Innovative drone and aircraft detection system

Abstract

The invention concerns System for detecting drones and aircraft, designed to be installed to defend a site or mobile platform against enemy / hostile drones and aircraft. Said system comprises an image / video acquisition system that includes: N modules (31, 32, 33, 34) arranged at 360° around a vertical central axis (y) of said image / video acquisition system, wherein said vertical central axis (y) is perpendicular to a horizontal plane of said image / video acquisition system, N being an integer greater than one; for each module (31, 32, 33, 34), at least one respective infrared or electro-optical sensor (21, 41, 42, 43, 44, 51) having a given Field of View in azimuth, i.e.parallel to the horizontal plane, and a given Field of View in elevation, i.e. parallel to the vertical central axis (y); for each module (31, 32, 33, 34), at least one respective mirror or prism associated with respective sensor (21, 41, 42, 43, 44, 51) and movable, preferably rotatable, so as to increase the given Field of View in azimuth of said respective sensor (21, 41, 42, 43, 44, 51), so that sensors (21, 41, 42, 43, 44, 51) of all modules (31, 32, 33, 34) provide, altogether, a Field of Regard in azimuth of 360°. The system for detecting drones and aircraft also comprises: a processing system based on artificial intelligence and trained to detect drones and / or aircraft on the basis of the images / video streams provided by the image / video acquisition system; and a control system configured to move the mirrors / prisms and actuate sensors (21, 41, 42, 43, 44, 51) of various modules (31, 32, 33, 34) in a synchronised manner with each other so as to obtain, for the image / video acquisition system as a whole, said Field of Regard in azimuth of 360°.

Description

[0001] INNOVATIVE DRONE AND AIRCRAFT DETECTION SYSTEM

[0002] Cross-Reference to Related Applications

[0003] This Patent Application claims priority from European Patent Application No . 23425050 . 4 filed on September 21 , 2023 , the entire disclosure of which is incorporated herein by reference .

[0004] Technical Sector of the Invention

[0005] The present invention concerns an innovative system designed to detect drones and aircraft and intended to be installed to defend a site or mobi le platform ( e . g . , a land vehicle or naval unit ) against enemy / hostile drones and aircraft .

[0006] Background Art

[0007] Systems of the C-UAS-type ( Counter-Unmanned Aircraft System) , also commonly known as anti-drone systems , can be considered, from a systems engineering point of view, as a system of systems that includes a set of actuators , a set of sensors and a command and control system for said actuators and sensors . Such systems include radars as one of their main sensors . However, under certain operating conditions , radars can cause false alarms , be a factor creating di f ficulties in detection and, being active elements , they can also increase the possibility of being detected by possible opponents .

[0008] In this context , the electro-optical subsystems currently used as sensors in the C-UAS systems operate based on radar designation, i . e . they do not perform any discovery functions .

[0009] Most of the electro-optical and infrared sensors in use today in the C-UAS systems of fer the possibility of on-axis rotation that permits to have a panoramic view in a given time ; this rotation is usually combined with a pan-and-tilt possibility, which allows the elevation thereof to be changed . Such sensors ( typically cameras ) are , therefore , dedicated to spotting a potential drone only after being designated by the radar . They also have a narrow Field of View ( FoV) , typically of the order of 4 ° , and cannot , therefore , spot or track obj ects outside this FoV, nor are they able to spot or track multi-drone attacks from multiple directions .

[0010] The electro-optical and infrared cameras in use in the current C-UAS systems are not capable of performing any discovery functions , as rotation on pan-and-tilt systems , in light of the reduced FoV, would take time that is not compatible with the operating requirements ; moreover, the continuous rotation of the system would cause premature wear of the system itsel f , leading to the need for continuous maintenance .

[0011] Object and Summary of the Invention

[0012] In light of the above , the Applicant felt the need to develop an innovative technology for the C-UAS systems capable of overcoming or alleviating, at least in part , the disadvantages and limitations of the technologies of known type , thus arriving at the present invention .

[0013] Therefore , aim of the present invention is to provide an innovative system for detecting drones and aircraft that can overcome or alleviate , at least in part , the advantages and limitations of currently known technologies .

[0014] This and other aims are achieved by the present invention as it relates to a system for detecting drones and aircraft , as defined in the appended claims .

[0015] In particular, the present invention concerns a system for detecting drones and aircraft , designed to be installed to defend a site or mobile platform against enemy / hostile drones and aircraft . Said system comprises an image / video acquisition system that includes : • N modules arranged at 360 ° around a vertical central axis of said image / video acquisition system, wherein said vertical central axis is perpendicular to a hori zontal plane of said image / video acquisition system, N being an integer greater than one ;

[0016] • for each module , at least one respective infrared ( IR) or electro-optical sensor having a given Field of View ( FoV) in azimuth, i . e . parallel to the hori zontal plane , and a given FoV in elevation, i . e . parallel to the vertical central axis ;

[0017] • for each module , at least one respective mirror or prism associated with the respective sensor and movable , preferably rotatable , so as to increase the given FoV in azimuth of said respective sensor, so that the sensors of all modules provide, altogether , a Field of Regard ( FoR) in azimuth of 360 ° .

[0018] The system for detecting drones and aircraft also comprises :

[0019] • a processing system based on arti ficial intelligence (Al ) and trained to detect drones and / or aircraft on the basis of the images / video streams provided by the image / video acquisition system; and

[0020] • a control system configured to move the mirrors / prisms and actuate the sensors of the various modules in a synchronised manner with each other so as to obtain, for the image / video acquisition system as a whole , said FoR in azimuth of 360 ° .

[0021] Brief Description of the Drawings

[0022] For a better understanding of the present invention, certain preferred embodiment , which are intended purely by way of non-limiting or non-binding examples , will now be described with reference to the accompanying drawings ( all not to scale ) , wherein Figures 1 - 9 schematically show various / di f f erent features of a system for detecting drones and aircraft and, more speci fically, of an image / video acquisition system of said system for detecting drones and aircraft according to preferred embodiment of the present invention .

[0023] Detailed Description of Possible Embodiments of the Invention

[0024] The following description is provided to enable a person skilled in the art to make and use the invention . Various modi fications to the embodiments set forth will be immediately clear to the persons s killed in the art and the general principles herein disclosed may be applied to other embodiments and applications without , however, departing from the protection scope of the present invention as defined in the enclosed claims .

[0025] Therefore , the present invention should not be understood as limited to the sole embodiments described and shown, but it must be given the widest scope of protection in accordance with the characteristics defined in the appended claims .

[0026] The present invention concerns a system for detecting drones and aircraft , designed to be installed to defend a site or mobile platform ( e . g . a land vehicle or naval unit ) against drones ( i . e . Unmanned Aerial Vehicles - UAVs ) and enemy / hostile aircraft . Said system comprises an image / video acquisition system that includes :

[0027] • N modules arranged at 360 ° around a vertical central axis of said image / video acquisition system, wherein said vertical central axis is perpendicular to a hori zontal plane of said image / video acquisition system, N being an integer greater than one ;

[0028] • for each module , at least one respective infrared ( IR) or electro-optical sensor having a given Field of View ( FoV) in azimuth, i . e . parallel to the hori zontal plane , and a given FoV in elevation, i . e . parallel to the vertical central axis ;

[0029] • for each module , at least one respective mirror or prism associated with the respective sensor and movable , preferably rotatable , so as to increase the given FoV in azimuth of said respective sensor, so that the sensors of all modules provide, altogether, a Field of Regard ( FoR) in azimuth of 360 ° .

[0030] The system for detecting drones and aircraft also comprises :

[0031] • a processing system based on arti ficial intelligence (Al ) and trained to detect drones and / or aircraft on the basis of the images / video streams provided by the image / video acquisition system; and

[0032] • a control system configured to move the mirrors / prisms and actuate the sensors of the various modules in a synchronised manner with each other so as to obtain, for the image / video acquisition system as a whole , said FoR in azimuth of 360 ° .

[0033] Conveniently, for each module , the respective mirror / prism is movable , preferably rotatable , so as to increase also the given FoV in elevation of the respective sensor, so that the sensors of all the modules altogether provide a predefined FoR in elevation greater than said FoV in elevation; conveniently, the control system is configured to move the mirrors / prisms and actuate the sensors of the various modules in a synchronised manner with each other so as to obtain, for the image / video acquisition system as a whole , also said predefined FoR in elevation .

[0034] Preferably, the image / video acquis ition system includes a master unit and a slave unit operationally synchronised with each other, wherein said master unit and slave unit each comprise N / 2 modules , and wherein N is an even integer .

[0035] Conveniently, the processing system includes neural networks trained to :

[0036] • analyse the image / video streams provided by the image / video acquisition system;

[0037] • detect the presence of a drone or aircraft in the analysed image / video streams ;

[0038] • confirm the presence of a drone / aircraf t in the image / video streams by analysing the kinematics of the detected obj ects ;

[0039] • calculate a future point of the detected obj ect to enable a correct pointing of kinetic actuators .

[0040] Preferably, the processing system includes neural networks trained to detect , identi fy and classi fy drones and aircraft on the basis of the images / video streams provided by the image / video acquisition system .

[0041] Preferably, the image / video acquisition system includes , for each module , a plurality of sensors comprising at least one infrared sensor and at least one electro-optical sensor .

[0042] As described above , in contrast to current solutions , the system according to the present invention preferably comprises a photocamera / camera that does not move via pan- and-tilt type systems , but is conveniently equipped with a set of optics arranged radially along a circumference for 360 ° coverage around the photocamera / camera . Conveniently, it is also possible to arrange the optics on several circular sectors to achieve 360 ° coverage even in the presence of observation obstructions , keeping these sectors synchronised by means of a synchronisation signal .

[0043] Preferably, the system includes two units , a master unit and a slave unit , wherein each unit comprises a suitable number of modules to cover a 180 ° Field of Regard ( FoR) . One of the two units is equipped with at least one Wide ( i . e . , wide FoV) photocamera / camera placed with the optical axis vertical to the hori zon . By way of example , in the case of four modules for each unit , each o f these would have an FoR of 45°, which can be conveniently obtained, as described below, as the sum of the various FoVs of the mirror in the different positions assumed by the latter as a result of movement .

[0044] Thus, for each electro-optical and / or infrared (IR) sensor :

[0045] FoR = FoV x numero posizioni possibili dello specchio sull'asse x .

[0046] In particular, each module can conveniently include:

[0047] • an IR sensor,

[0048] • an optic compatible with IR wavelength at fixed focal length and variable focus,

[0049] • a stepper motor,

[0050] • a mirror stand,

[0051] • N mirrors placed at different elevation angles (the number of mirrors N being conveniently estimated on the basis of data / predef ined operating requirements) ,

[0052] • an image processing module shared with two camera modules .

[0053] In this respect, Figure 1 schematically shows an example of a master unit 1 and a slave unit 2. In addition, as shown in Figures 2 and 3, a Wide 21 sensor (i.e., with wide FoV) is installed / arranged centred above one of the two system units (in Figures 2 and 3, merely by way of example, slave unit 2) with the optical axis y' that is vertical / orthogonal with respect to the horizontal plane of the system.

[0054] Preferably, a system unit (e.g. master unit 1 or slave unit 2) comprises four (or even more) modules 31, 32, 33, 34 arranged as shown in Figure 4 so as to guarantee that each provides an azimuth coverage of 45° and thus, altogether, 180° .

[0055] Thanks to the use of a rotating mirror system and to a subdivision of the rotation angle into portions, it is possible to generate a 360° image in a limited time. The video streams thus generated are conveniently analysed by arti ficial intelligence (Al ) modules - preferably based on neural networks - specially developed and trained for the detection, recognition and identi fication ( DRI ) of drones and aircraft .

[0056] In particular, the neural networks preferably allow to :

[0057] • recognise the presence of a drone ( or aircraft ) in the analysed streams ;

[0058] • confirm in advance of traditional Al techniques (based on the number of pixels occupied by the obj ect ) the presence of a drone / aircraf t in the video stream by analysing the kinematics of the detected obj ects ;

[0059] • calculate the future point of the detected obj ect to enable the correct pointing of kinetic actuators .

[0060] The aforementioned architecture allows quick replacement of the sensor block and optics , facilitating maintenance tasks and improving overall performance .

[0061] The architecture of the shooting system has been optimised to reduce the number of sensors required, in order to achieve a cost-ef fective configuration . Better results could be obtained by narrowing the FoV, i . e . by increasing the number of chambers or transitions of the mirrors , but this could lead to an increase in the composition time of the 360 ° mosaic . By way of example , the use of 8 high- resolution LWIR / SWIR ("Long Wavelength Infrared" / "Short Wavelength Infrared" ) or MWIR ("Medium Wavelength Infrared" ) sensors to cover the 360 ° side band and of one high- resolution LWIR / SWIR or MWIR sensor together with one high- resolution RGB ("Red Green Blue" ) sensor to cover the vertical areas could be conveniently envisaged . This configuration allows for a lateral band coverage that can be modulated in the elevation angle as a function of the detailed speci fications of the system, as shown in Figure 5 , which shows an example of the lateral shooting angles of a master or slave module of a system with a fixed camera and rotating mirror optics with the characteristics described previously . In particular, Figure 5 shows a master or slave 4 module equipped with 4 LWIR / SWIR or MWIR sensors 41 , 42 , 43 , 44 arranged along a semi-circle so that they each cover an angular sector in azimuth around master / slave module 4 of 45 ° , whereby the 4 LWIR / SWIR or MWIR sensors 41 , 42 , 43 , 44 provide , altogether, a coverage in azimuth of 180 ° . Furthermore , Figure 5 shows , for one of the sensors ( in particular sensor 41 ) , the subdivision of the relevant respective 45 ° sector into three successive 15 ° acquisitions made via an appropriate rotation of the mirrors .

[0062] LWIR / SWIR or MWIR sensors are conveniently connected to one or more GPUs ( Graphics Processing Units ) .

[0063] As explained above , the modular architecture for the various shooting sectors allows both a reduction in costs and greater maintainability of the system .

[0064] A stepper-motor-based motorisation with a multi-axis control system slaved to the data acquisition system can be conveniently used to move the mirrors .

[0065] The timing of the trigger is such that it synchronises the movement of the mirrors and the acquisitions of the chosen video channels .

[0066] The two master and slave units are both equipped with a synchronisation component ( referred to as a Trigger in Figure 1 ) of the mirror rotation motors .

[0067] The synchronisation component reported in Figure 1 keeps the downtimes and the movement times of all mirrors included in the system modules synchronised .

[0068] Conveniently, the inclination of the mirrors in elevation is such that at least 5px of overlap at a distance of 1 km between two contiguous images in elevation are provided .

[0069] In this regard, an example of mirror rotation trigger signals for an RGB sensor and an LWIR sensor is shown in Figure 6 .

[0070] Conveniently, in use , the system receives in input a start acquisition signal and returns images at a fixed update rate with appropriate labelling to be able to reconstruct a mosaic / panorama of images to cover 360 °x30 ° . Scanning is carried out by means of two types of steps of a mirror rotation motor :

[0071] • with a movement of a same mirror in azimuth ( in this case the steps are smaller and equal to m) ,

[0072] • with such a movement as to pass from one mirror to another by rotating between n mirrors tilted di f ferently in elevation .

[0073] Each sensor is therefore used to construct an image matrix equal to m x n images in azimuth and in elevation, respectively . An example of an image generation scheme based on the described architecture is shown in Figure 7 .

[0074] Preferably, for each module the FoV resulting from a scan is equal to 45 ° over at least 30 ° in azimuth and elevation, respectively, with a tolerance of ± 1 ° both in azimuth and in elevation . Each unit of fers images that can cover 180 ° over at least 30 ° of FoV .

[0075] In this respect , Figures 8 and 9 show a master or slave unit 5 wherein an IR sensor provides coverage ( i . e . FoV) of 45 ° in azimuth ( Figure 8 - top view) and of 30 ° in elevation ( Figure 9 - side view) .

[0076] From the above description, the many innovative features of the present invention and its countless technical advantages are immediately obvious to a technician in the field .

[0077] In particular, it is important to emphasise that the present invention makes it possible to improve the ability to discover, recognise and classi fy drones and aircraft over long distances , so as to aid the management of both military and civil operations , such as for example the protection of sensitive sites, like operations centres or command posts, and the protection of moving targets such as motorcades, convoys or individual vehicles.

[0078] The present invention defines an innovative architecture to be able to implement a photocamera / camera that makes it possible to drop the old pan-and-tilt concept, and that through a multi-optics with rotating mirrors / prisms allows a 360° view such that, through the integration of artificial intelligence modules, it is made capable of performing the detection, identification and classification of drones and aircraft with extreme accuracy, even over long distance.

[0079] The configuration of a camera capable of offering a 360° view through a rotating mirror multi-optics, thanks to the addition of specially trained artificial intelligence modules, makes it possible to reduce the DRI (i.e., detection, recognition and identification) time of drones and aircraft, reducing from a technical point of view the risk of wear and tear caused by the rotation and pan-and- tilt mechanisms of the cameras in use today, and from an operational point of view, it allows the system to remain electromagnetically silent, i.e., not to emit since it has no radar by default.

[0080] In conclusion, it is important to note that, while the above described invention refers in particular to very specific embodiments, it must not be intended as limited to such embodiments, including within its scope all the variants, modifications or simplifications covered by the enclosed claims.

Claims

CLAIMS1. System for detecting drones and aircraft, designed to be installed to defend a site or mobile platform against enemy / hostile drones and aircraft; wherein said system comprises an image / video acquisition system that includes:• N modules (31, 32, 33, 34) arranged at 360° around a vertical central axis (y) of said image / video acquisition system, wherein said vertical central axis (y) is perpendicular to a horizontal plane of said image / video acquisition system, N being an integer greater than one;• for each module (31, 32, 33, 34) , at least one respective infrared or electro-optical sensor (21, 41, 42, 43, 44, 51) having a given Field of View in azimuth, i.e. parallel to the horizontal plane, and a given Field of View in elevation, i.e. parallel to the vertical central axis (y) ;• for each module (31, 32, 33, 34) , at least one respective mirror or prism associated with respective sensor (21, 41, 42, 43, 44, 51) and movable, preferably rotatable, so as to increase the given Field of View in azimuth of said respective sensor (21, 41, 42, 43, 44, 51) , so that sensors (21, 41, 42, 43, 44, 51) of all modules (31, 32, 33, 34) provide, altogether, a Field of Regard in azimuth of 360°; and wherein the system for detecting drones and aircraft also comprises:• a processing system based on artificial intelligence and trained to detect drones and / or aircraft on the basis of the images / video streams provided by the image / video acquisition system; and• a control system configured to move the mirrors / prisms and actuate sensors (21, 41, 42, 43, 44, 51) of various modules (31, 32, 33, 34) in a synchronised manner with eachother so as to obtain, for the image / video acquisition system as a whole, said Field of Regard in azimuth of 360°.

2. The system according to claim 1, wherein, for each module (31, 32, 33, 34) , the respective mirror / prism is movable, preferably rotatable, so as to increase also the given Field of View in elevation of respective sensor (21, 41, 42, 43, 44, 51) , so that sensors (21, 41, 42, 43, 44, 51) of all modules (31, 32, 33, 34) altogether provide a predefined Field of Regard in elevation greater than said given Field of View in elevation; and wherein the control system is configured to move the mirrors / prisms and actuate sensors (21, 41, 42, 43, 44, 51) of various modules (31, 32, 33, 34) in a synchronised manner with each other so as to obtain, for the image / video acquisition system as a whole, also said predefined Field of Regard in elevation.

3. The system according to claim 1 or 2, wherein the image / video acquisition system includes a master unit (1) and a slave unit (2) operationally synchronised with each other, wherein said master unit (1) and slave unit (2) each comprise N / 2 modules (31, 32, 33, 34) , and wherein N is an even integer.

4. The system according to any one of claims 1 to 3, wherein the processing system includes neural networks trained to:• analyse the image / video streams provided by the image / video acquisition system;• detect the presence of a drone or aircraft in the analysed image / video streams;• confirm the presence of a drone / aircraf t in the image / video streams by analysing the kinematics of the detected objects;• calculate a future point of the detected object to enable a correct pointing of kinetic actuators.

5. The system according to any one of the precedingclaims, wherein the processing system includes neural networks trained to detect, identify and classify drones and aircraft on the basis of the images / video streams provided by the image / video acquisition system.

6. The system according to any one of the preceding claims, wherein the image / video acquisition system includes, for each module (31, 32, 33, 34) , a plurality of sensors (21, 41, 42, 43, 44, 51) comprising at least one infrared sensor and at least one electro-optical sensor.

7. The system according to any one of the preceding claims, where the system is installed on board a land vehicle or naval unit.

8. Platform equipped with the system for detecting drones and aircraft as claimed in any preceding claim.

9. The platform according to claim 8, wherein said platform is a mobile platform, conveniently a land vehicle or a naval unit.

Images