Method for detecting obstacles with a lidar obstacle sensor system for a rotorcraft

Abstract

The invention relates to a method for detecting obstacles of an aircraft (1) having at least one rotor (5). The method comprises: i) acquiring (STP1) a point cloud detected using at least one laser radar obstacle sensing device (21, 22, 23, 24); ii) positioning (STP2) each detected point in a standard orthogonal reference frame (100) attached to the aircraft, having a first axis (Z) coinciding with the axis of rotation (AXROT) of the rotor (5); iii) dividing (STP3) the surrounding space into volumetric units, one or more detected points present in the same volumetric unit being all replaced by a representative point; iv) segmenting the ground; v) determining (STP5) the presence or absence of at least one obstacle based on different representative points of the ground; and displaying (STP6) on a display a symbology showing the obstacles relative to the aircraft.

Description

[0001] Cross-references to related applications

[0002] This application claims the benefit of French application FR 24 13539, filed on December 6, 2024, the disclosure of which is incorporated herein by reference in its entirety. Technical Field

[0003] This invention relates to a method for detecting obstacles using a LIDAR obstacle sensing device system for rotorcraft.

[0004] An aircraft may include one or more systems to avoid collisions with obstacles (such as buildings, racks, cables, cranes, etc.) during flight. The term "obstacle" as used below refers to any element capable of colliding with the aircraft.

[0005] Various systems are known to avoid collisions during flight near obstacles, such as during rescue missions in mountainous or congested environments with towers, cranes, etc., or during maneuvers close to the ground.

[0006] Such a system may include an active obstacle sensor configured to detect one or more obstacles in flight, and a display showing the detected obstacles. The obstacle sensor may include one or more obstacle sensing devices of a known type, known by the acronym LIDAR (for light detection and ranging). A LIDAR system is equipped with a transmitter that sends light pulses (typically lasers) into a small-aperture detection area. When the pulse encounters an obstacle, it is reflected and captured by a receiver. The system is able to detect and calculate the distance to the obstacle by measuring the pulse return time.

[0007] If all obstacles present in the detection field are detected and signaled, the display showing these detection points may subsequently saturate.

[0008] However, obstacle avoidance warning systems for helicopters should not further burden pilots' already heavy workloads. Specifically, it is expected that continuous alarms will not be triggered when the aircraft is flying above the ground, but rather that the focus will be solely on actual obstacles. If this is not the case, the pilot can disable the system without frustration. Furthermore, the perception system must be able to rapidly analyze and understand the environment in real time. This ensures that each step in the processing chain of the sensing equipment executes within a pre-established timeframe, allowing the pilot to make informed decisions within a safe framework.

[0009] To improve obstacle detection systems, this system can implement real-time ground segmentation algorithms. More specifically, the pilot viewing the outside world actually has a mental representation of the ground, and the display of symbols showing other obstacles may be sufficient to assist him. The drawback of this technology is that it requires significant computational resources, and therefore may require heavy, expensive, and / or massive systems. Background Technology

[0010] In this context, patent document CN 116524219 A is significantly different from the field of this invention and relates to automotive systems. This document describes a method comprising preprocessing a point cloud obtained using sensors, filtering the point cloud by removing outliers, and then subsampling. The method includes first separating the point cloud corresponding to the ground surface, and then separating the point cloud corresponding to one or more obstacles, using a linear adjustment algorithm that takes into account the circular shape of the road.

[0011] References EP 4 390 439 A1, Wang Xianzhe et al., “Research on detection method of airborne obstacle avoidance lidar”, 20231218, Vol. 12963, December 18, 2023 (2023-12-18), pp. 1296318-1296318, XP060195426, CN 116 524219 A and CN 115 909 277 A are also known. Summary of the Invention

[0012] Therefore, the object of this invention is to propose an innovative method and system for detecting obstacles.

[0013] Therefore, the present invention relates to a method for detecting obstacles implemented by an aircraft having at least one rotor. The method for detecting obstacles includes:

[0014] • Use at least one LIDAR obstacle sensor mounted on the aircraft to acquire a detection point cloud;

[0015] • Position each detection point in a predetermined orthogonal reference frame attached to the aircraft, the reference frame having a first axis that coincides with the axis of rotation of the rotor;

[0016] • Divide the surrounding space into volume units with predetermined geometric shapes, and replace one or more detection points existing in the same volume unit with representative points;

[0017] • The ground is segmented through processing to detect each ground point among representative points that belongs to the ground; the processing includes, for each representative point, detecting whether the normal passing through that representative point forms an angle greater than a predetermined angle threshold with an axis parallel to the rotation axis, the normal being a surface normal including the representative point and adjacent points; and

[0018] • The presence or absence of at least one obstacle is determined by clustering representative points that are different from ground points, and if at least one obstacle is present, a symbolic system showing the obstacle relative to the aircraft is displayed on the screen.

[0019] Therefore, one or more LIDAR obstacle sensing devices enable the acquisition of point clouds (for convenience, they are referred to as "detection points").

[0020] The detected point cloud is then transformed using a method known to those skilled in the art as "voxelization." Each volume unit is typically referred to as a "voxel." All detection points present in a volume unit are replaced by a single point (referred to as a "representative point" for convenience). The size of each volume unit affects the number of points to be processed and thus the time and resources required to detect obstacles. Larger volume units will produce smaller representative point clouds but risk losing fine details (such as cable details), and vice versa. For example, each volume unit has the shape of a cube with sides between 25 cm and one meter, or even between 50 cm and one meter and, for example, 55 cm on each side, in order to achieve good accuracy while limiting the resources required to implement the method.

[0021] The method then includes a step of segmenting the ground to facilitate calculation. Due to the specific nature of rotorcraft and, for example, helicopters with low pitch and roll angles in flight, especially during the flight phase when the aircraft is flying near obstacles, all representative points associated with the normal to the axis of rotation that is substantially parallel to the rotor may belong to the ground.

[0022] Then, one or more obstacles are detected in the usual manner based on all representative points that are not considered to belong to the ground. For example, to detect and identify obstacles, the method can implement clustering algorithms, such as the algorithm known as "HDBSCAN" for "hierarchical density-based spatial clustering for noisy applications" or the Euclidean distance clustering algorithm.

[0023] Then, for example, one or more obstacles detected can be represented in two or three dimensions.

[0024] Therefore, the method of the present invention enables the rapid division of the ground using appropriate computer devices, which can be achieved with potentially lightweight, inexpensive and / or space-saving systems.

[0025] Methods for detecting obstacles may also include one or more of the following features, used individually or in combination.

[0026] One possibility is that the point representing a unit of volume can be the weighted centroid of the detection points of that unit of volume, or the centroid of the detection points of that unit of volume weighted by the light intensity of those detection points, the light intensity being provided by the LIDAR obstacle sensing device.

[0027] According to one possibility compatible with the aforementioned possibilities, the predetermined angle threshold can be equal to 20°.

[0028] Such a threshold enables the achievement of acceptable accuracy.

[0029] According to one possibility compatible with the foregoing, the symbol system may include symbols with colors that vary according to the danger level of an obstacle, which is determined based on the estimated distance between the obstacle and the aircraft or the time of impact.

[0030] Color coding indicating hazard levels, based on distance or time of impact assessed using the current velocity vector and distance, can aid pilots in visual interpretation and decision-making.

[0031] According to a possibility compatible with the foregoing possibilities, if for a representative point the angle is less than or equal to a predetermined angle threshold, then the representative point is a point of interest that may belong to the ground, and the process includes detecting each ground point among the points of interest that belongs to the ground.

[0032] The first filter then includes determining one or more points of interest that may belong to the ground based on the slope of the associated normal.

[0033] Normals can be determined using a conventional method based on decomposition into eigenvalues. For each representative point, neighboring points are identified, and then the centroid of the group of points including the representative point and its neighbors is determined. The centroid and neighboring points are used to determine the covariance matrix. This matrix captures the variation and orientation of these neighboring points relative to their centroid, thus providing information about the local spatial distribution. The eigenvectors of the covariance matrix are then computed, representing the principal directions of variation among neighboring points. The eigenvector associated with the smallest eigenvalue is then the normal being sought.

[0034] According to one possibility compatible with the foregoing possibilities, the neighboring points of the representative point to which the estimated normal points are directed may include all representative points located at a distance less than a predetermined distance from the representative point.

[0035] According to one possibility compatible with the aforementioned possibilities, detecting each ground point among the points of interest may include comparing the height of each point of interest with the average height of the points representing the vicinity. When the height of the point of interest minus the height of each point representing the vicinity is less than a predetermined limit and each point representing the vicinity is also a point of interest, the point of interest is a ground point, and the height of the point of interest is equal to the coordinate of the point of interest along the first axis.

[0036] Each point of interest (POI) is definitively considered to belong to the ground if it meets the following criteria: it meets the height proximity criterion if the POI and its nearest neighbors are substantially at the same height within a tolerance range, and it meets the proximity coherence criterion if the nearby points are also POIs. This method optimizes computation time. Furthermore, it enables accurate and robust identification of points in a point cloud that belong to the ground.

[0037] For example, the predetermined limit can be equal to 1.5 meters.

[0038] For example, the vicinity of a point of interest may include a predetermined number of the closest representative points within a sphere centered on that point of interest.

[0039] For example, the predetermined number is 15. Therefore, as an illustration, the vicinity of the point of interest under study includes the 15 representative points closest to that point of interest, and the study is conducted within a sphere centered on the point of interest under study.

[0040] According to another approach, detecting each ground point may include applying a random consistency algorithm through sampling, which determines at least one parameter of a plane model and each interest point consistent with the plane model, wherein each interest point consistent with the plane model is a ground point.

[0041] Random sampling consensus algorithms are also known as "RANdom Sampling Consensus" and the acronym "RANSAC". This algorithm is an iterative method for estimating the parameters of a predetermined model that best represents an initial point cloud, subtracting a set of outliers from that initial point cloud. Here, the model is an affine plane, and the parameters characterize its orientation and the points belonging to it. In a typical manner, the algorithm iteratively selects a subset of points of a predetermined size from the already filtered points of interest. This selection is performed randomly to estimate the plane model, and after a finite number of iterations, the algorithm selects the so-called optimal plane model that is closest to the random subset from which it is derived. Points that are more than a minimum distance from this plane are then retained as points of interest that do not belong to the ground.

[0042] The present invention also relates to an aircraft having at least one rotor, the aircraft having an obstacle detection system comprising at least one LIDAR obstacle sensing device and a controller and a display, the controller communicating with at least one LIDAR obstacle sensing device and the display.

[0043] The obstacle detection system is configured to implement the above method. Therefore:

[0044] • The at least one LIDAR obstacle sensing device is configured to perform the acquisition of the detected point cloud;

[0045] The controller is configured to perform the following actions: positioning each detection point in a predetermined orthogonal reference frame attached to the aircraft; dividing the surrounding space into volume units; segmenting the ground; and determining the presence or absence of at least one obstacle.

[0046] The display is configured to perform a symbolic representation of the obstacle relative to the aircraft in the presence of at least one obstacle. Attached Figure Description

[0047] The invention and its advantages will become more apparent from the following description of examples given by way of illustration with reference to the accompanying drawings, wherein:

[0048] Figure 1 This is a top view of the aircraft according to the present invention;

[0049] Figure 2 This is a diagram illustrating an obstacle detection system according to the present invention;

[0050] Figure 3 This is a flowchart illustrating the method implemented by such an obstacle detection system;

[0051] Figure 4 This is a diagram showing a display of such an obstacle detection system, illustrating two-dimensional obstacles; and

[0052] Figure 5 This is a diagram of a display showing such an obstacle detection system, illustrating three-dimensional obstacles. Detailed Implementation

[0053] Elements present in more than one figure are given the same reference numerals in each figure.

[0054] Figure 1 An aircraft 1 according to the invention is shown, which is equipped with a rotor 5 capable of rotating about the rotation axis AXROT.

[0055] For example, the aircraft 1 includes a fuselage 2, which extends laterally along its tilt axis from rear to front, from tail 4 to nose 3, from a first side to a second side, and in height from the hull to the apex. The rotor 5 is therefore capable of... Figure 1 It rotates above the fuselage 2.

[0056] Furthermore, the aircraft 1 includes an obstacle detection system 10 for detecting potential obstacles in the surrounding space and, if necessary, signaling them to the pilot. One or more detected obstacles are located in a predetermined orthogonal reference frame 100 attached to the aircraft 1. This reference frame 100 has a first axis Z coinciding with the rotation axis AXROT of the rotor 5, and a second axis X and a third axis Y perpendicular to the first axis Z. The second axis X extends longitudinally from the rear to the front, and vice versa, while the third axis Y extends laterally from one side to the other.

[0057] The obstacle detection system 10 includes an obstacle sensor 20 for detecting obstacles, for example, those that extend more than 360 degrees around a rotation axis AXROT.

[0058] Obstacle sensor 20 can be configured to scan for obstacles in a first volume that covers a 360-degree azimuth angle around the rotation axis AXROT of rotor 5 and has an opening in the height range of +10° above rotor 5 to -20° below rotor 5. Furthermore, obstacle sensor 20 can be configured to cover a second volume surrounding a portion of the first volume, extending, for example, more than 120 degrees in azimuth angle relative to the reference frame, and having an opening in the height range of +10° above rotor 5 to -50° below rotor 5.

[0059] The obstacle sensor 20 can be configured to detect obstacles with speeds up to 30 knots (approximately 55.56 km / h) and a maximum range of 200 meters. Therefore, this obstacle sensor 20 provides the pilot with an acceptable reaction time.

[0060] Therefore, the obstacle sensor 20 includes at least one LIDAR obstacle sensing device 21, 22, 23, 24. Figure 1In the example described, obstacle sensor 20 includes, for example, at least four LIDAR obstacle sensing devices 21, 22, 23, and 24 located near rotor 5. For instance, a first LIDAR obstacle sensing device 21 is carried by the fuselage below the rotor and points towards the front of the aircraft 1, for example, to cover the aforementioned second volume, and can be tilted downwards and forwards relative to the aircraft 1; two LIDAR obstacle sensing devices 22 and 23 are carried by the fuselage below the rotor and point towards volumes covering the respective sides of the two wings; and a fourth LIDAR obstacle sensing device 24 is carried by the fuselage below the rotor and points towards the rear of the aircraft 1. Optionally, the LIDAR obstacle sensing devices 21, 22, 23, and 24 may cover overlapping volumes. References should be made to obtain examples of sensors equipped with at least one LIDAR sensing device.

[0061] Each LIDAR obstacle sensing device 21 can emit light pulses and includes, for example, multiple laser diodes for this purpose.

[0062] refer to Figure 2 The obstacle detection system 10 also includes a controller 15, which includes at least one processing unit. Such a processing unit may include, for example, at least one processor and at least one memory, at least one integrated circuit, at least one programmable system, or at least one logic circuit; these examples do not limit the scope of the term "processing unit." The term "processor" may be used equally well to refer to a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processor (DSP), a microcontroller, etc.

[0063] The controller 15 communicates with the obstacle sensor 20, that is, with one or more obstacle sensing devices 21, 22, 23, 24.

[0064] Furthermore, the controller 15 can be connected to the velocity sensor 31 via wired or wireless means. The velocity sensor 31 may include at least one sensing device for determining the velocity vector of the aircraft 1. For example, the velocity sensor 31 may include a receiver for a satellite positioning device, a navigation system using the Doppler effect, an inertial unit, etc.

[0065] In addition, the controller 15 is connected to the display 25 via wired or wireless means. The display 25 may include a display device 26, such as a screen, helmet visor, eyeglass lenses, head-up collimator, etc.

[0066] The controller 15 may include, for example, a processing computer for implementing the present invention and a symbol generator computer for displaying symbols on a display within one or more processing units. The symbol generator computer 22 may be integrated into the display 25 or remotely. According to one example, the display 25 and the symbol generator computer may form the same device, and the processing computer may be a computer that is dedicated to the methods of the present invention or a computer that is not dedicated to the methods of the present invention.

[0067] Regardless of how the obstacle detection system 10 is implemented, Figure 3 A method for detecting obstacles according to the invention, implemented iteratively by such an aircraft 1, is shown.

[0068] The method includes acquiring a detected point cloud using obstacle sensing devices 20 during step STP1. Each LIDAR obstacle sensing device 21, 22, 23, 24 emits a beam 200. When beam 200 illuminates a point on an obstacle (referred to for convenience as a "detection point PT"), an echo 201 is reflected back to obstacle sensing devices 21, 22, 23, 24. Obstacle sensing devices 21, 22, 23, 24 infer localization data from this, enabling the location of the obstacle in reference frame 100, or even the light intensity of the echo 201. This localization data may include the distance DL between the detection point PT and obstacle sensing device 21, as well as the azimuth and elevation angles in the obstacle sensing device's reference frame.

[0069] The method then includes locating each detection point PT in reference frame 100 during step STP2 using controller 15. This step is then implemented, for example, by executing instructions stored in memory, to configure controller 15.

[0070] These two initial steps, STP1 and STP2, enable the detection points PT detected in the surrounding space to be constructed in a coherent manner.

[0071] The method then includes, during step STP3, using controller 15 to divide the surrounding space into volume units with predetermined geometries. All detection points PT existing within the same volume unit are additionally replaced by representative points.

[0072] Then, controller 15 applies a typical "voxarization" algorithm for this purpose. Then, controller 15 is configured to implement this step, for example, by executing instructions stored in memory.

[0073] According to one possibility, controller 15 calculates the coordinates of the representative point by considering that the representative point of each volume unit is the weighted centroid of the detection points of that volume unit. The coordinates of the representative point are calculated based on the coordinates of the detection points of the volume unit.

[0074] Alternatively, the controller 15 calculates the coordinates of the representative point for each volume unit in the reference frame 100 by considering that the representative point is the centroid of the detection points of that volume unit, which is weighted by the light intensity of the detection points. Therefore, the coordinates of the representative point are calculated based on the coordinates of the detection points of the volume unit, which are weighted by the light intensity of the detection points.

[0075] For example, this method, which uses various geometries such as cubes or parallelepipeds, simplifies the management of a very large number of detection points by balancing computational complexity and accuracy.

[0076] The method then includes segmenting the ground using controller 15. This step is then configured, for example, by executing stored instructions. Controller 15 identifies ground points among the representative points that belong to the flyby area to reduce computation time.

[0077] This segmentation of the ground includes the following process: during step STP4 implemented using controller 15, it is made possible to identify representative points POBS that do not belong to the ground and representative points belonging to the ground, referred to as "ground points".

[0078] During this process, controller 15 is configured to determine points of interest among the representative points during step STP41. The process also includes detecting each ground point belonging to the ground among the points of interest during step STP44.

[0079] To this end, the process includes, for each representative point, during step STP41 implemented using the controller, detecting whether the normal passing through that representative point forms an angle greater than a predetermined angle threshold with an axis parallel to the rotation axis. This normal is the normal to a surface that passes through the representative point and adjacent points. For example, the predetermined angle threshold is equal to 20°. If the normal forms an angle greater than the predetermined angle threshold with an axis parallel to the rotation axis, the representative point does not belong to the ground.

[0080] This filtering, performed by calculating normals, reduces the complexity of the process and quickly eliminates points that do not meet the plane criteria associated with the ground.

[0081] To this end, for each volume unit, controller 15 is configured to determine the normal to the surface passing through the representative point and its adjacent points at the representative point. For each representative point, adjacent points include all representative points located at a distance less than a predetermined distance from the representative point, such as 2 to 4 times the length of one side of the volume unit. Controller 15 then determines the angle at which the normal separates from the axis parallel to the axis of rotation and compares this angle to a predetermined angle threshold. If the angle is not less than or equal to the predetermined angle threshold, then according to arrow N1, the representative point is not a point belonging to the ground but a point that may belong to an obstacle, referred to as an "obstacle point POBS". Conversely, according to arrow Y1, the representative point is a point of interest that could be a ground point PSOL.

[0082] therefore, Figure 3 Two variations for performing step STP44 are described, designed to evaluate whether the point of interest is an obstacle point (POBS) or a ground point (PSOL).

[0083] According to the first variation shown by the continuous line, STP44 for detecting each ground point (PSOL) among the points of interest that belongs to the ground includes comparing the height of each point of interest with the average height of nearby representative points during step STP42. Therefore, controller 15 determines the difference between the height of each point of interest and the average height of nearby points. If this difference is less than a predetermined limit, for example, equal to 1.5 meters, then according to arrow N2, the point of interest is not a ground point (PSOL) but an obstacle point (POBS). Conversely, and according to arrow Y2, the point of interest can be a ground point (PSOL).

[0084] It should be noted that the vicinity of a point of interest includes a predetermined number (e.g., equal to 15) of the nearest representative points within a sphere centered on that point of interest.

[0085] After, simultaneously with, or before step STP42, during step STP43, controller 15 determines whether each point representing a point of interest is also a point of interest satisfying the criteria for the normal described above. If not, and according to arrow N3, the point of interest is not the ground point PSOL but the obstacle point POBS. Conversely, and according to arrow Y3, the point of interest can be the ground point PSOL.

[0086] Therefore, if all three conditions are met, the controller 15 estimates that the representative point is a ground point, namely: i) if the representative point is a point of interest, ii) the height of the representative point is substantially equal to the average height of the nearby representative points, and iii) the nearby representative points are points of interest.

[0087] According to the second variant shown by the dashed line, STP44 for detecting each ground point PSOL includes applying a random consensus algorithm by sampling during step STP45. This random consensus algorithm determines at least one parameter of the plane model and each point of interest consistent with the plane model, which is a ground point PSOL.

[0088] Regardless of the variation, the method includes determining the presence or absence of at least one obstacle based on representative points (i.e., obstacle points POBS) different from the ground during step STP5 implemented using controller 15. The controller 15 is then configured to implement this step, for example, by executing stored instructions.

[0089] Therefore, after ground segmentation, detection points PT that do not meet the ground criteria are clustered by controller 15. This step enables the clear identification and differentiation of potential obstacles in the environment by applying methods known to those skilled in the art.

[0090] Therefore, in the presence of at least one obstacle, the method includes displaying STP6 symbols on the display 25, which represent the obstacle relative to the aircraft 1.

[0091] For example, controller 15 sends a digital signal carrying the information to be displayed to display 25. The result can be displayed in two or three dimensions. In addition, the symbol representing an obstacle can have a color that changes according to the danger level of the obstacle, which is determined based on the impact distance or time between the obstacle and the aircraft 1.

[0092] For example, if an obstacle is located at a distance greater than a first predetermined corresponding value or at an estimated impact time, the obstacle is not considered dangerous; if an obstacle is located at a distance less than or equal to the first predetermined corresponding value and greater than a second predetermined corresponding value or at an impact time, the obstacle is considered moderately dangerous; if an obstacle is located at a distance less than or equal to the second predetermined corresponding value or at an impact time, the obstacle is considered dangerous.

[0093] Figure 4 An implementation method is shown that displays the results in two dimensions using polar coordinates.

[0094] The screen displays an aircraft symbol 60 representing aircraft 1. This aircraft symbol 60 is located at the center of concentric circles divided into corner sectors 65-67. Each corner sector displays a symbol that indicates the presence of an obstacle. If an obstacle is detected in a corner sector, that sector 65-67 may be highlighted based on the estimated hazard of the obstacle. For example, if the obstacle is not considered hazardous, corner sector 65 is colored green; if the obstacle is of moderate hazard, corner sector 66 is colored orange; and if the obstacle is considered hazardous, corner sector 67 is colored red.

[0095] Figure 5 An implementation showing the results in three dimensions is illustrated. Only obstacles are displayed via symbol 68, while the ground is removed or differently colored for better distinction. The symbols can be in the form of dots, with each symbol representing an obstacle point. Arcs 71-75 can be positioned on the ground plane to indicate the distance from the aircraft.

[0096] Naturally, many variations can be made to implement the present invention. Although several embodiments have been described above, it should be readily understood that it is not conceivable to exhaustively identify all possible embodiments. Of course, any of the described devices can be replaced with equivalent devices without departing from the scope of the invention and the claims.

Claims

1. A method for detecting obstacles, implemented by an aircraft (1) having at least one rotor (5), in, The method for detecting obstacles includes: • Use at least one LIDAR obstacle sensing device (21, 22, 23, 24) installed on the aircraft (1) to acquire (STP1) detection point cloud; • Position each detection point in a predetermined orthogonal reference frame (100) attached to the aircraft (1), the reference frame (100) having a first axis (Z) that coincides with the rotation axis (AXROT) of the rotor (5). • Divide the surrounding space (STP3) into volume units with predetermined geometric shapes, and replace one or more detection points existing in the same volume unit with representative points; • The ground is segmented by a process (STP4) to detect each ground point (PSOL) among the representative points; the process (STP4) includes: for each representative point, detecting (STP41) whether the normal passing through the representative point forms an angle greater than a predetermined angle threshold with an axis parallel to the rotation axis, the normal being the normal of the surface including the representative point and adjacent points; and • The presence or absence of at least one obstacle is determined by clustering representative points that are different from the ground point (STP5), and if at least one obstacle is present, a symbol system (65-68) of the obstacle relative to the aircraft (1) is displayed on the display (25) (STP6).

2. The method for detecting obstacles according to claim 1, in, The point representing a unit of volume is either the weighted centroid of the detection points within that unit of volume, or the centroid of the detection points within that unit of volume weighted by the light intensity of these detection points, provided by the LIDAR obstacle sensing device.

3. The method for detecting obstacles according to claim 1, in, The predetermined angle threshold is equal to 20°.

4. The method for detecting obstacles according to claim 1, in, The symbol system has symbols (65, 67) that have colors that vary according to the danger level of the obstacle, which is determined based on the estimated distance or impact time between the obstacle and the aircraft (1).

5. The method for detecting obstacles according to claim 1, in, The neighboring points of the representative point whose normal is estimated include all representative points located at a distance less than a predetermined distance from the representative point.

6. The method for detecting obstacles according to claim 1, in, If for a representative point, the angle is less than or equal to the predetermined angle threshold, then the representative point is a point of interest that may belong to the ground, and the process (STP4) includes detecting (STP44) each ground point among the points of interest that belongs to the ground.

7. The method for detecting obstacles according to claim 6, in, Detecting (STP44) each ground point among the points of interest that belongs to the ground includes: comparing (STP42) the height of each point of interest with the average height of points representing the nearby area; when the height of the point of interest minus the height of each point representing the nearby area is less than a predetermined limit and each point representing the nearby area of ​​the point of interest is also a point of interest, the point of interest is a ground point, and the height of the point of interest is equal to the coordinate of the point of interest along the first axis.

8. The method for detecting obstacles according to claim 7, in, The predetermined limit is equal to 1.5 meters.

9. The method for detecting obstacles according to claim 7, The vicinity of a point of interest includes: A predetermined number of closest representative points within a sphere centered on the point of interest.

10. The method for detecting obstacles according to claim 6, in, The detection (STP44) of each ground point among the points of interest that belong to the ground includes applying (STP45) a random consensus algorithm through sampling, which determines at least one parameter of the plane model and each point of interest that is consistent with the plane model, and each point of interest that is consistent with the plane model is a ground point.

11. An aircraft (1) having at least one rotor (5), the aircraft (1) having an obstacle detection system (10), the obstacle detection system (10) comprising at least one LIDAR obstacle sensing device (21, 22, 23, 24), a controller (15), and a display (25), the controller (15) communicating with the at least one LIDAR obstacle sensing device (21, 22, 23, 24) and the display, in, The obstacle detection system (10) is configured to implement the method according to claim 1: • The at least one LIDAR obstacle sensing device (21, 22, 23, 24) is configured to acquire the detected point cloud; The controller (15) is configured to: locate each detection point in a predetermined orthogonal reference frame attached to the aircraft, divide the surrounding space into volume units, segment the ground, and determine the presence or absence of at least one obstacle; and • The display (25) is configured to display a symbol system in the presence of at least one obstacle, the symbol system indicating the obstacle relative to the aircraft.

Images