Method for improving the time coverage of a beamforming sonar

Abstract

The invention relates to a method for improving the time coverage of a beamforming sonar (1) configured to be mounted on a carrier in an area to be imaged in order to characterise at least one target, the method comprising the steps of: - providing imaging, by coherent processing of the synthetic antenna aperture, of a seabed of the area, along a lateral line of sight (8) which is substantially perpendicular to the direction of movement (10) of the carrier; - providing an anticipated evaluation of the seabed, along a line of sight (9) that is misaligned with respect to the lateral line of sight (8); and - selecting one command among: - reducing the speed of the carrier (24) below a threshold value (VmaxSAS) that is specific to the coherent processing of the synthetic antenna aperture, according to the anticipated evaluation of the seabed, in order to maintain the coherent processing of the synthetic antenna aperture; or - switching from coherent processing to antenna processing that is not sensitive to the speed of the carrier, without reducing the speed below the threshold value (VmaxSAS) that is specific to the coherent processing of the synthetic antenna aperture.

Description

DESCRIPTION Method for improving the hourly coverage of a channel-forming sonar Technical field

[0001] The invention relates to a method for improving the hourly coverage of a beamforming sonar, as well as to a system capable of implementing the method. The invention is particularly, but not exclusively, suitable for sonars embedded in autonomous drones, in the context of mine warfare.

[0002] Underwater imaging systems mainly use synthetic aperture processing to benefit from high longitudinal resolution. The antenna is fixed on a lateral face of a sonar 1 (for example an underwater drone, see figure 1). Since the sonar 1 (more precisely the sonar carrier) moves in a longitudinal direction of movement 2, the same point is illuminated several times (swath 3), which makes it possible to obtain a series of data for each point under the sonar. By combining the variation in amplitude and phase of these returns, synthetic aperture processing makes it possible to obtain images of the observed areas as if a large antenna with very high resolution were used.

[0003] Hourly coverage can be defined as the area covered by the sonar in a given time. Figure 2 illustrates an area 4 (a seabed) covered by a sonar 1 . The sonar path is segmented into a plurality of predetermined lanes 5 (or rails), in order to map the entire area 4.

[0004] A disadvantage of synthetic antenna treatments is that they impose a physical limit on the carrier's advancement. Indeed, for an antenna of length L, the carrier cannot move more than L / 2 between two pings (sound pulses). The following demonstrates that the hourly coverage depends almost exclusively on the length of the physical antenna, for a synthetic antenna treatment.

[0005] Typically, and as illustrated in Figure 3, for a SAS (Synthetic Aperture Antenna) sonar with two sides 1 , i.e. illuminating on both sides of the longitudinal direction of movement, with overlap of the coverage hole under the sonar, the hourly coverage ACR (for "Area Coverage Rate" in English terminology) is:

[0006] ACR = (3P-Am).V / 2

[0007] A corresponds to the sonar altitude; P corresponds to the sonar range; m corresponds to the overlap margin to take into account positioning errors; V corresponds to the forward speed of the carrier.

[0008] SAS processing imposes a limit on the forward speed of the carrier. As a first approximation, this speed can be considered to be:

[0009] V maxSA s=L / (2.Tr)

[0010] L corresponds to the length of the receiving antenna; Tr corresponds to the inter-ping duration, linked to the range P by the following relation: P=c.Tr / 2 (and c, speed of sound in water).

[0011] The hourly coverage is then written:

[0013] Neglecting the margin, as well as the altitude (very low-level and / or long-range cases), we obtain:

[0015] This simplified formula shows that the hourly coverage, in the case of conventional synthetic antenna processing, is exclusively dependent on the length of the physical receiving antenna. Without modifying the sonar, it is therefore not possible to increase the hourly coverage (or only very marginally).

[0016] There are several solutions to increase hourly coverage.

[0017] A first solution consists of increasing the number of carriers in order to mechanically increase the hourly coverage, but with a proportional increase in the cost of the solution. Furthermore, this solution requires a handling a significant number of sonars, as well as complex management of interference between sonars.

[0018] A second solution is to apply a so-called "Fast SAS" processing, such as those described in

[0001] or [2]. This algorithm is also called "undersampled SAS Data" or "sparse SAS data". It consists of increasing the sonar's movement speed beyond the limit speed V maxSAs; the synthetic antenna becomes undersampled, causing grating lobes to appear, which degrade image quality. Fast SAS algorithms therefore aim to reduce these grating lobes. However, grating lobe reduction is not effective enough in prior art solutions.

[0019] Therefore, no software solution works properly. Only hardware solutions would solve the problem of exceeding the speed limit V. ma xSAs ■ But these solutions are not compatible with existing sonars, are expensive and bulky. Their compatibility with underwater drones with a reduced payload is very uncertain.

[0020] Thus, there is a need to provide a method for improving the hourly coverage of a channel-forming sonar, which does not require hardware modifications to existing sonars, and which overcomes the physical limit of the forward speed of the carrier. Summary of the invention

[0021] An object of the invention is therefore a method for improving the hourly coverage of a channel-forming sonar and configured to be embarked on a carrier in an area to be imaged in order to characterize at least one target, the method comprising the steps of: - providing imaging, by coherent synthetic antenna aperture processing, of a seabed in the area, along a lateral line of sight which is substantially perpendicular to the direction of movement of the carrier; - provision of an advance assessment of the seabed, along a line of sight offset from the lateral line of sight; - selection of an order from: - a reduction of the carrier speed below a threshold value specific to the coherent antenna synthetic aperture processing, based on the anticipated assessment of the seabed, in order to maintain the coherent antenna synthetic aperture processing, or - a transition from coherent processing to antenna processing insensitive to the carrier speed, without reducing the speed below the threshold value specific to coherent synthetic antenna aperture processing.

[0022] Advantageously, the antenna processing insensitive to the carrier speed comprises so-called incoherent processing, the incoherent processing comprising: - an estimation of the wearer's movement by a positioning module; - compensation of each reconstructed image based on the wearer's movement, so as to obtain compensated images; - an amplitude summation of the compensated images.

[0023] Advantageously, the transition from coherent processing to incoherent processing is conditioned by the passage of a correlation coefficient below a threshold value, the correlation coefficient being obtained by the correlation of the antenna signals from one ping to another.

[0024] Advantageously, the provision of an advance seabed assessment includes the steps of: - acquisition of measurements along the line of sight offset from the lateral line of sight, so as to obtain a so-called offset image; - detection of the target in the defocused image; - classifying the target according to at least one classification criterion, in order to determine whether the target is an object of interest; and wherein, if the classification is considered to be of good quality, the coherent processing switches to antenna processing insensitive to the carrier velocity, and if the classification is considered to be of poor quality, the carrier velocity is reduced below the threshold value specific to the coherent antenna synthetic aperture processing.

[0025] Advantageously, the classification of the object comprises an assignment of a quality score, the quality score being calculated by the product of a first probability and a second probability, the first probability being specific to the algorithm used for the classification, the second probability depending on the resolution of the defocused image and the dimensions of the target.

[0026] Advantageously, the method comprises calculating a fusion of quality scores, for a plurality of targets detected during the early assessment of the seabed, the control being selected according to the result of the fusion.

[0027] Advantageously, the carrier velocity insensitive antenna processing is performed provided that the anticipated seabed assessment determines that a target is located at a distance less than the resolution of the coherent synthetic aperture antenna processing.

[0028] Advantageously, the carrier velocity-insensitive antenna processing is performed if the anticipated seabed assessment includes detecting a target that has already been imaged by coherent antenna synthetic aperture processing in a rail separate from the current rail in which the carrier is located.

[0029] Advantageously, the method comprises the steps of: - receipt of a mission plan defining at least a first rail and a second rail to be taken by the carrier; - the anticipated assessment including a determination of the distance between the sonar and the target; - the command being executed in the first rail depending on the distance between the sonar and the target in the first rail and in the second rail.

[0030] The invention also relates to a system configured to improve the hourly coverage of a channel-forming sonar and configured to be embarked on a carrier in an area to be imaged in order to characterize at least one target, the system comprising: - an imaging device, by coherent processing of synthetic antenna aperture, of a seabed of the area, along a lateral line of sight which is substantially perpendicular to the direction of movement of the carrier; - a device for early assessment of the seabed, along a line of sight offset from the lateral line of sight; - a device for selecting an order from: - a reduction of the carrier speed below a threshold value specific to the coherent antenna synthetic aperture processing, based on the anticipated assessment of the seabed, in order to maintain the coherent antenna synthetic aperture processing, or - a transition from coherent processing to antenna processing insensitive to the carrier speed, without reducing the speed below the threshold value specific to coherent synthetic antenna aperture processing. Description of figures

[0031] Other characteristics, details and advantages of the invention will emerge from reading the description given with reference to the appended drawings given by way of example.

[0032] Figure 1, already described, illustrates the displacement and swath of a synthetic aperture sonar antenna.

[0033] Figure 2, already described, illustrates the trajectory of a synthetic aperture sonar antenna in a predetermined area.

[0034] Figure 3, already described, illustrates a geometric representation of sonar range and altitude.

[0035] Figure 4 illustrates the lateral boresight axis as well as the off-boresight axis of the sonar.

[0036] Figure 5 illustrates different embodiments of providing an advance assessment of the seabed.

[0037] Figure 6 illustrates the system according to the invention.

[0038] Figure 7 illustrates one embodiment of target classification.

[0039] Figures 8 to 11 illustrate different embodiments for implementing antenna processing insensitive to the carrier speed.

[0040] In Figure 4, a synthetic antenna side-scan sonar 1 is mounted on an underwater carrier. In addition to its conventional side-scan SAS imaging capability 6, the sonar is capable of providing a forward view 7 or "anticipatory" view of the bottom that will be seen later by side imaging. Thus, the sonar is capable of providing imaging, by coherent antenna synthetic aperture processing, of a seabed of the area, along a lateral line of sight 8 which is substantially perpendicular to the direction of movement 10 of the carrier 1. In addition, it can provide an anticipated assessment of the seabed, along an offset line of sight 9 relative to the lateral line of sight.

[0041] In Figure 4, the sonar is of the "single-board" type (the antenna is located only on one side of the carrier); preferably, the sonar can be of the "double-board" type (one antenna on each side of the carrier), which makes it possible to increase the hourly coverage of the area.

[0042] Depending on the result of the front view 7, the method provides either a reduction of the carrier's speed below the limit speed V ma xSAs (coherent processing characteristic), which allows to maintain the coherent processing of synthetic antenna aperture, i.e. a transition from coherent processing to an antenna processing insensitive to the carrier speed, without reducing the speed below the limit speed V max sAs-

[0043] This process allows for optimizing hourly coverage by automatically regulating the sonar speed.

[0044] As illustrated in Figure 5, the advance assessment of the environment can be carried out using one or more forward-looking sonar channels. Several embodiments are possible for achieving this forward view.

[0045] According to a first embodiment, illustrated in the left part of Figure 5, a single transmitter 16 having an electronic depointing capability can be used. This embodiment has the advantage of being compatible with the electronic scanning sonars available on the market. As a variant, illustrated by the central part of Figure 5, the sonar comprises an additional transmitter 12 mechanically depointed forward (fixed angle depointing). According to another variant, illustrated by the right part of Figure 5, a very wide transmitter 13 in emission can be used. A first sub-sector 14 perpendicular to the direction of movement of the carrier is used for imaging by coherent processing, and a second sub-sector 15, directed towards the front of the carrier, is used for the early assessment of the seabed.

[0046] In Figure 4, the depointing angle, which corresponds to the angle between the lateral line of sight 8 and the depointed line of sight 9, can be adjustable. A large angle allows for better anticipation of detections and more reaction time. On the other hand, the image will be of lower quality. An angle between 30° and 45° is a good compromise for good anticipation of detections and good image quality.

[0047] Figure 6 illustrates the system according to the invention. The system capable of implementing the method according to the invention comprises a sonar channel formation block 17, which comprises a “conventional” SAS channel processing unit 18, and an imaging unit 19 which makes it possible to estimate the forward environment in real time.

[0048] The "classic" SAS channel processing unit 18 can be configured, for example, to carry out a mission of detecting and classifying targets of interest, particularly in a mine warfare context.

[0049] The track formation block 17 is capable of providing two solutions that are not or only slightly sensitive to speed.

[0050] The first solution consists of carrying out so-called “natural” channel processing, which only uses the length of the physical antenna.

[0051] Natural processing is a classic channel formation. This formation is carried out, for each desired observation direction (i.e. for each channel) by summing the signals received by the sensors / hydrophones. The signals are previously delayed according to their position in space in order to put them in phase before summing in the desired listening direction.

[0052] A second solution consists of applying a so-called incoherent treatment, for the "broadside" sector (substantially perpendicular to the axis of movement) and / or for the front view sector.

[0053] For this, a positioning module 21 estimates the movement of the carrier 1. Advantageously, the positioning module 21 can be an inertial unit. Inertial units are capable of providing measurements without depending on external sources, and with a high refresh rate.

[0054] The sonar channel formation block 17 performs compensation of each reconstructed image according to the movement of the carrier, so as to obtain compensated images.

[0055] For incoherent SAS processing, the images acquired by the sensor 23 are amplitude integrated. These images must therefore be recalibrated with a precision close to the resolution of the images (generally fixed by the frequency band). To estimate the movements allowing this recalibration, the channel formation block 17 can use the information extracted from the positioning module 21, as described in particular in [3]. Thus, the positioning module 21 is used to estimate the movements necessary for the recalibration of the incoherent processing.

[0056] The "incoherent SAS" processing is particularly advantageous because, in addition to its performance, it is almost native to "classic" SAS processing, namely that it is sufficient to replace a sum of complex images (coherent SAS) by a sum of amplitude images (incoherent SAS).

[0057] The transition from coherent processing to incoherent processing can advantageously be conditioned by the passage of a correlation coefficient / z below a threshold value, the correlation coefficient being obtained by the correlation of the antenna signals from one ping to another. The correlation coefficient / z is therefore calculated by calculating the intercorrelation between a ping and the following ping. Indeed, when the carrier speed increases beyond the limit speed V ma xSAs, no sensor correlates with the other sensors anymore, and the correlation coefficient / z drops significantly. In this case, the images are summed in amplitude.

[0058] As a reminder, for coherent SAS processing, the images must be integrated (summed) in phase. This means that the images must be extremely precisely aligned with each other. For example, for a 300 kHz sonar, the wavelength is 5 mm. To phase align the images, the motion must be compensated with an accuracy that is a fraction of this wavelength, for example 0.5 mm (1 / 10 eme wavelength). This recalibration is carried out by a ping-to-ping cross-correlation method (or “P2C2” for “ping-to-ping cross-correlation”).

[0059] Thus, it is sufficient to base oneself on the correlation coefficient to switch from one mode to another. The channel formation processing will automatically and systematically place itself in the best possible mode (coherent or incoherent). Thus, a speed-robust imaging mode is advantageously created.

[0060] An example of an anticipated assessment of the seabed is illustrated in Figure 7. On this “before” image 25, a detection process is then carried out. The objective of this detection process may be, for example, to detect all objects 26 of mine size. This detection may be carried out by an artificial intelligence type process, or by a shadow detection process, as illustrated in Figure 7. The dimensions of these shadows provide direct information on the width l and the height h of the object, which then makes it possible to verify that these dimensions are compatible with the type of object sought.

[0061] The detections then pass into a classification stage, using ATR (Automatic Target Recognition) or AuDC (Automatic Detection and Classification) type algorithms, well known to those skilled in the art. The classification stage, implemented by the track formation block 17, associates the classification (for example mine / non-mine) with the position of the detection, obtained using the positioning module 21 and the sonar track pointing information calculated in the track formation block 17.

[0062] The system according to the invention, illustrated by Figure 6, also comprises a decision-making autonomy block 20, which has, as input, the detection / classification information provided by the lane formation block 17, and mission plan and / or mission directive information, provided by a command, control and communications system 22 (system called “C3”). The decision-making autonomy block 20 provides the carrier, as output, with a speed instruction, a time of application of the instruction (to manage the acceleration / deceleration constraints of the vehicle), and a duration of application of the instruction.

[0063] To simplify the description, two possible speed instructions are defined: - Vhigh: speed setpoint higher than the V limit max sAs imposed by coherent SAS processing. The actual speed depends on the vehicle's capabilities and the limits imposed by the lobe opening on transmission. - Vlow: speed setpoint lower than the V limit ma xSAs enforced by consistent SAS processing.

[0064] According to one embodiment, the decision-making autonomy block 20 determines whether the classification is considered to be of “good” or “bad” quality. The “good” or “bad” character can advantageously be set according to a quality score n q which is calculated by the product of a first probability P1 and a second probability P2. The first probability is specific to the algorithm used for classification. Thus, the first probability P1 depends on the algorithm used for classification (signal-to-noise contrast, complexity of the object's shape, size of the object, statistical approach, etc.).

[0065] The second probability P2 depends on the resolution of the defocused image and the dimensions of the target. Thus, a two-meter-long cylinder that would be viewed from the side does not need a very fine resolution to be correctly classified. According to one embodiment, the second probability P2 can be based on the Johnson criterion, described in more detail in [4].

[0066] Thus, using the second probability P2, in calculating the quality score n q , allows to take into account the “physical” characteristics of the target, as seen in the anticipated assessment of the seabed.

[0067] The quality grade n q can be compared to a threshold s q : - if n q > s q , then speed setpoint = V h igh - otherwise speed setpoint = V| 0W

[0068] The decision-making autonomy block 20 can also take into account, in making the speed instruction decision, the mission directives from the C3 system 22.

[0069] In particular, the C3 system 22 can transmit, to the decision-making autonomy block 20, a DirA directive to minimize the duration of the mission, at the cost of greater risk (which can be useful in a well-known area, for example a harbor of a port in which it is certain that no mines are located), or a DirB directive to minimize the risk, at the cost of greater mission duration (which can be useful in a war zone).

[0070] Thus, two thresholds can be defined, a first threshold SA associated with the DirA directive, and a second threshold SB associated with the DirB directive, with the following condition: s A < s B -

[0071] The hourly coverage is therefore higher in the case of a DirA directive compared to a DirB directive.

[0072] Figures 8 to 11 illustrate different embodiments of the invention.

[0073] According to a first embodiment, illustrated by Figure 8, the carrier velocity insensitive antenna processing is performed provided that the anticipated seabed assessment determines that a target is located at a distance less than the resolution of the coherent synthetic aperture antenna processing.

[0074] Indeed, depending on the sizing of the sonar, it is possible that at the beginning of the range, the resolution of the incoherent SAS processing, or of the natural SAS processing, is equal to the resolution of the coherent SAS processing. This occurs when a single ping is used at the beginning of the range. On the bottom portion 27, which corresponds to a dip distance ingbetween sonar 1 and the bottom, coherent SAS processing does not provide any resolution gain. Thus, for any detection that occurs at a distance less than di Ping , the speed setpoint can be the higher speed setpoint Vhigh , higher than the limit V ma xSAs-

[0075] In another embodiment, illustrated by Figure 9, the carrier velocity insensitive antenna processing is performed if the anticipated seafloor assessment includes detecting a target that has already been imaged by coherent antenna synthetic aperture processing in a rail separate from the current rail in which the carrier is located.

[0076] Indeed, the planned route for the sonar, composed of rails generally parallel to each other, provides for overlaps between the rails, in order to fill coverage gaps. Since the decision-making autonomy block 20 is aware of the mission plan via the C3 block 22, it is possible to use this plan to improve hourly coverage.

[0077] Thus, the embodiment consists of not taking into account a detection already processed during a previous rail. In Figure 9, if the target 28 has already been observed on rail 1, it is not necessary to slow down during rail 2. Thus, even if it is established that a target 28 is in rail 1, and requires a coherent SAS processing, the decision autonomy block 20 can transmit a command to switch from coherent processing to antenna processing insensitive to the carrier speed, without reducing the speed below the threshold value specific to the coherent antenna synthetic aperture processing.

[0078] Figure 10 illustrates another embodiment. Assuming that, with the anticipated assessment of the seabed, the target 28 was detected at the end of the range during a first pass (first rail 29) with insufficient quality (n q < s q ), it would therefore be necessary to slow down the vehicle (speed setting = V| 0W ), in order to provide imaging through consistent SAS processing.

[0079] However, the decision autonomy block 20, which has the mission plan information transmitted by the C3 block 22, can determine that the target 28 will be reviewed at the start of the range during the next pass (2eme pass 30), and therefore it will be faster to perform imaging there by coherent SAS processing. Thus, during the first pass, the decision-making autonomy block 20 can decide not to slow down, and to keep the speed setpoint at maximum V h igh, greater than the speed limit V maxS As-

[0080] Thus, receiving a mission plan defining at least two rails, and determining the distance between the sonar and the target during the early assessment, makes it possible to increase the sonar's hourly coverage.

[0081] Figure 11 illustrates another embodiment of the method according to the invention. In this figure, targets 31-35 are detected and classified as being of good quality, and target 36 is detected and classified as being of poor quality.

[0082] The method advantageously comprises calculating a fusion of quality scores for the plurality of targets detected during the early seabed assessment. The control (speed reduction or switching from coherent processing to antenna processing insensitive to carrier speed) is then selected based on the result of the fusion. The fusion of scores thus corresponds to a risk associated with the group of targets seen in a time slot by the early seabed assessment.

[0083] This embodiment is more “risky”, because it is not necessarily the lower speed setpoint V| 0W which overrides the speed instruction upper V h igh- However, the implementation method allows for increased hourly coverage.

[0084] With N detections (targets 31-36 in Figure 11), the overall quality is expressed by the following relationship:

[0086] nq (x) corresponds to the quality score n q for detection “x”, which is calculated by the product of a first probability P1 and a second probability P2, as described previously.

[0087] The command selection is expressed as follows:

[0088] if N n x e[i,N] n q( x ) then speed setpoint = V h igh

[0089] otherwise speed setpoint = V| 0W

[0090] s n is a predefined threshold.

[0091] The above embodiments can be combined to increase the hourly coverage of the sonar.

[0092] The method as described makes it possible to optimize the hourly coverage of an existing system by means of decision-making autonomy, while controlling overall performance. The method has the advantage that it does not require the development of a new, specific, expensive and complex sonar system. References cited:

[0093] [1] A Putney, “Reconstruction of Under-sampled SAS Data Using the WIPE Algorithm”, OCEANS 2005

[0094] [2] Shi Yl, “A Method of SAS with Sparse Aperture Based on Compressive Sensing”, IEEE 2020

[0095] [3] Michael P. Hayes et al., Synthetic Aperture Sonar: A Review of Current Status

[0096] [4] F. Langner et al., “Side Scan Sonar Image Resolution and Automatic Object Detection, Classification and Identification”, OCEANS 2009-EUROPE

Claims

CLAIMS 1. Method for improving the hourly coverage of a sonar (1) with channel formation and configured to be embarked on a carrier in an area to be imaged in order to characterize at least one target, the method comprising the steps of: - providing imaging, by coherent synthetic antenna aperture processing, of a seabed of the area, along a lateral sighting axis (8) which is substantially perpendicular to the direction of movement (10) of the carrier; - provision of an advance assessment of the seabed, along a sighting axis offset (9) from the lateral sighting axis (8); - selection of an order from: - a reduction of the carrier velocity (24) below a threshold value (VmaxSAs) specific to the coherent antenna synthetic aperture processing, depending on the anticipated assessment of the seabed, in order to maintain the coherent antenna synthetic aperture processing, or - a transition from coherent processing to antenna processing insensitive to the carrier speed, without reducing the speed below the threshold value (V ma xSAs) specific to coherent synthetic antenna aperture processing.

2. The method of claim 1, wherein the carrier velocity insensitive antenna processing comprises so-called incoherent processing, the incoherent processing comprising: - an estimation of the movement of the carrier by a positioning module (21); - compensation of each reconstructed image as a function of the movement of the wearer (24), so as to obtain compensated images; - an amplitude summation of the compensated images.

3. Method according to one of the preceding claims, in which the transition from coherent processing to incoherent processing is conditioned by the passage of a correlation coefficient below a threshold value, the correlation coefficient being obtained by the correlation of the antenna signals from one ping to another.

4. A method according to any preceding claim, wherein providing an advance assessment of the seabed comprises the steps of: - acquisition of measurements along the defocused sighting axis (9) relative to the lateral sighting axis, so as to obtain a so-called defocused image; - detection of the target in the defocused image; - classifying the target according to at least one classification criterion, in order to determine whether the target is an object of interest; and wherein, if the classification is considered to be of good quality, the coherent processing switches to antenna processing insensitive to the carrier velocity, and if the classification is considered to be of poor quality, the carrier velocity is reduced below the threshold value specific to the coherent antenna synthetic aperture processing.

5. The method of claim 4, wherein the classification of the object comprises an assignment of a quality score (n q ), the quality score being calculated by the product of a first probability (P1) and a second probability (P2), the first probability (P1) being specific to the algorithm used for classification, the second probability (P2) depending on the resolution of the depointed image and the dimensions of the target.

6. Method according to one of claims 4 or 5, comprising a calculation of a fusion of quality scores (n q ), for a plurality of targets (31-36) detected during the early seabed assessment, the control being selected based on the result of the fusion.

7. A method according to any preceding claim, wherein the carrier velocity insensitive antenna processing is performed subject to the anticipated seabed assessment determining that a target is located at a distance less than the resolution of the coherent antenna synthetic aperture processing (dlping)- 8. A method according to any preceding claim, wherein the carrier velocity insensitive antenna processing is performed if the anticipated seabed assessment includes detecting a target that has already been imaged by coherent antenna synthetic aperture processing in a rail distinct from the current rail in which the carrier is located.

9. Method according to one of the preceding claims, comprising the steps of: - receipt of a mission plan defining at least a first rail (R1) and a second rail (R2) to be taken by the carrier; - the anticipated evaluation comprising a determination of the distance between the sonar and the target; the command being executed in the first rail (R1) as a function of the distance between the sonar and the target in the first rail (R1) and in the second rail (R2).

10. System configured to improve the hourly coverage of a channel-forming sonar and configured to be embarked on a carrier in an area to be imaged in order to characterize at least one target, the system comprising: - an imaging device (18), by coherent processing of synthetic antenna aperture, of a seabed of the zone, along a lateral sighting axis which is substantially perpendicular to the direction of movement of the carrier; - a device for early assessment of the seabed (19), along a sighting axis offset from the lateral sighting axis; - a device for selecting a command (20) from: - a reduction of the carrier speed below a threshold value specific to the coherent antenna synthetic aperture processing, based on the anticipated assessment of the seabed, in order to maintain the coherent antenna synthetic aperture processing, or - a transition from coherent processing to antenna processing insensitive to the carrier speed, without reducing the speed below the threshold value specific to coherent synthetic antenna aperture processing.

Images