SAR (Synthetic Aperture Radar) range compressed domain ship detection method and device based on sub-aperture screening

Abstract

The invention relates to an SAR range compressed domain ship detection method and device based on sub-aperture screening, and the method comprises the steps: carrying out the range compression of ship echo data, arranging the data according to an azimuth time sequence to form a two-dimensional matrix, segmenting the two-dimensional matrix into a plurality of sub-aperture data strips according to the azimuth time sequence and a preset line number, and carrying out the segmentation of the sub-aperture data strips. Performing preliminary screening on azimuth lines in each sub-aperture data bar by using fractional Fourier transform, removing azimuth lines representing non-interested areas, estimating the consistency of adjacent sub-aperture data bars, performing secondary screening on the azimuth lines in the sub-aperture data bars subjected to preliminary screening, and performing secondary screening on the azimuth lines in the sub-aperture data bars subjected to secondary screening; and obtaining a plurality of target sub-aperture data bars representing the ship, and performing focusing according to each target sub-aperture data bar to obtain a plurality of super-aperture images so as to realize ship detection. By adopting the method, ship detection with good robustness and high efficiency can be realized.

Description

Technical Field

[0001] The present application relates to the technical field of SAR image target detection, and in particular to a SAR range compression domain ship detection method and device based on sub-aperture screening. Background Art

[0002] Synthetic Aperture Radar (SAR) is renowned for its all-day, all-weather imaging capabilities. Since the launch of the first spaceborne SAR satellite, SEASAT, it has been widely used for ocean monitoring. SAR is now recognized as an indispensable tool for ship surveillance and is crucial for maritime traffic management, environmental protection, fishery resource management, emergency rescue, coastal defense warning, and other areas. Unlike optical sensors, SAR data requires imaging processing before it can be called an "image." Generally speaking, SAR data consists of four forms, in the order of imaging: raw data, range-compressed data, single-look complex (SLC) products, and image products.

[0003] Among these methods based on different forms of SAR data, ship detection in range-compressed SAR data is the most promising and the preferred method to address the burden of data storage and transmission costs. It has three main advantages: (1) Range compression is a linear time-invariant system with a known correlation kernel. The entire process is very efficient and the range compression time is negligible. (2) The quality of range-compressed SAR data is significantly improved compared to raw SAR data. (3) It is consistent with the acquisition process of the SAR data source. The data on each range line can be detected and processed sequentially, without requiring a large amount of storage space.

[0004] However, the capabilities of range-compressed SAR data remain limited compared to fully focused 2D SAR images. Ship detection mechanisms using range-compressed SAR data have not been fully explored, resulting in lower-than-expected detection performance. For example, some methods may fail to detect ships with low radial velocities, even if the ships are large. Furthermore, deep learning-based detection methods rely heavily on labeled samples and have poor adaptability. When applied to other data sources, detection performance degrades dramatically. Therefore, there is an urgent need to explore new ship detection mechanisms using range-compressed SAR data. Summary of the Invention

[0005] Based on this, it is necessary to provide a SAR range compression domain ship detection method and device based on sub-aperture screening that can achieve robustness and efficient detection in response to the above technical problems.

[0006] A SAR range compression domain ship detection method based on sub-aperture screening, the method comprising:

[0007] Acquiring ship echo data, performing distance compression on the ship echo data, and arranging the data in azimuth and time order to form a two-dimensional matrix;

[0008] According to the azimuth time sequence, the two-dimensional matrix is ​​divided into a plurality of sub-aperture data strips according to a preset number of rows;

[0009] Performing a preliminary screening of the azimuth lines in each of the sub-aperture data strips using a fractional-order Fourier transform to remove azimuth lines representing regions of no interest;

[0010] The consistency of adjacent sub-aperture data strips is estimated, and the azimuth lines in the sub-aperture data strips after the initial screening are secondary screened to obtain multiple target sub-aperture data strips representing the ship;

[0011] Focusing is performed respectively according to each target sub-aperture data strip to obtain multiple super-aperture images to achieve ship detection.

[0012] In one embodiment, when performing distance compression on the ship echo data, the acquired ship echo data is gradually subjected to distance compression as the radar advances.

[0013] In one embodiment, when the two-dimensional matrix is ​​segmented, the number of segmentation rows is determined based on the ability to divide the full-aperture data into three or more sub-aperture data strips.

[0014] In one embodiment, the preliminary screening of the azimuth lines in each of the sub-aperture data strips by using fractional-order Fourier transform includes:

[0015] For each of the sub-aperture data strips, performing fractional Fourier transform on each direction line therein at different range gates to obtain corresponding transformation results;

[0016] The optimal rotation order sequence is estimated according to the transformation results corresponding to each of the direction lines, and the direction lines corresponding to the unstable optimal rotation order sequence are deleted from the sub-aperture data strips to complete the preliminary screening.

[0017] In one embodiment, when the azimuth lines in each sub-aperture data strip are preliminarily screened, sub-aperture data strips without target information are also removed.

[0018] In one embodiment, when screening out the azimuth lines in each of the sub-aperture data strips, the absolute value of the differential operation result of the optimal rotation order sequence is used to judge the stability, and a moving average operation with a sliding window is used in the optimal rotation order sequence to retain the azimuth lines with an absolute value less than a preset threshold, and screen out the azimuth lines with an absolute value greater than or equal to the preset threshold.

[0019] In one embodiment, the estimating the consistency of adjacent sub-aperture data strips and performing secondary screening on the azimuth lines in the sub-aperture data strips after the initial screening includes:

[0020] The azimuth lines with the same optimal rotation order in the adjacent sub-aperture data strips after preliminary screening are retained, and the other azimuth lines with inconsistent optimal rotation orders are removed to obtain the target sub-aperture data strips.

[0021] The present application also provides a SAR range compression domain ship detection device based on sub-aperture screening, the device comprising:

[0022] An echo data acquisition module is used to acquire ship echo data, perform distance compression on the ship echo data, and arrange the data in azimuth and time order to form a two-dimensional matrix;

[0023] a sub-aperture data strip segmentation module, configured to segment the two-dimensional matrix into a plurality of sub-aperture data strips according to a preset number of rows in azimuth time sequence;

[0024] An azimuth line preliminary screening module is used to perform preliminary screening of the azimuth lines in each of the sub-aperture data strips using a fractional-order Fourier transform, and remove azimuth lines representing areas of no interest;

[0025] The target sub-aperture data strip acquisition module is used to estimate the consistency of adjacent sub-aperture data strips, perform secondary screening on the azimuth lines in the sub-aperture data strips after the initial screening, and obtain multiple target sub-aperture data strips representing the ship;

[0026] The ship detection implementation module is used to focus according to each of the target sub-aperture data strips to obtain multiple super-aperture images to achieve ship detection.

[0027] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0028] Acquiring ship echo data, performing distance compression on the ship echo data, and arranging the data in azimuth and time order to form a two-dimensional matrix;

[0029] According to the azimuth time sequence, the two-dimensional matrix is ​​divided into a plurality of sub-aperture data strips according to a preset number of rows;

[0030] Performing a preliminary screening of the azimuth lines in each of the sub-aperture data strips using a fractional-order Fourier transform to remove azimuth lines representing regions of no interest;

[0031] The consistency of adjacent sub-aperture data strips is estimated, and the azimuth lines in the sub-aperture data strips after the initial screening are secondary screened to obtain multiple target sub-aperture data strips representing the ship;

[0032] Focusing is performed respectively according to each target sub-aperture data strip to obtain multiple super-aperture images to achieve ship detection.

[0033] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0034] Acquiring ship echo data, performing distance compression on the ship echo data, and arranging the data in azimuth and time order to form a two-dimensional matrix;

[0035] According to the azimuth time sequence, the two-dimensional matrix is ​​divided into a plurality of sub-aperture data strips according to a preset number of rows;

[0036] Performing a preliminary screening of the azimuth lines in each of the sub-aperture data strips using a fractional-order Fourier transform to remove azimuth lines representing regions of no interest;

[0037] The consistency of adjacent sub-aperture data strips is estimated, and the azimuth lines in the sub-aperture data strips after the initial screening are secondary screened to obtain multiple target sub-aperture data strips representing the ship;

[0038] Focusing is performed respectively according to each target sub-aperture data strip to obtain multiple super-aperture images to achieve ship detection.

[0039] The above-mentioned SAR range compression domain ship detection method and device based on sub-aperture screening performs range compression on the ship echo data, arranges them in azimuth-time order to form a two-dimensional matrix, then divides the two-dimensional matrix into multiple sub-aperture data strips according to a preset number of rows in azimuth-time order. The azimuth lines in each sub-aperture data strip are preliminarily screened using fractional Fourier transform, and the azimuth lines representing areas of no interest are removed. The consistency of adjacent sub-aperture data strips is estimated, and the azimuth lines in the preliminarily screened sub-aperture data strips are secondary screened to obtain multiple target sub-aperture data strips representing the ship. Focusing is performed on each target sub-aperture data strip to obtain multiple super-aperture images to achieve ship detection. This method can achieve robust and efficient ship detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 1 is a flow chart of a SAR range compression domain ship detection method based on sub-aperture screening in one embodiment;

[0041] Figure 2Schematic diagram of time domain and frequency domain of a single-component linear frequency modulation (LFM) signal and white noise in one embodiment;

[0042] Figure 3 FIG4 is a schematic diagram of a fractional Fourier transform (FRFT) result of a single-component linear frequency modulation (LFM) signal in one embodiment;

[0043] Figure 4 FIG4 is a schematic diagram of a fractional Fourier transform of a multi-component linear frequency modulation (LFM) signal in one embodiment;

[0044] Figure 5 Schematic diagram of a geometric model of a translationally moving point target in one embodiment;

[0045] Figure 6 FIG. 1 is a schematic diagram showing a comparison of fractional Fourier transform (FRFT) results of simulated single-component and multi-component linear frequency modulation (LFM) signals based on addition of white noise in one embodiment, wherein: Figure 6 (a) is a schematic diagram showing the results of the fractional Fourier transform (FRFT) of a simulated single-component linear frequency modulation (LFM) signal with white noise added. Figure 6 (b) Schematic diagram showing the fractional Fourier transform (FRFT) results of a multi-component linear frequency modulation (LFM) signal with white noise added;

[0046] Figure 7 1 is a schematic flow chart of a pre-detection imaging method according to an embodiment;

[0047] Figure 8 A flowchart of specific steps for implementing ship detection using the method in one embodiment;

[0048] Figure 9 1 is a structural block diagram of a SAR range compression domain ship detection device based on sub-aperture screening in one embodiment;

[0049] Figure 10 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0051] In order to achieve robust and efficient detection of SAR range compression data, in this application, as Figure 1 As shown, a SAR range compression domain ship detection method based on sub-aperture screening is provided, comprising the following steps:

[0052] Step S100 , obtaining ship echo data, performing distance compression on the ship echo data, and arranging the data in azimuth-time order to form a two-dimensional matrix.

[0053] Step S110 , dividing the two-dimensional matrix into a plurality of sub-aperture data strips according to a preset number of rows in azimuth and time sequence.

[0054] Step S120 , using fractional Fourier transform to preliminarily screen the azimuth lines in each sub-aperture data strip, and remove the azimuth lines representing areas of no interest.

[0055] Step S130 , estimating the consistency of adjacent sub-aperture data strips, performing secondary screening on the azimuth lines in the sub-aperture data strips after the preliminary screening, and obtaining a plurality of target sub-aperture data strips representing the ship.

[0056] Step S140 , focusing is performed according to each target sub-aperture data strip to obtain multiple super-aperture images to achieve ship detection.

[0057] In this application, by analyzing SAR range compression data and fractional Fourier transform (FRFT), the differences between ship clutter and sea clutter in frequency response analysis results are explained, and a method for ship detection based on FRFT and SAR range compression data is proposed.

[0058] Before explaining this method, we first explain why fractional Fourier transform (FRFT) can be used for target detection in SAR range-compressed data. This section first reviews SAR data acquisition modes, focusing on range-compressed SAR data, and then introduces the basic principles of FRFT.

[0059] SAR is a completely man-made device that transmits a linear frequency modulated (LFM) signal in the range direction at a constant frequency, called the pulse repetition frequency (PRF). The LFM signal transmitted by the radar has real-valued characteristics.

[0060] s pul (τ)=ω(τ)cos(2πf0τ+πK r τ 2 ) (1)

[0061] In formula (1), ω(τ) is the range envelope, f0 is the radar center frequency, K r is the linear modulation frequency, and τ is the fast time (distance time).

[0062] After receiving the backscattered signal, orthogonal demodulation is usually used to remove the radar carrier. The demodulated signal of a single point target is expressed in complex value form:

[0063]

[0064] In formula (2), A0 is a complex value (related to the backscatter coefficient of the target), ω a (·) is the azimuth envelope, R(·) is the instantaneous distance, R0(·) is the shortest distance, η is the slow time sampled at the PRF frequency, η c is the beam center crossing time. Since the linear FM signal is a designed signal, it can be easily compressed by a known correlation kernel. The output of range compression is shown below:

[0065]

[0066] In formula (3), p r represents the compressed pulse envelope (it is usually a function similar to the sinc function). In the case of low squint angles, if the aperture is not too large and the point target is stationary, the range equation can be approximately obtained:

[0067]

[0068] In formula (4), V r is the radar speed. Therefore, formula (3) can be rewritten as:

[0069]

[0070] From formula (5), it can be seen that the first phase term remains unchanged, and the second phase term is the second exponential term, that is, a function of 2. Therefore, the signal along the azimuth direction also has the LFM characteristic, such as Figure 1 The azimuth LFM rate is as follows:

[0071]

[0072] In addition, if the azimuth Fourier transform is performed on each range gate or azimuth line, data in the range time domain and azimuth frequency domain can be obtained, which is usually called range-Doppler data.

[0073] From the above analysis, we can see that compared to fully focused 2D SAR images, range-compressed data exhibits certain unique characteristics: First, while range resolution is high, azimuth resolution is extremely low. This creates a significant difference between these two dimensions. Second, due to the imaging geometry, the target is elongated along the azimuth direction, dispersing the energy across different range lines. Third, the lack of texture information makes it difficult to effectively extract features. Consequently, ships in range-compressed SAR data exhibit a low signal-to-clutter ratio and detectability. However, the LFM characteristics along the azimuth direction make ship identification within sea clutter feasible and effective.

[0074] Unlike Fourier transform, FRFT maps the signal into an orthogonal basis function space composed of LFM functions. FRFT has advantages that Fourier transform does not have and has a wide range of application prospects. The FRFT of the x(t) signal is defined as:

[0075]

[0076] In formula (7), K p is the kernel function, expressed as:

[0077]

[0078] In formula (8), n is an integer, α = pπ / 2 is the rotation angle, p is the rotation order of the FRFT, and δ(·) represents the impulse function. When p = 1, the FRFT becomes a Fourier transform; when p = -1, the FRFT becomes an inverse Fourier transform. P does not have to be an integer. Therefore, the FRFT can be considered a generalized form of the Fourier transform, transforming functions to any intermediate domain between time and frequency.

[0079] The single-component LFM signal x′(t)=exp(jkt 2 / 2) are as follows:

[0080]

[0081] The modulus function of formula (9) is expressed as:

[0082]

[0083] Formula (10) is symmetrical with the symmetry axis formula, where the symmetry axis formula is expressed as:

[0084] α=arccot(-k) (11)

[0085] The two sides of the symmetry axis are monotonically decreasing. Therefore, the FRFT results of single-component LFM signals of different orders are symmetrical and unilaterally monotonic.

[0086] In order to estimate the rate of an unknown LFM signal, as Figure 2 As shown, it is necessary to calculate the FRFT results of the signal at different rotation orders. Figure 3 As shown, the signal energy is distributed in two dimensions in the FRFT domain. The optimal rotation order of the signal is determined by searching for peak points on the two-dimensional plane, which can be expressed as

[0087]

[0088] It is worth noting that applying the FRFT with the optimal rotation order to the defocused LFM signal yields a focused signal, namely Figure 3Therefore, when the optimal rotation sequence is obtained, focusing is also completed.

[0089] However, it should be noted that when processing multi-component LFM signals, FRFT will fail due to mutual interference between the multiple components. Figure 4 This is the FRFT domain of the multi-component LFM signal. It is worth noting that there is no obvious peak point in the figure.

[0090] From the above analysis, it can be seen that FRFT is very suitable for analyzing range-compressed SAR data because it is an LFM signal along the azimuth direction.

[0091] In an ideal situation, if the target is stationary and noise-free, the LFM rate of the bearing line can be estimated from the range-compressed data, i.e., K a Combining formulas (6) and (11), we can obtain:

[0092]

[0093] From formula (13), we can see that for a certain SAR system, λ, R0 and V r Therefore, ideally, the estimated optimal rotation order should be unique.

[0094] However, Equation (6) is derived for stationary targets and is inconsistent with reality in two respects. First, real sea conditions and ships are not stationary. Therefore, there is a modulation in the radar velocity. Second, the bearing line used for optimal rotation order estimation is not caused by a single point target. It is a collection of many scattered points. Due to its different structure, scattering, and motion, ships exhibit different characteristics from sea clutter in the time and frequency domains.

[0095] For ocean monitoring over vast oceans, the synthetic aperture time is usually short and the resolution is limited, so the motion is mainly considered to be translational motion. Figure 5 As shown, the point target P is located at (x0,0,0), where x0 is the ground distance of the point target P. The velocity and acceleration in the distance direction are v r and a r The velocity and acceleration in the azimuth direction are v a and a a Therefore, the distance equation can be rewritten to obtain a new distance compression equation:

[0096]

[0097] It is worth noting that when the azimuth LFM rate is alternating, the signal along the azimuth direction also exhibits LFM characteristics:

[0098]

[0099] Because a ship is a rigid body, the velocity of any point on the ship at any given moment is the same, meaning the ship's azimuth signal can be modeled as a single-component LFM signal. More importantly, due to the ship's superior scattering, it can be considered a strong point target. The surrounding sea clutter has limited effect on the ship's amplitude modulation. Therefore, the ship's azimuth signal can be modeled as:

[0100] x ship (t) = A ship exp(jK a,p t 2 / 2)+wn(t) (16)

[0101] In formula (16), A ship is the dominant scattering amplitude of the ship, and wn(t) is the noise. From this, we can draw an important conclusion that the rotation order of the ship's bearing line at different range gates is the same, that is, a constant related to the fixed radar parameters and ship motion parameters.

[0102]

[0103] In addition, it should be noted that due to v a <<V r and Therefore, the effect of ship speed on the LFM rate and the optimal rotation order is quite limited, that is, the estimated result will be an approximate constant determined by (13), but modulated by the ship speed and noise within a limited range. It should also be noted that when processing real discrete SAR data, the estimated result is also related to the PRF and the number of azimuth sampling points.

[0104] However, due to the differences in the structure, scattering and motion of sea clutter, the conclusions for ships do not hold. First, sea clutter is not a rigid body. Therefore, the motion cannot be regarded as a simple translational motion. The azimuth velocity of each scattering point obeys a random distribution. There is evidence that a r Obeying Gaussian distribution, v a It follows a mixed distribution of Rayleigh and Gaussian distributions. This means that there is random phase modulation in the multi-component LFM signal. Secondly, sea clutter has no dominant scattering, and the scattering of each scattering point is comparable. The amplitude distribution of sea clutter usually follows the Rayleigh distribution. Therefore, there is also random amplitude modulation in the multi-component LFM signal. Therefore, the sea clutter signal along the azimuth direction can be modeled as:

[0105]

[0106] In formula (18), N is the number of LFM components, which is a larger number determined by the pixel samples used along the azimuth direction. clutterrepresents the amplitude of the nth LFM signal. It follows the Rayleigh distribution. K a,p is the rate of the nth LFM signal. r Obeying Gaussian distribution, v a A distribution that is a mixture of Rayleigh and Gaussian distributions.

[0107] For this multi-component LFM signal, FRFT cannot estimate the correct result due to the presence of cross terms. Therefore, the optimal rotation order estimated from the actual SAR data will be quite unstable. Therefore, the sea clutter will show completely different characteristics in the time-frequency domain obtained by FRFT. Figure 6 As shown, the optimal rotation orders of the multi-component LFM signals are inconsistent, while the optimal rotation orders of the single-component LFM signals are consistent.

[0108] In summary, Table 1 summarizes the differences between ships and sea clutter in range-compressed SAR data acquired using FRFT. This provides a theoretical basis for ship detection in range-compressed SAR data.

[0109] Table 1 Differences between ship and sea clutter

[0110]

[0111] Furthermore, this method uses a pre-detection imaging step to achieve ship target detection. Pre-detection imaging is a non-imaging paradigm that directly detects ships before two-dimensional focusing, combining target detection with imaging. The process is as follows: Figure 7 shown.

[0112] Returning to this method, in step S100, as the radar advances, ship echo data is gradually collected. Because range compression is a linear time-invariant system with a known correlation kernel, echoes can be range compressed in real time. The collected ship echo data is arranged in azimuth-time order to form a two-dimensional matrix. At this point, this matrix cannot be called an image because it is difficult for the human eye to interpret. However, compared to the original SAR data, the quality of the range-compressed SAR data is significantly improved. Strong targets can usually be distinguished from the range-compressed SAR data.

[0113] In this embodiment, when performing distance compression on the ship echo data, as the radar advances, the acquired ship echo data is gradually subjected to distance compression.

[0114] Furthermore, as the radar advances, when a certain number of range lines are obtained along the azimuth angle, adjacent azimuth lines are detected. Based on the detection mechanism, consistency can be checked by estimating the optimal rotation order sequence. Data with a stable optimal rotation order sequence is identified as a ship.

[0115] When detecting SAR range compression data based on FRFT, radar echo input data is collected incrementally, with the range being compressed as the radar advances. The full-aperture range compression data for a target is not acquired simultaneously. In this embodiment, the two-dimensional matrix is ​​processed in multiple strips, eliminating the need for complete target echo data before detecting the target. According to synthetic aperture imaging theory, each data strip can be considered a portion of the full aperture.

[0116] In step S110, when the two-dimensional matrix is ​​segmented, the number of segmentation rows is determined based on the ability to divide the full aperture data into 3 or more sub-aperture data strips. Figure 8 The received data is arranged into subaperture data strips, as shown in the figure. Therefore, regardless of where a target enters the beam footprint, the target's main echo will be covered by at least two complete subaperture data strips. Since the full aperture length is known based on radar parameters, the subaperture data strip length is fixed for a given SAR system. Slicing the two-dimensional matrix mitigates the effects of range bin shift on LFM signal estimation. Furthermore, this makes it possible to detect the target from the outset.

[0117] It is important to note here that the detected ships in the range compression data are data strips along the azimuth direction, rather than a small patch in the SAR image detection task, because they are not focused in two dimensions.

[0118] What needs to be explained here is that step S100 and step S110 are not two steps with an absolute order of sequence. Since the radar acquisition data is continuous, when processing the radar echo data, after obtaining a certain amount of echo data, distance compression can be performed to obtain a sub-aperture data strip. Therefore, step S100 and step S110 are implemented in an alternating cycle, and after obtaining a sub-aperture data strip, the operation of step S120 can be directly performed.

[0119] Furthermore, in step S120, the azimuth lines in each sub-aperture data strip are screened out, and the process is as follows: Figure 8 As shown in the section Pre-screening along the distance direction.

[0120] In this embodiment, since this method is primarily used for detecting large and medium-sized ships in the open ocean, a single ship will appear in multiple range gates, exhibiting consistent characteristics in the frequency domain. When a subaperture data strip is received, the frequency response Fourier transform of each position line at different range gates is first calculated, and the optimal rotation order is estimated. The consistency of this optimal rotation order is then verified.

[0121] Specifically, the azimuth lines in each of the sub-aperture data strips are preliminarily screened using fractional Fourier transform, including: for each sub-aperture data strip, performing fractional Fourier transform on each directional line at different range gates, using the above formula (8) to obtain the corresponding transformation result, estimating the optimal rotation order sequence based on the corresponding transformation result of each directional line, and using the above formula (12) to delete the directional lines corresponding to the unstable optimal rotation order sequence from the sub-aperture data strip to complete the preliminary screening.

[0122] In this embodiment, during the initial screening of the azimuth lines in each sub-aperture data strip, sub-aperture data strips without target information are removed. If a sub-aperture data strip lacks a stable optimal rotation order sequence, it indicates that there is no target signal in that sub-aperture data strip and that it is entirely sea clutter. This data strip can then be deleted. In effect, the initial screening process can extract signals related to ship targets.

[0123] Furthermore, when screening out the azimuth lines in each sub-aperture data strip, the absolute value of the differential operation of the optimal rotation order sequence is used to judge the stability, and a moving average operation with a sliding window is used in the optimal rotation order sequence to retain the azimuth lines with an absolute value less than a preset threshold, and screen out the azimuth lines with an absolute value greater than or equal to the preset threshold.

[0124] Specifically, when filtering the sub-aperture data strips, the azimuth lines with large deviations are eliminated, and the azimuth lines of the consistent portion where the ship target may be located are retained. The threshold value can be based on the test parameters determined by the SAR system.

[0125] Due to the diversity of the clutter background, some false clutter detections are inevitable in step S120. However, due to decoherence caused by random motion, sea clutter also exhibits different scattering characteristics within adjacent subapertures. As a result, the estimated rotation order of sea clutter can vary significantly. Conversely, the optimal rotation order sequence for ships in adjacent subapertures will be highly consistent. This helps mitigate false positives caused by sea clutter.

[0126] In step S130, the consistency of adjacent sub-aperture data strips is estimated, and the azimuth lines in the sub-aperture data strips after the preliminary screening are secondary screened, including: retaining the azimuth lines with consistent optimal rotation orders in the adjacent sub-aperture data strips after the preliminary screening, and removing other azimuth lines with inconsistent optimal rotation orders to obtain the target sub-aperture data strips.

[0127] In step 140, each target sub-aperture data strip can be focused into multiple super-aperture images using the same optimal rotation order sequence. Afterwards, the super-aperture images can be stitched together in sequence to obtain a complete target focused image according to the requirements of the specific task.

[0128] The aforementioned SAR range-compression domain ship detection method based on subaperture screening proposes a new ship detection mechanism for range-compressed SAR data. Its core concept is that, due to the different structure, scattering, and motion of sea clutter, sea clutter is not in the time-frequency domain, while ships have stable characteristics. Based on this new ship detection mechanism, this method sequentially processes adjacent continuous azimuth data strips and performs target detection based on the characteristic differences captured by the frequency-domain Fourier transform results. This detection method is closely integrated with azimuth focusing of the detected range-compressed SAR data strips. Based on the FRFT parameters estimated during the detection process, it can effectively focus not only stationary ships but also moving ships. This eliminates the unfocusing effect of moving ships. Furthermore, based on clear physical principles, this method has good generalization performance and strong interpretability.

[0129] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. In addition, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0130] In one embodiment, Figure 9 As shown, a SAR range compression domain ship detection device based on sub-aperture screening is provided, comprising: an echo data acquisition module 200, a sub-aperture data strip segmentation module 210, an azimuth line preliminary screening module 220, a target sub-aperture data strip acquisition module 230, and a ship detection implementation module 240, wherein:

[0131] The echo data acquisition module 200 is used to acquire ship echo data, perform distance compression on the ship echo data, and arrange the data in azimuth and time order to form a two-dimensional matrix;

[0132] The sub-aperture data strip segmentation module 210 is configured to segment the two-dimensional matrix into a plurality of sub-aperture data strips according to a preset number of rows in azimuth time sequence;

[0133] An azimuth line preliminary screening module 220 is configured to perform preliminary screening of the azimuth lines in each of the sub-aperture data strips using a fractional-order Fourier transform, and remove azimuth lines representing regions of no interest;

[0134] The target sub-aperture data strip obtaining module 230 is used to estimate the consistency of adjacent sub-aperture data strips and perform secondary screening on the azimuth lines in the sub-aperture data strips after the initial screening to obtain multiple target sub-aperture data strips representing the ship;

[0135] The ship detection implementation module 240 is used to focus according to each of the target sub-aperture data strips to obtain multiple super-aperture images to implement ship detection.

[0136] Regarding the specific definition of the SAR range compression domain ship detection device based on sub-aperture screening, please refer to the definition of the SAR range compression domain ship detection method based on sub-aperture screening above, and will not be repeated here. The various modules in the above-mentioned SAR range compression domain ship detection device based on sub-aperture screening can be implemented in whole or in part through software, hardware, or a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of the above-mentioned modules.

[0137] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 10 As shown. The computer device includes a processor, memory, network interface, display screen, and input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a SAR range compression domain ship detection method based on subaperture screening is implemented. The display screen of the computer device can be a liquid crystal display or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a key, trackball, or touchpad provided on the computer device housing, or an external keyboard, touchpad, or mouse.

[0138] Those skilled in the art will understand that Figure 10 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0139] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0140] Acquiring ship echo data, performing distance compression on the ship echo data, and arranging the data in azimuth and time order to form a two-dimensional matrix;

[0141] According to the azimuth time sequence, the two-dimensional matrix is ​​divided into a plurality of sub-aperture data strips according to a preset number of rows;

[0142] Performing a preliminary screening of the azimuth lines in each of the sub-aperture data strips using a fractional-order Fourier transform to remove azimuth lines representing regions of no interest;

[0143] The consistency of adjacent sub-aperture data strips is estimated, and the azimuth lines in the sub-aperture data strips after the initial screening are secondary screened to obtain multiple target sub-aperture data strips representing the ship;

[0144] Focusing is performed respectively according to each target sub-aperture data strip to obtain multiple super-aperture images to achieve ship detection.

[0145] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0146] Acquiring ship echo data, performing distance compression on the ship echo data, and arranging the data in azimuth and time order to form a two-dimensional matrix;

[0147] According to the azimuth time sequence, the two-dimensional matrix is ​​divided into a plurality of sub-aperture data strips according to a preset number of rows;

[0148] Performing a preliminary screening of the azimuth lines in each of the sub-aperture data strips using a fractional-order Fourier transform to remove azimuth lines representing regions of no interest;

[0149] The consistency of adjacent sub-aperture data strips is estimated, and the azimuth lines in the sub-aperture data strips after the initial screening are secondary screened to obtain multiple target sub-aperture data strips representing the ship;

[0150] Focusing is performed respectively according to each target sub-aperture data strip to obtain multiple super-aperture images to achieve ship detection.

[0151] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0152] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0153] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A SAR range compression domain ship detection method based on sub-aperture screening, characterized by: The method comprises: Acquiring ship echo data, performing distance compression on the ship echo data, and arranging the data in azimuth and time order to form a two-dimensional matrix; According to the azimuth time sequence, the two-dimensional matrix is ​​divided into a plurality of sub-aperture data strips according to a preset number of rows; Performing a preliminary screening of the azimuth lines in each of the sub-aperture data strips using a fractional-order Fourier transform to remove azimuth lines representing regions of no interest; The consistency of adjacent sub-aperture data strips is estimated, and the azimuth lines in the sub-aperture data strips after the initial screening are secondary screened to obtain multiple target sub-aperture data strips representing the ship; Focusing is performed respectively according to each target sub-aperture data strip to obtain multiple super-aperture images to achieve ship detection.

2. The SAR range compression domain ship detection method based on sub-aperture screening according to claim 1 is characterized in that: When performing distance compression on the ship echo data, as the radar advances, the acquired ship echo data is gradually subjected to distance compression.

3. The SAR range compression domain ship detection method based on sub-aperture screening according to claim 2 is characterized in that: When the two-dimensional matrix is ​​segmented, the number of segmentation rows is determined based on the ability to divide the full-aperture data into three or more sub-aperture data strips.

4. The SAR range compression domain ship detection method based on sub-aperture screening according to claim 3 is characterized in that: The preliminary screening of the azimuth lines in each of the sub-aperture data strips by using fractional-order Fourier transform includes: For each of the sub-aperture data strips, performing fractional Fourier transform on each direction line therein at different range gates to obtain corresponding transformation results; The optimal rotation order sequence is estimated according to the transformation results corresponding to each of the direction lines, and the direction lines corresponding to the unstable optimal rotation order sequence are deleted from the sub-aperture data strips to complete the preliminary screening.

5. The SAR range compression domain ship detection method based on sub-aperture screening according to claim 4 is characterized in that: When the azimuth lines in each of the sub-aperture data strips are preliminarily screened, sub-aperture data strips without target information are removed.

6. The SAR range compression domain ship detection method based on sub-aperture screening according to claim 4 is characterized in that: When screening out the azimuth lines in each of the sub-aperture data strips, the absolute value of the differential operation result of the optimal rotation order sequence is used to judge the stability, and a moving average operation with a sliding window is used in the optimal rotation order sequence to retain the azimuth lines with an absolute value less than a preset threshold, and screen out the azimuth lines with an absolute value greater than or equal to the preset threshold.

7. The SAR range compression domain ship detection method based on sub-aperture screening according to claim 6 is characterized in that: The estimating the consistency of adjacent sub-aperture data strips and performing secondary screening on the azimuth lines in the sub-aperture data strips after the preliminary screening includes: The azimuth lines with the same optimal rotation order in the adjacent sub-aperture data strips after preliminary screening are retained, and the other azimuth lines with inconsistent optimal rotation orders are removed to obtain the target sub-aperture data strips.

8. A SAR range compression domain ship detection device based on sub-aperture screening, characterized in that: The device comprises: An echo data acquisition module is used to acquire ship echo data, perform distance compression on the ship echo data, and arrange the data in azimuth and time order to form a two-dimensional matrix; a sub-aperture data strip segmentation module, configured to segment the two-dimensional matrix into a plurality of sub-aperture data strips according to a preset number of rows in azimuth time sequence; An azimuth line preliminary screening module is used to perform preliminary screening of the azimuth lines in each of the sub-aperture data strips using a fractional-order Fourier transform, and remove azimuth lines representing areas of no interest; The target sub-aperture data strip acquisition module is used to estimate the consistency of adjacent sub-aperture data strips, perform secondary screening on the azimuth lines in the sub-aperture data strips after the initial screening, and obtain multiple target sub-aperture data strips representing the ship; The ship detection implementation module is used to focus according to each of the target sub-aperture data strips to obtain multiple super-aperture images to achieve ship detection.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images