating to the final approach segment FAS (from the English final approach segment), and even in the longer term to be enriched by information relating to the types of approach light patterns and therefore the number of transverse lights (with, for example, an enumeration of up to 16 values), relating to the spacing between longitudinal lights (with, for example, an enumeration of the 2 or 3 standardized values), relating to the effective width of the runway, including asphalt or concrete approaches (for example, to the nearest 1 m),relating to the length of the offset if the runway threshold is offset (the runway threshold being essential for localization), relating to specific features of the approach scene such as the turning area and taxiways at the end of the runway, the PAPI (Precision Approach Path Indicator) or the indicators of an ILS (Instrument Landing System) namely the runway centerline (called "localizer" - LOC), relating to the spacing between runways of the same aerodrome, relating to the taxiways of an aerodrome particularly in terms of location (i.e. side and distance from the runway) and characteristics (i.e. width, taxiways, lights...).

[0150] Such pilot inputs are also suitable for use in classification 50, for example to add elementary categories, and in determining elementary score 52, for example in terms of predetermined condition or weighting.

[0151] During substep 54, said elementary scores are combined CoMB, according to a predetermined logic, to form said approach scene recognition score.

[0152] More specifically, during substep 54, the objective is to determine, from the elementary scores, whether the detected approach and landing scene is indeed the one associated with the intended landing runway, (i.e. whether the approach scene is recognized or not).

[0153] This includes determining a probability of detection based on a confidence rate (i.e. the recognition score from said combination) and a probability of false alarm.

[0154] As an optional addition, said predetermined combination logic is specific to vary according to the coordinates associated with each detection (i.e. according to the approach phase; indeed, below one hundred feet in height, the approach lights are no longer visible; the requirement on the confidence level is lower above two hundred). feet than below, etc.), and / or depending on the presence of runway lights, approach lights on said landing runway.

[0155] For example, up to 150 feet high on the runway with approach lights, said predetermined logic consists of combining the elementary scores of the aforementioned categories: Cl to Cil according to a logical "AND", whereas without approach lights or below 150 feet high relative to the ground, only the elementary scores of categories C7 to Cil (i.e. relating to the runway) are combined according to a logical "AND" for example.

[0156] Note that said predetermined logic (i.e. comprising a set of logical equations) and, optionally, the weighting of their components (i.e. of each elementary score) is likely to be based on the analysis of normative texts (for example the electronic Code of Federal Regulations e-CFR and in particular paragraph §91.176 entitled "Visual Reference Requirements", DO-341 entitled "Minimum Aviation System Performance Standards (MASPS) for an Enhanced Flight Vision System to Enable All-Weather Approach, Landing and Roll-Out to a Safe Taxi Speed" and in particular paragraph §3.1.3.1, AC90-106A entitled "Enhanced Flight Vision Systems" and in particular Table 1, etc.), or to be based on the expertise of representatives of pilots in order to take into account human factors and the "automatic" aspect (which will lead to over-constraining the normative requirements).

[0157] To prevent a risk to the statistical independence of the elementary scores, the logical equations of said predetermined logic are advantageously suited to separately group the criteria (i.e., elementary category, class) relating to the runway lights (C1 to C6) and those relating to the runway itself (C7 to C1). Alternatively, such a constraint of statistical independence can be applied upstream to the image capture device (for example, with a quantified objective greater than 99%) instead of being taken into account at the level of said predetermined logic. Failing this, the correlations between the elementary scores will be estimated or measured by testing, and the logical equations of said predetermined logic will be characterized by testing.

[0158] Preferably, in said predetermined logic implemented, the elementary scores are suitable for being weighted to take into account their relative importance.

[0159] More specifically, to obtain (i.e. calculate) the (i.e. overall) approach scene recognition score, the aforementioned elementary scores are each specific to be:

[0160] - lowered to take into account the lower limit of their confidence interval appropriate (said confidence interval being predetermined by considering the level of integrity specific to each step of the approach), and / or

[0161] - as previously stated, multiplied by:

[0162] - the binary value 0 or 1 when the predetermined condition associated with the category elementary of said elementary score is respectively unverified (multiplication by 0) and verified (multiplication by 1), and / or

[0163] - a predetermined associated weighting.

[0164] In addition, depending on the predetermined logic implemented, said elementary scores are multiplied or added together (under the assumption of statistical independence of the components of the logical equation), applying in particular a multiplication in the case of logical "AND" and an addition in the case of logical "OR".

[0165] Note that the logical "OR" is only valid depending on the configurations of the aerodromes, so for a fixed aerodrome, there are only "AND"s to consider, the logical "OR" is therefore probably not considered much unless we consider an indeterminate aerodrome (i.e. airport) for which the presence or absence of approach lights is an unknown.

[0166] The result of this multiplication and / or addition operation according to the predetermined logic implemented is further normalized (i.e. divided) by the theoretical maximum approach scene recognition score defined according to a standard (e.g., document of the International Civil Aviation Organization ICAO, Annex 14, Volume I) or, preferably, as stored in the database accessible by said decision support device in the approach phase for landing 14 for said destination runway of the current approach.

[0167] As an optional complement, said predetermined logic is representative of at least one pattern of characteristic reference elements of said landing strip.

[0168] More precisely, according to this optional complement, as an extension of the logical constructions set out above, the construction of suitable subsets of logical equations is suitable for identifying (more globally) known patterns characterizing an approach scene.

[0169] For example, a first subset of equations is used to identify one or more approach light patterns with respect to possible configurations (for example, those defined by TOACI, in the aforementioned document of the International Civil Aviation Organization ICAO, Annex 14, Volume I, paragraphs 3.1, 4.1, 5.2 and 5.3) or with respect to an expected configuration on the envisaged approach (i.e., a pattern from the aforementioned database accessible by said decision support device during the approach phase for landing 14). This first subset thus aims to provide a recognition of an approach scene and assurance of lateral and longitudinal positioning.

[0170] For example, to identify a Category 1 approach light pattern with five transverse rows of approach lights (i.e., two more than shown in [Fig.2]), the logical equation used is a weighted average of the elementary scores or elementary score elements associated with the aforementioned CIA, C1B, C2A, C3A, C3B, C3C, C4, C5A, C6A and C6B elements.

[0171] A second subset of logic equations is for example suitable for identifying a degree of symmetry of approach and runway lights and uses for this purpose the elementary score from the aforementioned elementary categories C3, C7 and Cil, in order to provide recognition of an approach scene and assurance of lateral positioning.

[0172] A third subset of logic equations is for example specific to identifying the runway as such by the contrast of its surface and by its threshold, edge and runway centerline lights, in order to provide recognition of an approach scene and assurance of lateral and longitudinal positioning.

[0173] A fourth subset of logical equations is for example suitable for identifying a pattern of predetermined points of interest representative of an expected passive or active ground signature from the aforementioned database accessible by said decision support device in the approach phase for landing 14, in order to provide recognition of an approach scene and assurance of lateral and longitudinal positioning.

[0174] As previously stated, in step 46, the resulting approach scene recognition score (CoMP) is compared to a predetermined decision threshold value. This threshold value represents the minimum value that must be met to allow the approach to continue (i.e., proceed).

[0175] Note that the said predetermined decision threshold value is likely to vary depending on the height of the aircraft relative to the ground and is established beforehand by consensus with the pilots and the competent authorities.

[0176] According to another optional supplement, illustrated in dotted lines on [Fig.2], the method 40 further includes a preliminary step 56 in which the decision threshold value is determined as a function of a degree DEG of similarity of said reference approach scene to a set of elements surrounding said landing runway.

[0177] For example, when the approach scene associated with the runway in question closely resembles a section of surrounding motorway, or when an airport has two distinct but similar (i.e., similar and close) runways (e.g., for example, at Lyon Saint-Exupéry Airport, in Nice with a disused runway, and in Rome with a backup runway), the decision threshold is tightened in view of of this degree of resemblance of said reference approach scene to a set of elements surrounding said landing runway, to avoid confusing the approach scene of the landing runway considered proper to the current flight with a "false" surrounding landing runway.. In these cases, differentiating points of interest (POIs) could optionally also be included in the database and taken into account in terms of elementary category to be checked.

[0178] More specifically, the decision threshold is to be established by taking into consideration predetermined target performance of probability of detection and false alarm Pd and Pfa and optionally the environment of the approach scene with the degree of similarity of the surrounding scenes mentioned above.

[0179] The definition of the detection probability Pd and the associated thresholds depends on the objective regarding the false alarm probability Pfa. The false alarm probability Pfa is determined based on the degree of similarity of the surrounding scenes to the scene of the approach under consideration. The radius of the analysis is determined based on the uncertainty of the localization solution that led to the approach for landing, for example: a GNSS (Global Navigation Satellite System) and / or IRS / GNSS (Inertial Reference System) protection radius, taking into account cases of spoofing.

[0180] Typically, the probability of detection Pd is pre-set to 99% (i.e. limiting to 1% the number of cases of approaches not wrongly interrupted) and the probability of false alarm Pfa to 103 corresponding to the number of unnecessary go-arounds.

[0181] It should be noted that these values ​​depend on the system architecture of the aircraft in question. For example, the probability of detection Pd is set at "only" 99% under the assumption of a simple confirmation function with respect to (independent) information provided by GPS (Global Positioning System) satellite geo-positioning or by a Category I ILS (Instrument Landing System).

[0182] During decision aid step 48, the result of this comparison is presented to the pilot (human or automatic) in the approach phase for landing.

[0183] More specifically, presented by "rendering" via the display (via a screen or part of a screen) of the comparison result, or via the audible reproduction (via a loudspeaker) of the comparison result, or via the transmission, radio or otherwise, of a message containing the comparison result. Alternatively, the comparison result is specifically designed to be provided to a pilot after being transformed into a literal form of advice or even an instruction to be applied, namely in the form of a message inviting (i.e., authorizing) the continuation of the approach phase if the approach scene recognition score is higher than the aforementioned decision threshold value. predetermined, or to interrupt the approach phase and resume takeoff if the score is lower than said predetermined decision threshold value.

[0184] Such a reproduction is common, typically at a frequency of 1 to 10 Hz and / or at at least a predetermined height Hk of the aircraft relative to the ground, such as:

[0185] Hk - DA or DH + L50 feet, with k an integer # [0 .. 8] with DA or DH a "decision height" known via the aeronautical acronym DH (from the English Decision Height) or DA (from the English Decision Altitude), notably at least 250 and 200 feet.

[0186] A person skilled in the art will understand that the invention is not limited to the embodiments described, nor to the particular examples of the description, the embodiments and variants mentioned above being capable of being combined with each other to generate new embodiments of the invention.

[0187] The present invention thus makes it possible to propose an automatic solution for the recognition (i.e. identification) of an approach scene for landing, whether intended for a human or automatic pilot, by providing him with an approach scene recognition score whose value is compared to a decision value, so that the result of this comparison serves the pilot (human or automatic) to decide whether or not he should interrupt the approach phase by going around, and thus relieve his mental load.

[0188] This decision support solution is intended to complement and enhance, in terms of integrity, existing location solutions based, for example, on radar-type measurements such as those already described in applications FR1912940 or FR2309055.

Claims

Demands

1. A method (40) for decision support during the approach to landing of an aircraft (12) on a runway (10), the runway (10) having a longitudinal axis equidistant from the longitudinal edges of the runway, called the runway axis (Y), the runway (10) comprising a set of characteristic features (16), the aircraft comprising at least one image capture device (15) configured to detect, during the approach to landing, within at least one image, at least one characteristic feature (16) of the runway, each detection being associated with coordinates, the method (40) being implemented by an electronic decision support device (14) for the approach to landing, and comprising the following steps: - obtaining (42) said characteristic features detected via said at least one image capture device,- determination (44) of an approach scene recognition score associated with said detected characteristic elements, based on a reference approach scene associated with said landing runway; - comparison (46) of said approach scene recognition score to a predetermined decision threshold value corresponding to the minimum value to be respected to authorize the continuation of the approach for landing; - decision support (48) by providing the result of the comparison to a pilot, or to an autopilot system, capable of continuing the approach phase in the event of a score higher than said predetermined decision threshold value, or of interrupting the approach phase, by going around, in the event of a score lower than said predetermined decision threshold value.

2. Method (40) according to claim 1, wherein said determination (44) of an approach scene recognition score comprises: - from said detected characteristic elements, a classification (50) of information associated with said detected characteristic elements, within a set of predetermined distinct elementary categories; - for each elementary category, a determination (52) of an elementary score associated with the information classified in said category; - a combination (54), according to a predetermined logic, of said elementary scores to form said approach scene recognition score.

3. Method (40) according to claim 2, wherein said set of predetermined distinct elementary categories comprises at least three distinct subsets of distinct elementary categories relating respectively to: - elements relating to approach lights and runway lights; - elements relating to the runway, distinct from the elements relating to approach lights and runway lights; - other elements specific to the approach scene and distinct from the elements relating to approach lights, runway lights and the runway.

4. Method (40) according to claim 2 or 3, wherein each elementary score is: - normalized according to the information corresponding to said elementary category and associated with said reference approach scene, and / or - multiplied by one when said information associated with said detected characteristic elements satisfies a predetermined condition and by zero otherwise; and / or - weighted according to a predetermined weighting.

5. Method (40) according to any one of the preceding claims 2 to 4, wherein said predetermined combination logic is adapted to vary according to the coordinates associated with each detection, and / or according to the presence of runway lights, approach lights on said landing runway.

6. Method (40) according to any one of the preceding claims 2 to 5, wherein said predetermined logic is representative of at least one pattern of characteristic reference elements of said landing strip.

7. A method (40) according to any one of the preceding claims, wherein said decision threshold value is predetermined (56) as a function of a degree (DEG) of similarity of said reference approach scene to a set of elements surrounding said landing runway.

8. A computer program comprising software instructions which, when executed by a computer, implement a method (40) for decision support in the approach phase for landing of an aircraft according to any one of the preceding claims.

9. Electronic decision support device (14) for the approach to landing of an aircraft (12) on a runway (10), the runway having a longitudinal axis equidistant from the longitudinal edges of the runway, called the runway centerline, the runway comprising a set of characteristic features (16), the aircraft (12) comprising at least one image capture device (15) configured to detect, during the approach to landing, within at least one image, at least one characteristic feature of the runway, each detection being associated with coordinates, the electronic decision support device (14) for the approach to landing comprising: - an acquisition module (22) configured to obtain said characteristic features detected via said at least one image capture device,- a determination module (24) configured to determine an approach scene recognition score associated with said detected characteristic elements, based on a reference approach scene associated with said landing runway; - a comparison module (26) configured to compare said approach scene recognition score to a predetermined decision threshold value corresponding to the minimum value to be met to authorize continuation of the approach for landing; - a decision support module (28) configured to assist in the decision by providing the result of the comparison to a pilot, or to an autopilot system, capable of continuing the approach phase in the event of a score higher than said predetermined decision threshold value, or of aborting the approach phase by going around in the event of a score lower than said predetermined decision threshold value.

10. Aircraft (12) comprising at least one image capture device (15) configured to detect, during the approach for landing, within at least one image, at least one characteristic feature of the runway, the image capture device being an element belonging to the group comprising: - a radar, and / or - a camera measuring visible or infrared radiation; the aircraft further comprising a decision support device (14) for the approach phase of an aircraft on a runway according to claim 9.

Images