Method for determining, using an optronic system, positions and orientations in a scene, and associated optronic system and vehicle

Abstract

The present invention relates to a method for determining, using an optronic system (18), at least one position and at least one orientation in a scene (10), the scene (10) comprising reference elements (12) of known geographical coordinates, the method comprising an initialization phase that comprises the following steps: - acquiring, using an omnidirectional imager, a panoramic datum in relation to the scene (10), - displaying the panoramic datum on a display element, - collecting reference data for at least three reference elements (12) in the scene (10), and - determining a position and orientation, referred to as close-up position and close-up orientation, of the optronic system (18) on the basis of the reference data stored for the at least three reference elements (12).

Description

[0001]DESCRIPTION TITLE: Method for determining positions and orientations by an optronic system in a scene, associated optronic system and vehicle The present invention relates to a method for determining at least one position and at least one orientation by an optronic system in a scene. The present invention also relates to such an optronic system and an associated vehicle. Collaborative combat, the prerogative of modern armed forces, aims to share the tactical situation with all the actors in the theater of operations thanks to information-valuation. The first information to be shared between the actors is the position of each. Since the 90s, the use of satellite positioning systems (GNSS for Geolocation and Navigation by a Satellite System) has become widespread. GNSS (Global Navigation Satellite System) systems cover GPS (from the English "Global Positioning System" which can be translated into French as "Satellite geopositioning") and other constellations such as GLONASS, GALILEO, BEIDOU. However, such systems are susceptible to being jammed or deceived, which does not allow them to be used reliably. This is referred to as "GNSS denied", i.e. denial of service to the use of satellite location. In addition, it has been shown that it is possible to destroy a satellite in orbit, and therefore services based on constellations of several satellites. Finally, GNSS systems do not work when the signal strength becomes too weak, for example inside buildings or under cover. We therefore return to the use of conventional navigation solutions, in particular by means of an inertial unit or by using a map and a manual or digital magnetic bearing compass. In particular, the use of an inertial unit allows, after initial positioning, to maintain acurrent position by measuring all the wearer's movements. This system still requires regular recalibration by an external means (typically a GPS recalibration) to reduce drifts. In addition, its quality on land deteriorates with increasing latitude and in a lower gravity environment. The use of a map and a manual or digital magnetic bearing compass remains the most rustic way of positioning. It involves leaving the vehicle to avoid biasing the field measurement by the metal masses of the machine and also requires a certain technical skill. This method is therefore long, tedious and risky because it requires exposure to potential external threats, and furthermore remains imprecise (positioning to a few dozen meters, use of Magnetic North requiring declination corrections). There is therefore a need for an optronic system to determine positions and orientationsin a more precise and robust manner. To this end, the present description relates to a method for determining at least one position and at least one orientation by an optronic system in a scene, the scene comprising reference elements of known geographical coordinates, the optronic system comprising the following elements integrated in the optronic system: - an omnidirectional imager capable of acquiring panoramic images of the scene, the panoramic images being formed of pixels identified by pixel coordinates, - a memory in which is stored, for at least each reference element of the scene, an indicator representative of said reference element associated with the geographical coordinates of said reference element, - an element for displaying the indicators stored in the memory and the panoramic images acquired by the omnidirectional imager, - a calculation unit, the method being implemented by the elements integrated in theoptronic system and comprising an initialization phase comprising the steps of: - acquisition of a panoramic image or a stream of panoramic images of the scene by the omnidirectional imager, the panoramic image or the stream of panoramic images being called panoramic data, - display of the panoramic data on the display element, - collection of reference data for at least three reference elements of the scene, the collection step comprising for each of the at least three reference elements: ^ the acquisition, by the calculation unit, of the coordinates of at least one pixel of the panoramic data representative of the reference element considered, ^ the determination, by the calculation unit, of an orientation for the reference element considered as a function of the pixel coordinates of the at least one pixel (pointed), ^ the pointing, on the display element, among the stored indicators, of an indicator representative of the reference element considered,^ the acquisition, by the calculation unit, of the geographical coordinates associated with the pointed indicator, ^ the storage of a datum, called reference, comprising the determined orientation and the geographical coordinates acquired for the reference element considered, - determination of a position and an orientation, called approximate position and approximate orientation, of the optronic system as a function of the reference data stored for the at least three reference elements. According to other advantageous aspects of the invention, the method comprises one or more of the following characteristics, taken individually or in all technically possible combinations: - the initialization phase comprises a step of determining the approximate orientation of the Geographic North and / or the Cartographic North as a function of the reference data stored for the at least three reference elements; - the initialization phase comprises a step of determining the poseapproximate position of the optronic system as a function of the determined approximate orientation, and of the roll and pitch of the omnidirectional imager in the local geographical reference frame, the approximate pose comprising the approximate position and the approximate attitude; - the roll and pitch of the omnidirectional imager in the local geographical reference frame are obtained by a measurement from a measuring device, such as an inclinometer or an inertial unit, or are obtained from the image information alone as a function of the approximate position of the optronic system and the reference indicators stored in the memory; - the optronic system comprises a steerable reduced field imager equipped with a zoom, the reduced field imager being capable of acquiring images of the scene, called small field images, with a variable field of vision of less than 360°, the small field images having a resolution greater than the resolution of the panoramic images acquired by the imageromnidirectional, during the collection step, the acquisition of the coordinates of pixels representative of at least one reference element being implemented in the following manner: - the acquisition of a first small field image of the scene by the reduced field imager, the first small field image being a zoomed image of a portion of the scene, - the display of the first small field image on the display element, - the pointing, on the display element, of at least one pixel imaging the reference element considered on the first small field image to obtain pixel coordinates, - the acquisition of a second small field image of the scene by the reduced field imager, the second small field image being a zoomed-out image comprising at least a part of the portion of the scene imaged on the first small field image, the reduced field imager having an orientation substantially equal to the orientation corresponding to the acquisition of the first and secondsmall field image, - determining the pixel coordinates of at least one pixel of the second small field image corresponding to the pixel pointed at in the first small field image, and - acquiring the pixel coordinates of at least one pixel of the panoramic image corresponding to one of the pixels of the second small field image whose pixel coordinates have been determined; - the method comprises a phase of optimizing the determined approximate position and approximate orientation, the optimization phase comprising: - repeating the step of collecting reference data of the initialization phase for other reference elements so as to obtain reference data for a number of reference elements strictly greater than three, and - determining a position and an orientation, called optimized position and optimized orientation, of the optronic system as a function of all the stored reference data, the optimized position having aaccuracy greater than the accuracy of the approximate position, the optimized orientation having an accuracy greater than the accuracy of the approximate orientation, - preferably, determining an optimized orientation of the Geographic North and / or Cartographic North as a function of all the stored reference data; - the method comprises a phase of updating a position determined for the optronic system during a movement of the optronic system, the updating phase comprising the steps of: - setting up automatic tracking of the movement, on the panoramic data, of the reference elements corresponding to the stored reference data, - updating the reference data as a function of the pixel coordinates of the pixels imaging the tracked reference elements and of the previous reference data, and - updating the at least one position and the at least one orientation of the optronic system as a function of the reference data updatedup to date; - the method comprises, during the update phase, determining the speed of movement of the optronic system as a function of the current position determined, the previous position and the dating of said information, for establishing navigation tracking; - the optronic system is deployed on a platform together with a navigation instrument, such as an inertial unit, the method comprising a phase of resetting the navigation instrument as a function of a position determined by the optronic system; - each indicator stored in the memory is associated with a category relating to the type of the corresponding reference element, during the reference data collection step, the pointing of an indicator stored in the memory comprising: - the display on the display element of the indicators stored in the memory, and - the highlighting on the display element of the indicators whose categorycorresponds to the type of the reference element considered; - during the reference data collection step, at least one reference element considered comprises a rectilinear structure, the acquisition of the pixel coordinates of pixels representative of the reference element on the panoramic data being carried out by pointing on the display element a straight line segment representative of a portion of the rectilinear structure of the reference element considered; - a preliminary solution, comprising a preliminary position and a preliminary attitude, has been obtained for the optronic system, the step of acquiring the coordinates of at least one pixel of the panoramic data, representative of the reference element considered, comprising the determination of search zone(s), on the panoramic data, of the at least one pixel to be acquired, as a function of the preliminary solution and the geographical coordinates of the reference elements, and the display of the zone(s)search determined on the panoramic data; - during the optimization phase, the determination of the optimized position and the optimized orientation of the optronic system is also a function of the pixel coordinates on the panoramic data of a pixel imaging a celestial body and the date and time of acquisition of the panoramic data; - the acquisition of the coordinates of each pixel on the panoramic data is associated with an acquisition uncertainty, called the first uncertainty, the pointing of each reference element on the display element is associated with a pointing uncertainty, called the second uncertainty, each geographic coordinate is associated with an uncertainty on said geographic coordinate, called the third uncertainty, the phase of determining an approximate position and an approximate attitude of the optronic system comprising the estimation of an uncertainty on the approximate position and on the approximate orientation determined as a function of theless than the first, second and third uncertainty. The present description further relates to an optronic system for determining at least one position and at least one orientation by an optronic system in a scene, the scene comprising reference elements of known geographical coordinates, the optronic system (comprising elements integrated in said optronic system and configured to implement a method as described above, the integrated elements comprising at least the following elements: - an omnidirectional imager capable of acquiring panoramic images of the scene, the panoramic images being formed of pixels identified by pixel coordinates, - a memory in which is stored, for at least each reference element of the scene, an indicator representative of said reference element associated with the geographical coordinates of said reference element, - an indicator display elementstored in the memory and panoramic images acquired by the omnidirectional imager, and - a calculation unit. The present description also relates to a vehicle, such as a land vehicle, comprising an optronic system according to the second aspect of the invention. Other characteristics and advantages of the invention will appear on reading the following description of embodiments of the invention, given by way of example only, and with reference to the drawings which are: - [Fig 1] figure 1, a schematic representation of a scene comprising reference elements (landmarks), as well as objects of unknown coordinates, an optronic system integrated in a vehicle is also present on the scene, - [Fig 2] figure 2, a schematic representation of an example of an optronic system comprising elements integrated in said system, - [Fig 3] figure 3, a flowchart of an example of implementation of a method for determining positions andof orientations in a scene, - [Fig 4] figure 4, an example of a graph illustrating transformations of change of reference frame, and - [Fig 5] figure 5, an example of a graph illustrating angles and components of a vector between two reference frames. In the following description, the absolute (geographic) orientation of an object in a scene refers to the vector joining the system to the object; the associated direction being the straight line supporting this vector. This absolute orientation is characterized by angles expressed in relation to a geographical reference. The most used are the azimuth angle which expresses the orientation in a locally horizontal plane (tangent to the ellipsoid associated with the geoid) in relation to the local geographical meridian and the elevation angle (or inclination angle) which expresses the orientation in a vertical plane, in relation to the locally horizontal plane. A compass typically allows you to measure an azimuth. An inclinometer typically allows you to measure aelevation. A relative orientation is defined with respect to another orientation (i.e. an angular difference between two orientations), characterized by the bearing angles in the horizontal plane and the elevation angles in the vertical plane. A goniometer typically measures a bearing and an elevation. The attitude of an element refers to the information allowing the element to be completely oriented in a frame of reference covering the 3 dimensions of geographic space (for example, minimally with the 3 Euler angles of roll, pitch and yaw). A scene 10 is illustrated as an example in Figure 1. A scene designates a theater of operations, i.e. the place where an action takes place. The scene is therefore an extended space with sufficient dimensions to allow an action to take place. The scene is typically an outdoor space. The scene 10 includes reference elements 12, also called landmarks or reference structures, having coordinatesknown geographical coordinates. The scene 10 also includes elements having unknown coordinates, also called objects 14. To implement the method according to the invention, an optronic system 18 is located in the scene 10. The optronic system 18 is therefore also an object 14 of the scene 10. The optronic system 18 is preferably mounted on a land vehicle 19 as in the example of Figure 1. The vehicle is for example of the military type, such as an assault tank. Such a military vehicle is particularly suitable for comprising a plurality of weapons and for protecting the operator(s) installed inside the vehicle. Advantageously, a navigation system, such as an inertial unit, is also mounted on the vehicle. Each reference element 12 is a fixed and remarkable object of the scene 10. The coordinates (latitude, longitude) of each reference element 12 are known. Optionally, the altitude of each reference element 12 is alsoknown. The reference elements 12 are, for example, points belonging to the following elements: a construction (building, bell tower, lighthouse, road, bridge, etc.) whose coordinates can be found on an ortho-image of satellite or airborne origin or a topographic map and a natural element (mountain, rock, hilltop, vegetation, tree, etc.) whose coordinates can be found on an ortho-image of satellite or airborne origin. In the example illustrated by Figure 1, the reference elements 12 are constructions and trees. Each other element of the scene 10 different from a landmark is an object 14 of unknown position. In the example illustrated by Figure 1, the objects 14 are vehicles, as well as the optronic system 18 itself. Those skilled in the art will understand that the term "object" is used in the broad sense, and also includes individuals present in the scene 10. The optronic system 18 comprises elements integrated into saidoptronic system 18. By the term "integrated", it is understood that the elements are incorporated physically and software-wise in said optronic system 18. Such elements therefore form a single block in the optronic system 18. The elements integrated in the optronic system 18 presented in FIG. 2 comprise at least the following elements: an omnidirectional imager 20, a memory 22, a display element 24 and a calculation unit 28. Optionally, the optronic system 18 further comprises a reduced field imager 29 and an attitude measurement module 30. The omnidirectional imager 20 (also referred to as "sensor" in the description) is capable of acquiring panoramic images of the scene 10. The native images acquired by the omnidirectional imager can be resampled in a specific geometry or projection; the resulting image is indifferently referred to as a panoramic image in the following. The panoramic images are formed ofpixels identified by pixel coordinates. The pixels are preferably all associated with the same angular value. An imager is said to be omnidirectional when the imager is capable of providing images over 360° in bearing. The elevation range of the objects in the image is, for example, between 75° and -15°. Advantageously, the omnidirectional imager 20 is capable of operating in several spectral bands, for example, in the visible and in the infrared. The omnidirectional imager 20 is, for example, formed by a single camera. This allows for optimal spatial coherence and temporal coherence for the acquired images. Alternatively, the omnidirectional imager 20 is formed by a set of cameras. In one example, the omnidirectional imager 20 is the ANTARES sensor developed by THALES. Data is stored in the memory 22. The data includes in particular, for at least each reference point 12 of the scene 10, an indicatorrepresentative of said reference element 12 associated with the geographic coordinates of said element 12. The indicators are typically visual elements displayable on the display element 24 and making it possible to identify the corresponding reference elements 12. The indicators are, for example, symbols, textual data (name of the reference element 12) or geographic data, also called geographic products. The geographic products include one or more of the following elements: a map, an ortho-image (of satellite or airborne origin), a digital terrain model (DTM) or a digital elevation model (DEM), ephemerides and astrometric catalog. In particular, the DTM is internal data making it possible to obtain the altitude of a point with known coordinates in latitude and longitude, or to measure a distance from the sensor to the scene by ray tracing. The ephemerides and astrometric catalogs make it possibleto precisely determine the orientations of a large number of celestial objects in a geocentric reference frame. In the case where the indicators are geographic products, the optronic system 18 further comprises a geographic information system (GIS) which groups these products (the data) and the software(s) allowing them to be used (visualize, manipulate, etc.). Advantageously, the geographic information system integrates functionalities allowing the display to be modified, for example, to: - center a displayed image on a position (using an actuator, such as a button, a joystick, a mouse pointer, a stylus, a touch, an eye tracker, etc.), - modify the scale or zoom of the data display (increasing or decreasing the zoom) by any means (mouse wheel, touch, joystick / buttons, eye tracker, etc.), - automatically reload the data in the event of a modification of the center of the geographic product (map,ortho-image…) or the display zoom, - point at an element of the map to obtain its geographical coordinates (using an actuator), or - display over the geographical product (in overprint) thematic layers containing different types of information (for example the terrain intervisibility calculated from the observation post). The geographical coordinates of the reference elements 12 are, for example, in the form of metadata associated with said reference elements 12. The geographical coordinates are, for example, expressed by latitude data, longitude data and optionally altitude data (provided by the digital terrain model for example). The precision errors associated with these data are also provided. In one example, only the indicators of the reference elements 12 are stored in the memory 22, all of the indicators then forming a reference book. Such a reference book is specificto be completed by the operator, for example, during a mission preparation phase. This mission preparation phase can be carried out: - either directly with the optronic system 18 which allows references to be created and their coordinates to be entered, - or via an external mission preparation system. In this case, the optronic system 18 has the means for importing data (USB key, Wi-Fi, etc.). Thus, those skilled in the art will understand that the references are either predefined (in mission preparation), or developed in situ, by selecting them on the geographic information system (GIS) or by directly entering their coordinates. For example, the positions of the reference element 12 are noted by the operator on a geographic product as defined previously and recorded in the form of a list (reference book) in the memory 22. In another example, the indicators stored in the memory 22 are geo-referenced points inthe geographic information system, which provides latitude, longitude (and altitude if the digital terrain model is embedded for example) data. Preferably, the memory 22 comprises, in addition to the indicators of the reference elements 12, indicators of all the geo-referenced points on the stored geographic products. In other words, geo-referenced data (such as an ortho image or a map) is data in which each element (pixel, element) is associated with geographic coordinates. The display element 24 is suitable for displaying images from the omnidirectional imager 20 and / or data stored in the memory 22, such as the indicators of the reference elements 12. The display element 24 is, for example, a display, such as an OLED screen. The calculation unit 28 is suitable for receiving data from the other elements integrated in the optronic system 18, in particular images fromthe omnidirectional imager 20 and the reduced field imager 29, data stored in the memory 22 and measurements made by the attitude measurement module 30, or external navigation instruments, for example, on board the vehicle on which the optronic system 18 is mounted. The computing unit 28 is, for example, a processor. In one example, the computing unit 28 interacts with a computer program product which comprises an information medium. The information medium is a medium readable by the computing unit 28. The readable information medium is a medium adapted to memorize electronic instructions and capable of being coupled to a bus of a computer system. For example, the readable information carrier is an optical disc, a CD-ROM, a magneto-optical disc, a ROM memory, a RAM memory, an EPROM memory, an EEPROM memory, a magnetic card, an optical card or a USB key. On the information carrierthe computer program product comprising program instructions is stored. The computer program is loadable onto the computing unit 28 and causes the implementation of a method for determining positions and orientations in a scene 10, when the computer program is implemented on the computing unit 28 as will be described in the remainder of the description. The reduced field imager 29 is orientable (in elevation and in bearing) and equipped with a zoom. The reduced field imager 29 is capable of acquiring images of the scene 10, called small field images, with a variable field of vision less than 360°. The variable field of vision depends on the zoom and is typically between 2° and 20°. The small field images typically have a resolution higher than the resolution of the panoramic images acquired by the omnidirectional imager 20. Advantageously, the reduced field imager 29 is able to operate in several spectral bands, for example, in thevisible and in the infrared. The reduced field imager 29 is, for example, a pan tilt zoom (PTZ) camera. The attitude measurement module 30 is capable of carrying out measurements relating to the reference elements 12 or to the objects 14 of the scene 10. Preferably, the attitude measurement module 30 comprises at least one element, such as a sensor, chosen from: a goniometer, a rangefinder, an inclinometer and a dating system (millisecond, second, day, month, year). The rangefinder is, for example, a laser rangefinder. The operation of the optronic system 18 resulting in the implementation of a method for determining positions and orientations in a scene 10 will now be described with reference to the flowchart in FIG. 3. The determination method is implemented by the elements integrated in the optronic system 18. INITIALIZATION (PHASE 100) The determination method comprises an initialization phase 100 at the end of which a position is obtainedapproximate and an approximate orientation of the optronic system 18, and possibly an approximate attitude. The initialization phase comprises the following steps. Acquisition of panoramic data (step 110) The initialization phase comprises a step 110 of acquiring a panoramic image or a stream (video, for example real time at the image frequency) of panoramic images of the scene 10 by the omnidirectional imager 20. The panoramic image or the stream of panoramic images are called panoramic data in the remainder of the description. In the case of the acquisition of a panoramic image, the acquisition can take place with the carrier stationary or in motion. The speed of the movement must be compatible with the integration time of the omnidirectional imager 20 so as not to introduce blur into the panoramic image. For example, a speed of around 50 km / h can be reached with an integration time of a few milliseconds. Optionally, if the carrier is equipped with a centralinertial, simultaneously with the acquisition of the panoramic image, the position maintained by the inertial unit will be acquired. In the case of the acquisition of a flow, the acquisition takes place with the carrier stationary. Display of panoramic data (step 120) The initialization phase comprises a step 120 of displaying the panoramic data on the display element 24. Collection of reference data (step 130) The initialization phase comprises a step 130 of collecting reference data for at least three reference elements 12 (landmark) of the scene 10. For each of the at least three reference elements 12, the collection step is implemented in the following manner. Acquisition of pixel coordinates on the panoramic data The collection step 130 comprises the acquisition, by the calculation unit 28, of the coordinates of at least one pixel of the panoramic data representative of the reference element 12 considered. In an example implementation,the acquisition is carried out by pointing at the display element 24, at least one pixel of the panoramic data representative of the reference element 12 considered. The pointing is, for example, carried out by an operator. Alternatively, the pointing is carried out automatically (via an automatic landmark recognition algorithm for example). Alternatively, the pointing is carried out semi-automatically (for example by an algorithmic pre-selection, followed by validation by the operator). Preferably, when the panoramic data is a stream of panoramic images, the reference elements 12 are pointed at the same panoramic image (thus giving both temporal coherence and spatial coherence). Alternatively, the reference elements 12 are pointed in the stream of images (in this case a positioning error due to vibrations of the vehicle could reduce the accuracy of the measurement). In another example of implementation, when the systemoptronics 18 comprises a reduced field imager 29, the acquisition is carried out in the following manner: - acquisition of a first small field image of the scene 10 by the reduced field imager, the first small field image being a zoomed image of a portion of the scene 10, - displaying the first small field image on the display element 24, - pointing, on the display element 24, at least one pixel imaging the reference element 12 considered on the first small field image to obtain pixel coordinates, - acquisition of a second small field image of the scene 10 by the reduced field imager, the second small field image being a zoomed-out image comprising at least a part of the portion of scene 10 imaged on the first small field image, the reduced field imager having a similar orientation (substantially equal to within a few degrees, for example to within 5 degrees) to that corresponding to the acquisition of the first and second imagessmall field. In practice, more than two images can be acquired with the reduced field imager, we then have a video in the form of a zoom ramp going from the smallest field to a large field closer to that of the omnidirectional sensor. - the determination of the pixel coordinates of at least one pixel of the second small field image corresponding to the pixel pointed at on the first small field image. This is done, preferably automatically, by matching the pixels of the first and second small field images. For example, for these images acquired at different scales but in the same spectral band and similar orientations, point-type primitives with SIFT or KAZE type descriptors make it possible to establish correspondences between the images. In order to automatically reject bad correspondences, a RANSAC type algorithm (RANdom SAmple Consensus of Fischler & Bolles 1981) is used. The model ofimage shooting is a perspective or pinhole model whose focal length varies for each image. The focal length value is known a priori and can be refined using a generalized beam adjustment algorithm with a variable focal length that can integrate several images (at least 2) acquired with variable focal length and allows the relative attitude of the images of the zoom ramp to be determined. - the acquisition of the pixel coordinates of at least one pixel of the panoramic image corresponding to one of the pixels of the second small field image whose pixel coordinates have been determined. For example, to attach the larger field image of the reduced field imager to a part of the omnidirectional imager, the same type of point primitive and the previous descriptors can be used as long as the 2 sensors have detectors with neighboring spectral bands; the oriented primitives of the omnidirectional imager are inserted into the beam adjustment forconfer an absolute attitude to the entire sequence of images of the zoom ramp of the reduced field imager. The absolute orientations of all the objects visible in the images of the reduced field imager are thus deduced. If the 2 sensors have detectors with distinct spectral bands (example IR and visible color). In an exemplary implementation, an automatic recognition algorithm is implemented to automatically display on the panoramic data the indicators stored in the memory 22 when the corresponding reference elements 12 are recognized on the panoramic data. In an exemplary implementation, a preliminary solution, comprising a preliminary (approximate) position and a preliminary (approximate) attitude, was obtained for the optronic system. The preliminary solution was, for example, obtained by a GNSS device (possibly in degraded operation), or during maintenance of the position andthe attitude of the optronic system by an inertial unit on board the vehicle or in the optronic system, or by a previous approximate calculation. In this case, the reference elements in the field of vision of the omnidirectional imager are determined, as a function of the preliminary solution and the geographical coordinates of the reference elements, as well as one or more search zones for these reference elements on the panoramic data. The search zone is an area (for example an ellipse or a curve materializing the predicted azimuth for the reference) on the panoramic data to be favored for the search for the reference element. Alternatively, a search zone is determined and displayed only for the reference element considered as a function of the preliminary solution and the geographical coordinates of said reference element. For example, the following sub-steps are implemented to identify the reference elements 12: -Extraction of the ellipsoidal coordinates (longitude, latitude) of a landmark after designation on an embedded ortho-image using its metadata, - Calculation of the altitude of the landmark based on its planimetric coordinates; using an embedded digital terrain model (DTM), a priori its ground coordinates for a current level 2 DTM but which may be its effective altitude if the DTM has the nature of a digital elevation model (DEM) or a surface model (SEM); - Calculation of the spatial orientation linking the position of the landmark to the approximate position of the sensor; - Calculation of visibility using the digital terrain model (DTM) in the external dimensions of the corridor in order to assess the chance of the landmark being visible or hidden from the position of the sensor; - Calculation of the orientation of the landmark in the local geographic reference (RGL) based on the approximate position of the sensor and its approximate attitude; - Display on the panoramic data of the point / pixel corresponding to thisorientation; - Calculation of the error on the orientation of the landmark in the RGL according to the error covariances on the coordinates of the landmark of the approximate position and attitude of the sensor; - display on the panoramic data of an ellipse representative of this error covariance in order to prioritize the user's search in this area. In another example, from a landmark point notebook, or a landmark database (BDDA) with 3D coordinates, the following sub-steps are implemented: - Display all the landmarks in the BDDA located within a pre-fixed radius around the position of the sensor, - Calculate as in the previous example, the landmarks having a chance or not of being visible from the position of the sensor. Preferably, once two landmarks have been identified, the choice of the third landmark is made by materializing an exclusion zone (circle) around which the third landmark is preferably not searched for, and by displaying,near each landmark in a landmark database, the accuracy (CE90) that would be obtained for the position by choosing this landmark as the third landmark. The third landmark is chosen from among the landmarks associated with the best accuracy, and not located in the exclusion zone. The preceding examples can be combined with each other. Determining an absolute orientation for the reference element considered The collection step 130 comprises the determination, by the calculation unit 28, of an (absolute) orientation for the reference element 12 considered as a function of the pixel coordinates of the at least one pixel. In particular, in one example, the orientation (in elevation and in bearing) is obtained by multiplying the pixel number (pixel coordinates) by the angular value of the pixel. This is possible because the pixels are extracted from a panoramic image (360° in bearing) and are each associated with the same angular value. Pointing an indicator for the reference element considered The stepcollection step 130 comprises pointing, on the display element 24, among the stored indicators, an indicator representative of the reference element 12 in the scene 10. During this step, the pointing designates the alignment of a reference (digital pointer, stylus) on the indicator of the reference element 12 or the selection of the reference element 12 from a list (reference book). In an exemplary implementation, the pointing step 130 comprises the display on the display element 24, in parallel or successively or superimposed: - the image of the scene 10 comprising the reference element 12 pointed by the digital imager 20, and - the indicators stored in the memory 22. For example, when the display is done in parallel, a part of the display element 24 displays the image of the scene 10, and another part displays the indicators. For example, when the display is done successively, the image of scene 10 on the one hand,and the indicators on the other hand, are likely to be displayed on the entire display element 24. For example, when the display is done in superposition, the indicators of the reference elements 12 are displayed in superposition (approximately) on the image of the scene 10 (by projection in the space of the scene 10). In an exemplary embodiment, each indicator stored in the memory 22 is associated with a category relating to the type of the corresponding reference element 12 (for example, water tower, bell tower, electricity / telephone pylon, wind turbine, building, characteristic tree, mountain peak). In this case, the pointing of an indicator stored in the memory 22 comprises: - the display on the display element 24 of the indicators stored in the memory 22, and - the highlighting on the display element 24 of the indicators whose category corresponds to the type of the reference element 12 considered. For example, in practice,after having designated the landmark on the panoramic image, the operator selects a type of landmark (for example: pylon). The system will then indicate on a map / ortho-image (or equivalent) all the positions of this type (pylons) in the landmark book (or equivalent). He will then simply have to select by clicking on it the landmark that seems good. The system will then acquire the coordinates previously entered in the landmark book. In an exemplary embodiment, during the reference data collection step, at least one reference element 12 considered comprises a rectilinear structure (e.g.: buildings in an urban or peri-urban environment). In this example, the pointing of 2 pixels along the reference element 12 on the panoramic data is carried out by selecting on the visualization element 24 a straight line segment representative of a portion of the rectilinear structure of the reference element 12 considered, the pixels pointed at being the pixels correspondingsaid straight line segment. In particular, in urban or peri-urban environment contexts, the use of straight line segments is particularly effective for both identifying structures in the image and in the ortho-image and designating them more quickly without having to worry about finding a salient point and then strictly matching a point in the image with the corresponding point in the ortho-image. For their use, only the orientation of the designated segments matters and it is not required that their ends correspond two by two. Furthermore, a segment type support has a good chance of presenting a designation error in the images lower than that of a point element at least in the direction perpendicular to the segment where it is used. Thus, the types of correspondences that can be established by the user are the following (ORTHO designates an ortho-image or equivalent, and PANO designates panoramic data): -Classic correspondence between a point (ORTHO) and a point (PANO), - Correspondence of a segment (ORTHO) with a segment (PANO), for example a roadside structure or an immaterial direction, - Correspondence between a point (ORTHO) and a segment (PANO) as for a base of a building partially masked on the grazing view of the omnidirectional imager 20, and - Correspondence of a segment (ORTHO) with a point (PANO). The ability to not seek to strictly match a point of the ortho-image with a point of the image (PANO) of the sensor brings a gain in precision and designation time and this for several reasons: - Less time to search for interesting primitives in the 2 images because finding a point marked in one image already requires special attention without guarantee of finding it in the other image, and - Speed ​​of designation in the image because even if it is necessary to enter 2 points to define the direction of the segment, it is not useful to applyto position the point precisely in the 2 directions of the image. Thus, from a vehicle on a road for example the curbs or sidewalks will constitute elements of choice in this matter. In an exemplary embodiment, the indicators stored in the memory 22 are points geo-referenced on geographical data. During the reference data collection step, the pointing of an indicator stored in the memory 22 comprises the display on the display element 24 of the geographical data and the highlighting on the geographical data of the indicators stored in the memory 22. In another exemplary embodiment, and once the position is obtained, ephemerides are used to determine the direction of bodies in the solar system. During the day the sun is very often visible in the image acquired by the omnidirectional sensor. Its center is automatically extracted in the panoramic image and the ephemerides make it possible to obtain its directionabsolute. This information has at least 2 advantages: on the one hand, improving the position of the sensor already obtained by overabundance of measurements and on the other hand, improving the bearing or attitude of the sensor by having a precise direction equivalent to a long-distance landmark. The previous examples can be combined with each other. Acquisition of geographical coordinates associated with the pointed indicator The collection step 130 comprises the acquisition, by the calculation unit 28, of the geographical coordinates associated with the pointed indicator. The acquisition of the geographical coordinates is, for example, carried out following the implementation of an acquisition command. The acquisition command is, for example, a validation carried out by an operator of the optronic system 18, for example, via an actuator. Storage of reference data The collection step 130 comprises the storage of a piece of data, called reference data, comprising the determined orientation and the coordinatesgeographical positions acquired for the reference element 12 considered. Thus, in the memory 22, the known geographical positions of the reference elements 12 pointed are associated with the orientations obtained for said reference elements 12 via the omnidirectional imager 20. Determination of an approximate solution of the optronic system (step 140) In the following, the term determination of a solution in the broad sense designates a determination of all or part of the following elements: - a 3D position, possibly 2D if reduced to the plane, - a 3D attitude, possibly only an orientation limited to the bearing if reduced to the plane and - the details of the preceding quantities. The initialization phase 100 comprises a step 140 of determination, by the calculation unit 28, of a position, called approximate position, of the optronic system 18 as a function of the reference data stored for the at least three reference elements 12. In an exemplary implementation, theapproximate position of the optronic system 18 is located at the intersection of arcs of circles C1, C2. Each arc of circle C1, C2 passes through the two reference elements 12 of a pair of reference elements 12, which form the ends of the arc of circle, and its radius is such that each point of the arc of circle C1, C2 is the vertex of an angle (signed) ΦAC, ΦAB equal to the angle measured between the two reference elements 12. In practice this approach uses 3 landmarks each pair providing a goniometric circle of the plane. The position of the sensor is located at the position minimizing the distances to the 3 circles, taking into account the measurement errors the 3 circles do not intersect in a single position. Once the position is obtained, the bearing or orientation corresponding to a column of the panoramic image is easily obtained as a solution of a linear system. Other approaches are described in the literature of topographic procedures among the methodsbarycentric, from Dalembert, Italy. Another approach also consists of exploiting an explicit analytical expression of the bearing using 3 landmarks and then determining the position as a solution of a linear system. The choice of an analytical expression facilitates the calculation allowing to predict the error on the position and the bearing taking into account the geometry of the landmarks (angular deviations), the errors on the measurements of their coordinates as well as the error on the angular readings. The angular readings consist of determining for each landmark 2 angles of relative bearing and elevation taking into account their coordinates in the image of the omnidirectional sensor. The bearing obtained characterizes the absolute orientation in the plane of the geographical representation used. In the case of a local geographical reference we obtain the orientation as geographical azimuth (true north); in the case of a cartographic representation we obtain a cartographic azimuth (grid north).Whatever the choice of representation, the transition from one azimuth to another is done by adding the local convergence of the meridian, the angle between the vertical axis of the map projection and the local meridian. This quantity can be calculated without error for a given position and projection. In addition, with a view to achieving positioning in the shortest possible time and in the case of using a P&T camera with a rangefinder, it is possible to limit the number of ranged landmarks to 2 and without resorting to any other measuring instrument than the PANO image and to determine among the 2 position solutions at the intersection of the 2 circles of radii the distance measurements based on the 2 landmarks, the good solution using the equation of the circle passing through the 2 landmarks and the position of the sensor with a radius depending on the position of the landmarks and the angle measured between the 2 landmarks according to the image. Determination of an approximate pose (position and attitude) of the systemoptronics (step 150) The approximate orientation determined previously typically corresponds to the bearing angle of the omnidirectional imager 20 (and therefore of the optronic system 18). In the following, we detail the obtaining of the approximate attitude of the omnidirectional imager 20, the attitude comprising, in addition to the bearing already determined, the attitude of the omnidirectional imager, i.e. the roll and pitch of the omnidirectional imager 20 in the local geographical reference frame, as well as the refinement of the position of the system. When the omnidirectional imager 20 is not inclined relative to the horizontal, the attitude of the omnidirectional imager is zero. When this remains less than a few degrees, its influence disturbs the determination of the directions / angles relatively little and therefore does not alter the position and bearing solution obtained as indicated previously. When the omnidirectional imager 20 is tilted relative to the horizontal local plane(typically inclined vehicle), the attitude of the omnidirectional imager 20 is non-zero. The attitude is in this case determined by a measurement from a measuring device, such as an inclinometer or an inertial unit. Alternatively, if the measurement module 30 does not include an inclinometer or if the attitude data cannot be communicated by the carrier to the system 18, the determination of the pose of the system is carried out based on the image / visual information alone. For this, the initialization phase 100 comprises a step 150 of determination, by the calculation unit 28, of the approximate attitude of the optronic system 18 as a function of the approximate position of the optronic system 18 and the reference indicators stored in the memory 22. In this case, the transformation is determined making it possible to pass from the reference frame of the omnidirectional imager 20 to the local geographic reference point ℛ ^^, as illustrated in Figure 4. In an exemplary implementation, the attitude determination step 150 implements a batch least squares estimation technique or a Kalman-type recursive estimator after initializing a solution of adequate quality. More precisely, in an exemplary implementation, the algorithmic solution adds two angles (roll and pitch) describing the attitude of the omnidirectional imager 20 in space. The three landmarks considered make it possible to estimate the 6 parameters (3 position coordinates and 3 attitude Euler angles) describing the pose of the sensor in space. Knowing the harmonization or pose of the sensor relative to the vehicle, it is possible to deduce the pose of the vehicle after estimating that of the sensor. The calculation of the present example also makes it possible to calculate a theoretical error of the geographic position and the attitude.To proceed, if we do not have a priori inclination information: - an approximate position is obtained with 3 image points on 3 objects of the reference scene 10, a position and a bearing are thus obtained with one of the topographic procedures mentioned. - an approximate attitude of the sensor reference is then calculated using a TRIAD type algorithm for example, this uses the 2 best reference orientations among the 3 available above. - The image pointing errors are translated into angular orientation errors and their consideration increases that of the errors on the coordinates of the landmarks in the observation equations.Alternatively, if approximate information is available on the "flatness" of the sensor around the roll and pitch axes of the horizontal plane thanks to the measurement module 30 or communicated by the carrier, in particular available by means of accelerometric measurements, the previous approximate attitude calculation is optional and the approximate attitude is obtained by completing the roll and pitch measurements of the bearing. Having an approximate pose solution, the local geographic reference (RGL) can be fixed on this position and the observation equations can be written in a form bringing together the following two orientations: - a reference orientation ^^. ^^ in RGL space, fixed according to the coordinates of a landmark and the approximate position: Where: o ^^0, ^^0, ^^0 denote the position coordinates of the sensor, and o ^^ ^^ , ^^ ^^ , ^^ ^^ designate the position coordinates of the landmark considered. - an orientation ^^ ^^ ( ^^^^ , ^^ ^^ ) in the corresponding sensor space, after designating the pixel ( ^^ ^^, ^^ ^^) on the image:^^ ^^ = ^^ ^^ ^^^^ ( ^^ ^^, ^^ ^^) cos ^^ ^^ cos ^^ ^^ (sin ^^ ^^ cos ^^ ^^ )− sin ^^ ^^Where: o ^^ ^^ , ^^ ^^ denote the coordinates in the raw image of the omnidirectional sensor, o ^^ ^ ^ ^ ^ ^^ denotes the function giving the azimuth in sensor reference from the coordinates in the raw image, o ^^ ^ ^ ^ ^ ^^ denotes the function giving the elevation in sensor reference from the coordinates in the raw image, o ^^ ^^ denotes the azimuth in sensor reference, i.e. an azimuth relative to the reference axes of the sensor or the image. For example, aligned with the central column of the raw image of the omnidirectional sensor, o ^^ ^^denotes the elevation in the sensor reference frame. For a horizontal sensor, the origin or zero elevation corresponds to a particular circle in the raw image or a particular line in a so-called dewarped image resampled in cylindrical geometry. In the general case of a non-horizontal sensor base, we work with the image attitude and the reference axes of the sensor attitude correspond to the central column of the image for the azimuth, for the elevation the complement to pi / 2 of the zenith angle corresponding to the angle between the direction of the pixel considered and the ascending vertical axis perpendicular to the base plane. The orientation ^^ ^^ ( ^^ ^^ , ^^ ^^ ) is then transformed from the reference frame from the sensor space to the RGL ℛ reference frame ^^from an attitude matrix ^^, of which we know in practice an approximate initial value ^^0 that we seek to estimate to improve the precision of the approximate attitude. The matrix R0 is obtained as the product of 3 elementary rotations around the X, Y, Z axes of angles φ0, θ0, ψ0. To initialize R0 several approaches are possible: A first solution consists of initializing it by assuming that the sensor is slightly inclined, i.e. φ0 = dφ0 ≅ 0, θ0 = dθ0 ≅ 0. The angle ψ0, corresponding to the bearing in the horizontal plane, takes a value on the interval [−π, π]. It cannot be approximated by 0. In practice, it is initialized to an approximate value ψ0 obtained from a previously indicated topographic procedure. Thus R0 = R φ0,θ0,ψ0 = R0,0,ψ0.Another solution is to determine R0 from 2 reference directions. We then use a TRIAD type algorithm as described by Bar-Itzhack in 1996. In practice the origin of ℛ ^^ can be brought back to that of ℛ ^^; then ( ^^0 = ^^0 = ^^0 = 0). The observation equation on a landmark ' ^^', is written in ℛ ^^ in the form: Where: -^^ ^^ is the distance from the landmark to the sensor: -The rotation increment ^^ ^^0 is fixed by three small rotation angles to be estimated, ^^0, ^^0, ^^0 acting around each axis of the RGL trihedron, ^^ ^^0 = ^^ ^^ ^^ ^^, ^^ ^^ ^^, ^^ ^^ ^^.- ( ^^ ^^ ^^ , ^^ ^^ ^^ , ^^ ^^ ^^ ) a vector representing the angular error between the direction towards the landmark written in the frame ℛ ^^ and that obtained after rotation of the direction on the landmark in reference ℛ ^^by the sensor attitude matrix ^^. In this expression we note the 6 unknowns of position and attitude increment to be estimated. 2 of these 3 equations are independent due to the unity constraint on the norm of the directions. Also, the system is solved with at least 3 directions towards 3 landmarks in a least-squares approach. Another way to express it is to consider the unknown of distance to the landmark ^^ ^^as unknown for each matching providing 3 equations, or for each 2 independent relations. In detail, a direction in sensor space is obtained from the image coordinates and knowledge of the geometry or optical architecture of the sensor. It may be preferable to resample the raw image of the sensor in order to facilitate the interpretation of the panoramic image and reduce the recognition of landmarks and their designation in the image and then to designate the landmarks in this image since it is more naturally interpretable by relying more on the appearance in the ortho-image but especially on the direct vision of the user. This is done for example by means of a cylindrical type projection with several possibilities aimed at preserving: - distances or scale in certain orientations, - angles via for example the Equatorial Mercator projection, - surfaces via for example a Healpix type projection.It does not matter which projection is used for this transformation linking the coordinates of the panoramic image (called OMNI) to the resampled panoramic image (called. PANO) : The transformation has the following characteristics: - this transformation is bijective and invertive, - the transformation can be sampled in the form of a grid also called LUT (Look Up Table), in order to reduce and control the calculation times - it allows to exchange with a negligible error, less than half a pixel, the pixels between OMNI and PANO images - a designation error in the PANO image characterized by the covariance Λ ^^ translates into error Λ ^^ in the OMNI image; the latter transforming into error Λ^^ on the angles of the orientation ^^ ^^ in sensor space with a covariance Λ ^^ ^^: S oit Λ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ ^^ = ^^ ^^ ^^ ^^ ^^ ^^. Λ ^^. ( ^^ ^^ ^^ ^^ ^^ ^^)Where: - ^^^ ^ ^ ^ is the Jacobian of the transformation ^^ ^ ^ ^ ^ , known from the projection equations used or the LUTs; - ^^ ^ ^ ^ ^ is the Jacobian of the transformation of the coordinates of the raw OMNI image to the azimuth, elevation in the sensor frame: ^^ ^^ ^^^^ = ^^ ^^ ( ^^ ^^ , ^^ ^^) ; ^^ ^^ =^ ^ ^^ ^^ ^^ ( ^^ ^^, ^^ ^^) ; these transformations can themselves involve assembly or simbleautage parameters, also called signatures if they are known. - ^^ ^ ^ ^ ^ ^^ is the Jacobian of the transformation passing from the azimuth elevation values ​​to the 3D components of the orientation in the sensor frame: In practice, we (the calculation unit 28) solve the system by means of several observations on N (^3) landmarks: - We calculate the initial solution, - We linearize the observation system and proceed by Newton-Rapshon type iterations, - We weight these equations according to the errors on the landmarks and designations according to the variances allowing us to determine the Ω ^^ , - Alternatively, one can use available a priori information in order to constrain the estimated values ​​in a domain consistent with the a priori information, - At each iteration, increments are calculated on the 6 parameters of the solution - Convergence is detected by a threshold adapted to the estimated increments of position and / or orientation; typically, at iteration 'i', one calculates: And we stop the iterations with suspicion if i reaches a significant value ^^ = ^^ ^^ ^^ ^^ ^^ ^^ ^^ prefixed and with confidence if before this value ^^ ^^^^ et ^^ ^^^^check:^^ ^^ ^^ < ^^ ^^ ^^ ^^ ^^ ^^ ; ^^ ^^ ^^ < ^^ ^^ ^^ ^^ ^^ The values ​​^^^^ ^^ ^^ ^^ ^^ and ^^^^ ^^ ^^ ^^ are set according to the required precision; for example, we choose: ^^^^ ^^ ^^ ^^ ^^= 10 −1 ^^ and ^^^^ ^^ ^^ ^^ ^^= 10 −4 ^^ ^^ ^^. Then, having the precision on the approximate solution, we can add constraint equations to the observation equations in order to process the estimation of the parameters in Bayesian form. - Denoting by Λ ^^0 the covariance on the position, we add 3 equations of the form: ^ ^0 √ Λ ^^0 ( ^^ 0) = 0 ^^0 So that the product by their transpose expresses the constraint: - Designating by Λ ^^0 the covariance on the angles we add 3 equations of the type: So that the product by their transpose expresses the constraint: In such an example of resolution, the solution Θ0 = (P0, R0) is obtained after some 3 to 4 iterations aiming at each step to update Θ0 by the increment ^^Θ0 obtained as solution of the system. After convergence we have the position and attitude of the sensor which best match the set of directions expressed in RGL to the directions of the sensor frame to which the transformation with the estimated rotation and translation are applied. We note DoF the number of degrees of freedom to be estimated corresponding to the number of equations reduced by the number of unknowns. Here for each landmark, we have 3 equations but one unknown distance, i.e. 2 effective equations. Having 6 parameters to estimate we have for N landmarks: - In a classic weighted least-squares approach: ^^ ^^ ^^ = 2 ^^ − 6; in this case at least 3 landmarks are necessary to solve the system.-In a Bayesian approach, adding the 6 equations mentioned: ^^ ^^ ^^ =2^^; in this case 1 landmark is enough to start improving the initial solution. The improvement mainly concerns the parameter(s) with the highest sigma(s). An approach by batch estimation of all measurements on the landmarks, called batch estimation, converges in 2 or 3 iterations and requires as many matrix inversions of the parameter dimension. Alternatively, rather than using a batch estimation approach, it is possible to use a recursive Kalman-type estimator. The implementation of this approach is simpler to integrate the a priori information; it requires both the value information and the approximate covariance for all the parameters to be estimated. The number of matrix inversions of the parameter dimension is this time equal to the number of landmarks N.For a small number of landmarks, for example 3 landmarks, this type of approach will have the same order of complexity as a batch approach; on the other hand, if the number of landmarks becomes larger, it will require more inversions than a batch approach. Another delicate point of this approach lies in the behavior of the filter during an update with a poorly weighted observation, covariance on the non-representative observation. The divergence of the filter after updating with an erroneous observation is restored after integration of other correct observations but this process can be more or less rapid and more difficult to detect than via the analysis of the residuals at the end of a batch estimation. Preferably, as illustrated by figure 5, the harmonization / simbleautage characterizing the mounting of the optronic system on the vehicle is determined according to: - the 3 angles of the attitude matrix ^^. ^ ^ ^ ^defining the passage between the axes of the sensor reference frame to the axes of the vehicle reference frame ℛ ^^ - the 3 components of a vector ^^ ^ ^ ^ ^ allowing the origin to be translated between these two reference points. Knowing the similarity of the optronic system, we deduce from the approximate pose of the optronic system an approximate pose for the vehicle in ℛ ^^ . This can be obtained from image information alone. More precisely, the vehicle's attitude is obtained simply as ^^ ^^ = ^^ ^ ^ ^ ^ ^^ ^^ = ^^ ^ ^ ^ ^ ^^0and the position of the vehicle in the RGL of the optronic system ^^ ^^by applying the attitude matrix obtained to the lever arm ^^ ^^^^ between the 2 reference positions of the optronic system and the vehicle as ^^ ^^ =^^ ^^ ^^ ^^^^ . As the positions of the optronic system and the vehicle are close (a few meters at most), we can neglect the passage relationship between the axes of the RGL attached to the optronic system and the RGL attached to the vehicle. Strictly speaking, we would introduce a complementary rotation depending on the 2 respective positions. This amounts to assuming that the axes of the 2 RGLs are parallel. The solution can also be merged with that of a CNI by means of the approximate position and attitude covariances. Determination of an uncertainty on the approximate position and where appropriate on the approximate attitude of the optronic system (step 160) Preferably, the initialization phase 100 comprises a step 160 of determination, by the calculation unit 28, of the precision of the determined approximate position,and where appropriate the determined approximate attitude. The accuracy is determined based on the accuracy errors associated with the geographic coordinates of the reference elements 12, and the errors related to the extraction of information on the panoramic data and / or when pointing the indicators of the reference elements on the display element 24 as indicated previously. Typically, the overall uncertainty is obtained by a propagation of the uncertainties. OPTIMIZATION (PHASE 200) Optionally,the method comprises a phase 200 of optimizing the determined approximate position and approximate orientation or attitude. The optimization phase 200 comprises repeating the step 110 of collecting reference data from the initialization phase for other reference elements 12 so as to obtain reference data for a number of reference elements 12 strictly greater than three. The optimization phase 200 also comprises determining a position, called the optimized position, and an attitude (or at least one orientation), called the optimized attitude – i.e. an optimized pose or solution to be determined – of the optronic system 18 as a function of all the stored reference data. The optimized position has a precision greater than the precision of the approximate position. The optimized attitude has a precision greater than the precision of the approximate attitude. In particular, in the example cited above,from four reference elements 12 (three circular arcs), the intersection of the circular arcs is not done at a single point (taking into account the errors on the angles, and on the position of the landmarks). The position retained is, for example, the result of an optimization of the nonlinear equations resulting from a problem describing the geometry of the example. Preferably, during the optimization phase, the determination of the optimized pose of the optronic system 18 is also a function of the pixel coordinates on the panoramic data of a pixel imaging a celestial body and of the date and time of acquisition of the panoramic data. The celestial body is, for example, the sun or the moon. This makes it possible to increase the number of degrees of freedom (DoF) without designating additional landmarks. In particular, this makes it possible to redundantly solve the solution and participate in the detection of faults,and also, if necessary, to improve the attitude and in particular the bearing or yaw. In an example of implementation, as soon as an approximate solution is available, i.e. in a DoF=0 situation, the following sub-steps are implemented by the calculation unit 28: - With the date / time, the orientation of the celestial bodies in the celestial reference frame is determined, - With the approximate position, the orientation of the bodies in the local reference frame is calculated, - With the errors, the error on this orientation is determined, - With the orientation of the bodies in the local reference frame, those which are visible are determined, i.e. on the horizon and in the visibility sector of the omnidirectional imager 20, - With the orientation of a visible body, its position in the panoramic image and the error on this position are determined, - With the orientation error, the probable area of ​​the body in the image is determined, - With the type of body,an adapted image processing is used to extract the center of the body and determine its precise position in the image, and, - With this information, a new observation is obtained with direction correspondence in image space and in RGL. Thus the proposed resolution mechanism provides the following advantages: - Calculation of an approximate solution including on terrain presenting a declination and this without the need for a priori information, - Improvement of an approximate solution regardless of the number of landmarks available, - Calculation of the precision of the solution, - Optimality of the solution, - Ability to detect a coordinate error or designation with a redundant number of measurements. Knowing the harmonization / simbleautage characterizing the mounting of the optronic system on the vehicle,an optimal pose for the vehicle is deduced from the optimal pose of the optronic system. This can be obtained from the image information alone. The solution can also be merged with that of a CNI by means of the optimal position and attitude covariances. UPDATE (PHASE 300) Optionally, the method comprises a phase 300 of updating, by the calculation unit 28, a position and an attitude (or orientation) determined for the optronic system 18 during a movement of the optronic system 18. The update phase 300 comprises a step 310 of setting up automatic tracking of the movement, on the panoramic data, of the reference elements 12 corresponding to the stored reference data. The tracking is, for example,performed by an image processing and tracking algorithm. The update phase 300 comprises a step 320 of updating the reference data according to the pixel coordinates of the pixels imaging the reference elements 12 tracked and the previous reference data. The update phase 300 comprises a step 330 of updating the position and orientation of the optronic system 18 according to the updated reference data. Preferably, the update phase 300 comprises determining the speed of movement of the optronic system according to the dating of the information, the current position determined and the previous position for establishing navigation tracking. Knowing the harmonization / simbling characterizing the mounting of the optronic system on the vehicle,A navigation solution for the vehicle is deduced from the maintenance of the optronic system's navigation solution. This can be obtained from image information alone. The solution can also be merged with that of a CNI using position, attitude and speed covariances. DETERMINING THE POSITION OF AN OBJECT IN THE SCENE (PHASE 400) Optionally, the method comprises a phase 400 of determining, by the calculation unit 28, the position of an object 14 in the scene 10 as a function of a position determined for the optronic system 18 (approximate position, optimized position or updated position, preferably the last position obtained for the optronic system 18), of an absolute orientation obtained from the object 14 relative to the optronic system 18 and of a distance obtained between the object 14 and the optronic system 18. The distance is obtained for example either by ray tracing with a DTM,or more precisely with telemetry when a rangefinder is integrated into the reduced field sensor on the pan & tilt. The object 14 considered is visible from the optronic system 18 (within range and not masked). The position of the object 14 is then obtained by calculation, by the calculation unit 28, of the geographical coordinate located at the end of the vector having as origin the position of the optronic system 18, for orientation the absolute orientation of the object 14 and for length the distance between the optronic system 18 and the object 14. Advantageously, the precision on the position of the object 14 is calculated as a function of: - the precision on the position of the optronic system 18, - the precision on the absolute orientation of the object 14, and - the value of the distance and the precision on the distance between the optronic system 18 and the object 14. For the determination of the distance between the object 14 and the optronic system 18, according to an example,when at least one element of the measurement module 30 is a rangefinder, the distance between the object 14 and the optronic system 18 is obtained by a measurement acquired by the rangefinder during pointing of the object 14 by the digital imager 20. According to another example, when the object 14 is on the ground, the distance between the object 14 and the optronic system 18 is obtained by a ray tracing method from a digital terrain model of the scene 10. The distance obtained is then the distance between the determined position of the optronic system 18 and the intersection of a predetermined straight line with the ground of a digital terrain model. The predetermined half-line passes through the determined position of the optronic system 18 and has the orientation obtained from the object 14 relative to the optronic system 18. RECALIBRATION OF NAVIGATION INSTRUMENTS (PHASE 500) Optionally, when the optronic system 18 is deployed on a platform in conjunction with a navigation instrument,such as an inertial unit, the method comprises a phase 500 of recalibrating the navigation instrument as a function of a position determined for the optronic system 18 (approximate position, optimized position or updated position, preferably the last position obtained). When the platform is in motion, the inertial unit will be recalibrated by the error between the position calculated by the method and the position maintained by the inertial unit acquired simultaneously with the acquisition of the panoramic image, as described in step 100. Thus, the present method makes it possible to use a completely autonomous geographic positioning system by image, that is to say independent of an external means such as an inertial unit, GNSS, rangefinder. This method is, moreover, more precise and rapid than an “à la carte” survey. It is also discreet (because it does not require an electromagnetic emission such as a telemetry laser), without the obligation to leave the vehicle,of the accuracy class of a GNSS system without correction (i.e. a positioning accuracy of 3 to 5 meters), which can be achieved in a few tens of seconds in a time shorter than the provision of GNSS information from a receiver starting hot. As a reminder, the time to obtain a GPS measurement reaches 5 to 10 seconds for a simple temporary cut of the signal; 30 to 45 seconds for a hot start which retains the last position information and up to 50 to 80 seconds for a cold start corresponding to a factory exit without pre-operation or lost in a space. In particular, this process uses a precise 360° omnidirectional imager coupled with an on-board mapping system, making it possible to find the geographical coordinates of the position of the optronic system 18 with good accuracy, in a reasonable time, without the crew members of the machine having to dismount. If necessary,it also allows to determine the attitude including the bearing of the sensor (or the orientation of the Geographic North in the panoramic imagery of the vision system). The attitude calculation makes it possible to work on the entire range of terrain inclination that can be encountered. More specifically, the omnidirectional imager makes it possible to have an instantaneous goniometer and controlled local coherence throughout the field of vision. This process is suitable for use on a moving vehicle due to the extraction of input data (pixel coordinates) from panoramic data. In particular, the data obtained can also be used to recalibrate an inertial unit or any other instrument (in practice component / sensor / effector) on board the vehicle. The system allows to correct the drift of the inertial unit by recalibration when bearings are possible,and the inertial unit making it possible to maintain the position of the vehicle when crossing an area that does not allow for bearing: crossing a forest or tunnel for example. Those skilled in the art will understand that the order of the different phases is given by way of example. For example, the determination phase 400 can be implemented after any one of the phases 100, 200, 300 or 500. The same is true for the recalibration phase 500. Similarly, within each phase, the order of the steps is also given by way of example (for the reference data collection step, it is for example possible to start by designating indicators corresponding to chosen reference elements, and then extract the data from the panoramic image, or to do the reverse). Those skilled in the art will understand that the previously described embodiments can be combined to form new embodiments provided that they are technically compatible.,

Claims

CLAIMS 1. Method for determining at least one position and at least one orientation by an optronic system (18) in a scene (10), the scene (10) comprising reference elements (12) of known geographical coordinates, the optronic system (18) comprising the following elements integrated in the optronic system (18): - an omnidirectional imager (20) capable of acquiring panoramic images of the scene (10), the panoramic images being formed of pixels identified by pixel coordinates, - a memory (22) in which is stored, for at least each reference element (12) of the scene (10), an indicator representative of said reference element (12) associated with the geographical coordinates of said reference element (12), - a display element (24) for displaying the indicators stored in the memory (22) and the panoramic images acquired by the omnidirectional imager (20), - a calculation unit (28),the method being implemented by the elements integrated in the optronic system (18) and comprising an initialization phase comprising the steps of: - acquisition of a panoramic image or a stream of panoramic images of the scene (10) by the omnidirectional imager (20), the panoramic image or the stream of panoramic images being called panoramic data, - display of the panoramic data on the display element (24), - collection of reference data for at least three reference elements (12) of the scene (10), the collection step comprising for each of the at least three reference elements (12): ^ the acquisition, by the calculation unit (28), of the coordinates of at least one pixel of the panoramic data representative of the reference element (12) considered, ^ the determination, by the calculation unit (28), of an orientation for the reference element (12) considered as a function of the pixel coordinates of the at least one pixel, ^ the pointing,on the display element (24), among the stored indicators, of an indicator representative of the reference element (12) considered, ^ the acquisition, by the calculation unit (28), of the geographical coordinates associated with the pointed indicator, ^ the storage of a piece of data, called reference data, comprising the determined orientation and the geographical coordinates acquired for the reference element (12) considered, - determination of a position and an orientation, called approximate position and approximate orientation, of the optronic system (18) as a function of the reference data stored for the at least three reference elements (12).

2. Method according to claim 1, in which the initialization phase comprises a step of determining the approximate orientation of the Geographic North and / or the Cartographic North as a function of the reference data stored for the at least three reference elements (12). 3.Method according to claim 1 or 2, wherein the initialization phase comprises a step of determining the approximate pose of the optronic system (18) as a function of the determined approximate orientation, and the roll and pitch of the omnidirectional imager in the local geographic reference frame, the approximate pose comprising the approximate position and the approximate attitude.

4. Method according to claim 3, wherein the roll and pitch of the omnidirectional imager in the local geographic reference frame are obtained by a measurement from a measuring device, such as an inclinometer or an inertial unit, or are obtained from the image information alone as a function of the approximate position of the optronic system (18) and the reference indicators stored in the memory (22). 5.Method according to any one of claims 1 to 4, in which the optronic system (18) comprises a steerable reduced field imager equipped with a zoom, the reduced field imager being capable of acquiring images of the scene (10), called small field images, with a variable field of vision less than 360°, the small field images having a resolution greater than the resolution of the panoramic images acquired by the omnidirectional imager (20), during the collection step, the acquisition of the coordinates of pixels representative of at least one reference element (12) being implemented in the following manner:. - acquiring a first small field image of the scene (10) by the reduced field imager, the first small field image being a zoomed image of a portion of the scene (10), - displaying the first small field image on the display element (24), - pointing, on the display element (24), at least one pixel imaging the reference element (12) considered on the first small field image to obtain pixel coordinates, - acquiring a second small field image of the scene (10) by the reduced field imager, the second small field image being a zoomed-out image comprising at least a portion of the portion of the scene (10) imaged on the first small field image, the reduced field imager having an orientation substantially equal to the orientation corresponding to the acquisition of the first and second small field images,- determining the pixel coordinates of at least one pixel of the second small field image corresponding to the pixel pointed at in the first small field image, and - acquiring the pixel coordinates of at least one pixel of the panoramic image corresponding to one of the pixels of the second small field image whose pixel coordinates have been determined.

6. Method according to any one of claims 1 to 5, wherein the method comprises a phase of optimizing the determined approximate position and approximate orientation, the optimization phase comprising: - repeating the step of collecting reference data of the initialization phase for other reference elements (12) so as to obtain reference data for a number of reference elements (12) strictly greater than three, and - determining a position and an orientation, called optimized position and optimized orientation,of the optronic system (18) as a function of all the stored reference data, the optimized position having a precision greater than the precision of the approximate position, the optimized orientation having a precision greater than the precision of the approximate orientation, - preferably, the determination of an optimized orientation of the Geographic North and / or Cartographic North as a function of all the stored reference data., 7. Method according to any one of claims 1 to 6, in which the method comprises a phase of updating a determined position for the optronic system (18) during a movement of the optronic system (18), the updating phase comprising the steps of: - setting up automatic tracking of the movement, on the panoramic data, of the reference elements (12) corresponding to the stored reference data, - updating the reference data as a function of the pixel coordinates of the pixels imaging the reference elements (12) tracked and of the previous reference data, and - updating the at least one position and the at least one orientation of the optronic system (18) as a function of the updated reference data.Method according to claim 7, in which the method comprises, during the update phase, determining the speed of movement of the optronic system as a function of the current position determined, the previous position and the dating of said information, for establishing a navigation track.

9. Method according to any one of claims 1 to 8, in which the optronic system (18) is deployed on a platform together with a navigation instrument, such as an inertial unit, the method comprising a phase of resetting the navigation instrument as a function of a position determined by the optronic system (18). 10.Method according to any one of claims 1 to 9, in which each indicator stored in the memory (22) is associated with a category relating to the type of the corresponding reference element (12), during the step of collecting reference data, the pointing of an indicator stored in the memory (22) comprising: - the display on the display element (24) of the indicators stored in the memory (22), and - the highlighting on the display element (24) of the indicators whose category corresponds to the type of the reference element (12) considered.

11. Method according to any one of claims 1 to 10, in which during the step of collecting reference data, at least one reference element (12) considered comprises a rectilinear structure, the acquisition of the pixel coordinates of pixels representative of the reference element (12) on the panoramic data being. carried out by pointing on the display element (24) a straight line segment representative of a portion of the rectilinear structure of the reference element (12) considered.

12. Method according to any one of claims 1 to 11, in which a preliminary solution, comprising a preliminary position and a preliminary attitude, has been obtained for the optronic system, the step of acquiring the coordinates of at least one pixel of the panoramic data, representative of the reference element considered, comprising the determination of search zone(s), on the panoramic data, of the at least one pixel to be acquired, as a function of the preliminary solution and the geographical coordinates of the reference elements, and the display of the search zone(s) determined on the panoramic data. 13.Method according to any one of claims 1 to 5 or 7 to 12 in its dependence on claim 6, in which during the optimization phase the determination of the optimized position and the optimized orientation of the optronic system (18) is also a function of the pixel coordinates on the panoramic data of a pixel imaging a celestial body and of the date and time of acquisition of the panoramic data. 14.Method according to any one of claims 1 to 13, in which the acquisition of the coordinates of each pixel on the panoramic data is associated with an acquisition uncertainty, called the first uncertainty, the pointing of each reference element (12) on the display element (24) is associated with a pointing uncertainty, called the second uncertainty, each geographic coordinate is associated with an uncertainty on said geographic coordinate, called the third uncertainty, the phase of determining an approximate position and an approximate attitude of the optronic system (18) comprising the estimation of an uncertainty on the approximate position and on the approximate orientation determined as a function of at least the first, second and third uncertainty. 15.Optronic system (18) for determining at least one position and at least one orientation by an optronic system (18) in a scene (10), the scene (10) comprising reference elements (12) of known geographic coordinates, the optronic system (18) comprising elements integrated in said optronic system (18) and configured to implement a method according to any one. of claims 1 to 14, the integrated elements comprising at least the following elements: - an omnidirectional imager (20) capable of acquiring panoramic images of the scene (10), the panoramic images being formed of pixels identified by pixel coordinates, - a memory (22) in which is stored, for at least each reference element (12) of the scene (10), an indicator representative of said reference element (12) associated with the geographical coordinates of said reference element (12), - a display element (24) of the indicators stored in the memory (22) and of the panoramic images acquired by the omnidirectional imager (20), and - a calculation unit (28).

16. Vehicle, such as a land vehicle, comprising an optronic system (18) according to claim 15.

Images