Miniature sar imaging method, system and medium

Abstract

The invention discloses a miniature sar imaging method and system and a medium, and relates to the technical field of synthetic aperture radar imaging, and the method comprises the steps: configuring a multi-channel receiving array for a target miniature platform, and selecting miniature sar radar sensors for integration; radar echo signals are received, and a multi-channel radar echo signal sequence array is obtained; carrying out intra-sequence disparity separation identification; carrying out homomorphic low-pass filtering processing and heteromorphic difference retention processing; performing intra-sequence information fusion to obtain a multi-channel region sar image set; and carrying out image feature multi-channel cross enhancement fusion to obtain a target area sar image. According to the invention, the technical problems of insufficient processing precision and low imaging resolution of multi-channel radar echo signals under a micro platform in the prior art are solved, and the technical effects of realizing accurate processing and fusion of the multi-channel radar echo signals under the micro platform and improving the sar imaging resolution of a target area are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of synthetic aperture radar imaging, and particularly relates to a micro sar imaging method, system and medium. BACKGROUND

[0002] As an active microwave remote sensing device, synthetic aperture radar (SAR) has the ability to work all day and all weather, and is widely used in the fields of environmental monitoring, disaster assessment, military reconnaissance, etc. With the rapid development of micro platforms such as unmanned aerial vehicles and small satellites, micro SAR systems have become a research hotspot due to their advantages of flexible deployment and low cost. However, due to the load limitation of micro platforms, existing micro SAR systems usually face problems such as small antenna size, low transmit power, and limited number of channels, which leads to the radar echo signal being easily disturbed by noise and the matching precision of multi-channel signals being insufficient, ultimately affecting the imaging resolution and image quality, and making it difficult to meet the high-precision observation requirements.

[0003] The prior art has the technical problems of insufficient processing precision of multi-channel radar echo signals under a micro platform and low imaging resolution. SUMMARY

[0004] The present application provides a micro SAR imaging method, system and medium, which is used to solve the technical problems of insufficient processing precision of multi-channel radar echo signals under a micro platform and low imaging resolution in the prior art.

[0005] In view of the above problems, the present application provides a micro SAR imaging method, system and medium.

[0006] In a first aspect of the embodiments of the present application, a micro SAR imaging method is provided, and the method comprises: The target micro-platform is configured with a multi-channel receiving array, and a micro-sar radar sensor is selected for integration; the target area is scanned according to a preset scanning path by the integrated micro-sar radar, and a multi-channel radar echo signal sequence array is obtained by parallel receiving of the radar echo signals through the multi-channel receiving array; according to the signal multi-dimensional features of each multi-channel radar echo signal sequence in the multi-channel radar echo signal sequence array, sequence-intra heterogeneous separation identification is performed on each multi-channel radar echo signal sequence, and a multi-channel identified radar echo signal sequence array is obtained, wherein the identification corresponding to each multi-channel identified radar echo signal is homogeneous signal identification and heterogeneous signal identification; based on the homogeneous signal identification and the heterogeneous signal identification, a signal processing double channel is called to perform homogeneous low-pass filtering processing and heterogeneous difference reservation processing on the multi-channel identified radar echo signal sequence array, and a multi-channel identified filtered radar echo signal sequence array is obtained; through a synthetic aperture radar processing method, sequence-intra information fusion is performed on the multi-channel identified filtered radar echo signal sequence array, and a multi-channel regional sar image set is obtained; and image feature multi-channel cross enhancement fusion is performed on the multi-channel regional sar image set, and a target area sar image is obtained.

[0007] In a second aspect, the embodiment of the present application provides a micro-sar imaging system, which comprises: An array configuration module is configured to configure a multi-channel receiving array for a target micro-platform, and select a micro-sar radar sensor for integration; an echo signal receiving module is configured to scan a target area according to a preset scanning path by an integrated micro-sar radar, and obtain a multi-channel radar echo signal sequence array by parallel receiving of radar echo signals through the multi-channel receiving array; a heterogeneous separation identification module is configured to perform sequence-intra heterogeneous separation identification on each multi-channel radar echo signal sequence according to the signal multi-dimensional features of each multi-channel radar echo signal sequence in the multi-channel radar echo signal sequence array, and obtain a multi-channel identified radar echo signal sequence array, wherein the identification corresponding to each multi-channel identified radar echo signal is homogeneous signal identification and heterogeneous signal identification; an echo signal sequence array acquisition module is configured to call a signal processing double channel to perform homogeneous low-pass filtering processing and heterogeneous difference reservation processing on the multi-channel identified radar echo signal sequence array based on the homogeneous signal identification and the heterogeneous signal identification, and obtain a multi-channel identified filtered radar echo signal sequence array; a sar image set acquisition module is configured to perform sequence-intra information fusion on the multi-channel identified filtered radar echo signal sequence array through a synthetic aperture radar processing method, and obtain a multi-channel regional sar image set; and an image fusion module is configured to perform image feature multi-channel cross enhancement fusion on the multi-channel regional sar image set, and obtain a target area sar image.

[0008] In a third aspect, the present application provides a computer readable storage medium storing a computer program for executing the micro-sar imaging method.

[0009] The one or more technical solutions provided in the present application have at least the following technical effects or advantages: The target micro-platform is configured with a multi-channel receiving array, and a micro-sar radar sensor is selected for integration. The target area is scanned according to a preset scanning path by the integrated micro-sar radar, and the multi-channel receiving array is used to receive radar echo signals in parallel to obtain a multi-channel radar echo signal sequence array. According to the signal multi-dimensional features of each multi-channel radar echo signal sequence in the multi-channel radar echo signal sequence array, intra-sequence hetero-separation identification is performed on each multi-channel radar echo signal sequence to obtain a multi-channel identified radar echo signal sequence array. Based on homo-signal identification and hetero-signal identification, a signal processing dual-channel is called to perform homo-low-pass filtering processing and hetero-difference reservation processing on the multi-channel identified radar echo signal sequence array to obtain a multi-channel identified filtered radar echo signal sequence array. Through a synthetic aperture radar processing method, intra-sequence information fusion is performed on the multi-channel identified filtered radar echo signal sequence array to obtain a multi-channel regional sar image set. Image feature multi-channel cross-enhancement fusion is performed on the multi-channel regional sar image set to obtain a target area sar image. The technical effect of realizing accurate processing and fusion of multi-channel radar echo signals under a micro-platform is achieved, and the target area sar imaging resolution is improved. BRIEF DESCRIPTION OF DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0011] Figure 1 A flowchart of a micro-sar imaging method provided by an embodiment of the present application.

[0012] Figure 2 A structure diagram of a micro-sar imaging system provided by an embodiment of the present application.

[0013] Legend of the drawings: array configuration module 10, echo signal receiving module 20, hetero-separation identification module 30, echo signal sequence array acquisition module 40, sar image set acquisition module 50, image fusion module 60. DETAILED DESCRIPTION

[0014] The application provides a micro SAR imaging method, system and medium, which are used to solve the technical problems of insufficient processing precision and low imaging resolution of multi-channel radar echo signal under a micro platform in the prior art.

[0015] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the application.

[0016] Embodiment one, as shown in the figure, the application provides a micro SAR imaging method, the method comprises: Figure 1 Step S100: configuring a multi-channel receiving array for a target micro platform, and selecting a micro SAR radar sensor for integration. Step S200: scanning a target area according to a preset scanning path by the integrated micro SAR radar, and receiving radar echo signals in parallel by the multi-channel receiving array to obtain a multi-channel radar echo signal sequence array.

[0017] Specifically, first, the specific type and operating parameters of the target micro platform are determined, the specifications of the multi-channel receiving array are determined according to the weight limit, the space size that can be carried, and the actual imaging scene demand of the platform, including the number of antenna units, the arrangement mode and the signal receiving bandwidth, to ensure that the array can be installed in the limited space of the platform and meet the spatial resolution requirements of subsequent imaging. Then, the micro SAR radar sensor that meets the adaptation standard of the micro platform is selected, and the volume, power consumption, weight and signal output stability of the sensor are mainly considered to ensure that it is compatible with the power supply system and data transmission interface of the target micro platform. Subsequently, the selected multi-channel receiving array and the micro SAR radar sensor are integrated, the circuit connection between them is completed to realize signal transmission and synchronous control, and calibration testing of the integrated system is carried out to ensure that each unit of the multi-channel receiving array can accurately receive the echo of the radar sensor after the transmitted signal is reflected by the target, forming a cooperative micro SAR radar hardware system, which lays a hardware foundation for subsequent target area scanning and echo signal acquisition.

[0018] Step S200: scanning a target area according to a preset scanning path by the integrated micro SAR radar, and receiving radar echo signals in parallel by the multi-channel receiving array to obtain a multi-channel radar echo signal sequence array.

[0019] Specifically, in the micro-sar imaging process, after hardware integration, the integrated micro-sar radar carries out systematic scanning on the target area according to the pre-planned scanning path. During scanning, the radar actively emits electromagnetic waves, and when the electromagnetic waves contact various objects in the target area, they produce echoes. The multi-channel receiving array works synchronously and captures these echo signals in parallel. Different channels receive echoes: 1) amplitude information, reflecting the target reflection intensity, such as the strong echo amplitude of a metal surface and the weak echo of a smooth or distant target; 2) phase information, representing the relative position of the target and the radar, used to calculate the distance; 3) Doppler shift, speed information, generated by the relative motion of the target, which can determine the relative speed and is crucial for monitoring dynamic targets such as aircraft and vehicles; 4) spatial direction information, which helps determine the target azimuth and elevation angles, and multi-channel reception from different angles improves the spatial resolution of the image. These echo signals are arranged in time sequence and channel sequence, ultimately forming a multi-channel radar echo signal sequence array, providing raw data containing rich target features for subsequent signal processing.

[0020] Step S300: According to the signal multi-dimensional features of each multi-channel radar echo signal sequence in the multi-channel radar echo signal sequence array, perform intra-sequence hetero-separation identification of each multi-channel radar echo signal sequence, and obtain a multi-channel identification radar echo signal sequence array, wherein each multi-channel identification radar echo signal corresponds to a same signal identification and a hetero signal identification.

[0021] Specifically, based on the obtained multi-channel radar echo signal sequence array, the signal multi-dimensional features of each sequence are extracted, including target reflection intensity reflected by echo amplitude, target relative position represented by phase, target relative speed associated with Doppler shift, and target azimuth and elevation angles determined with the aid of spatial direction information. Then, according to these multi-dimensional features, the signals within a single sequence are finely discriminated. Signals with background region characteristics and dominated by low-frequency information are marked as same signals and assigned a same signal identification; signals presenting target region characteristics and dominated by high-frequency information are marked as hetero signals and added with a hetero signal identification. By performing the above operations on all sequences in the multi-channel radar echo signal sequence array, a multi-channel identification radar echo signal sequence array is finally integrated and formed. This array can be used as input data for subsequent synthetic aperture radar imaging, distinguishing between background and target, and providing a foundation for improving imaging accuracy and enhancing target detection and resolution.

[0022] Step S400: Based on the same signal identification and the hetero signal identification, a signal processing dual-channel is called to perform same low-pass filtering processing and hetero difference reservation processing on the multi-channel identification radar echo signal sequence array, and a multi-channel identification filtered radar echo signal sequence array is obtained.

[0023] Specifically, based on the output of the multi-channel identification radar echo signal sequence array, for the sequence with the same signal identification, the low-pass filter channel is called, the Butterworth low-pass filter algorithm is used, the cutoff frequency is set according to the signal spectrum characteristics, the low-frequency signal reflecting the stable characteristics of the target such as the reflection amplitude of the static terrain and the fixed position phase information is allowed to pass smoothly, and at the same time, the high-frequency noise is attenuated, so that the same signal is more pure while retaining the basic characteristics; for the sequence with different signal identification, the difference reservation channel is triggered, the singular value decomposition method is used to decompose the matrix constructed by the different signal, and the characteristic vector corresponding to the larger singular value is extracted to strengthen the key information different from the same signal such as the unique Doppler shift of the dynamic target and the spatial direction difference of the special target, and the difference is highlighted. After the dual-channel processing is completed, all sequences are integrated to obtain a multi-channel identification filtered radar echo signal sequence array, which provides optimized and more clear characteristic signal support for subsequent imaging.

[0024] Step S500: Through the synthetic aperture radar processing method, the multi-channel identification filtered radar echo signal sequence array is traversed for sequence information fusion, and a multi-channel regional sar image set is obtained.

[0025] Specifically, the synthetic aperture radar processing method is used to traverse and process the obtained multi-channel identification filtered radar echo signal sequence array, for each multi-channel identification filtered radar echo signal sequence in the array, first, distance compression processing is performed, the size of the echo signal in the range direction is compressed through the matched filter algorithm, the signal energy of the target in the range direction is focused, and the range resolution is improved; then, azimuth focusing processing is carried out, the Doppler shift information is used to correct the azimuth phase of the signal, the phase error caused by the platform motion is eliminated, the signal energy of the target in the azimuth direction is gathered, the imaging clarity is further optimized, and a multi-channel identification filtered preprocessed radar echo signal sequence array is obtained. Subsequently, the sequence information of each multi-channel identification filtered preprocessed radar echo signal sequence is spliced and fused, the signal fragments collected at different times and different positions in the sequence are integrated according to the time and space correlation, the information gap of the single signal is made up, and a complete sequence fusion signal is formed. Then, the one-dimensional echo signal after fusion is converted into a two-dimensional space image by the back projection method, each multi-channel identification filtered radar echo signal sequence corresponds to generate a sar image covering a specific region, and the sar images generated by all sequences are summarized to obtain a multi-channel regional sar image set.

[0026] Step S600: Image feature multi-channel cross enhancement fusion is performed on the multi-channel regional sar image set to obtain a target regional sar image.

[0027] Specifically, the image feature extractor is used to extract features of each image in the multi-channel regional SAR image set respectively, and a multi-channel regional SAR image feature set containing image texture details, gray distribution, edge contour and scattering intensity information is obtained. Then, all image features in the feature set are enumerated and combined two by two to form a multi-channel regional SAR image feature combination set. For each feature combination in the set, image feature similarity analysis is first carried out, the similarity coefficients between the features are calculated to form an image feature similarity coefficient set, and then the similarity coefficient set is subjected to softmax processing to construct a cross-enhancement fusion matrix according to the processing result. Based on the matrix, each image feature in the feature combination is enhanced respectively to obtain the corresponding cross-enhancement fusion feature combination, and all combinations are summarized as a multi-channel regional SAR image cross-enhancement fusion feature combination set. Subsequently, the original multi-channel regional SAR image feature set is used as an index to calculate the mean value of the multiple enhanced fusion features corresponding to the same image feature in the cross-enhancement fusion feature combination set to generate a multi-channel regional SAR image cross-enhancement fusion feature mean value set. According to the mean value set, the multi-channel regional SAR image set is subjected to inversion image enhancement to improve the detail definition and contrast of the image, and an enhanced multi-channel regional SAR image set is obtained. Finally, the enhanced image set is subjected to image mean filtering to eliminate local noise interference, and finally a target regional SAR image that can clearly present the features of the target regional ground objects without over-smoothing problem is obtained.

[0028] In one possible implementation manner, the step S300 further includes: The step S310: according to a preset signal multi-dimensional feature, feature extraction is performed on each multi-channel radar echo signal of the first multi-channel radar echo signal sequence in the multi-channel radar echo signal sequence array, and a first multi-channel radar echo signal multi-dimensional feature sequence is obtained, wherein the preset signal multi-dimensional feature includes a frequency domain feature, a time domain feature, a Doppler shift feature and a spatial direction feature.

[0029] The step S320: a multi-layer perception same matching prototype library is obtained, each multi-channel radar echo signal multi-dimensional feature in the first multi-channel radar echo signal multi-dimensional feature sequence is subjected to similarity identification with each multi-layer perception same matching prototype in the multi-layer perception same matching prototype library, when a similarity degree greater than or equal to a preset similarity identification result threshold exists in the similarity identification result, a corresponding multi-channel radar echo signal is subjected to same matching signal identification, and a first multi-channel same matching signal identification radar echo signal sub-sequence is obtained.

[0030] The step S330: when a similarity degree greater than or equal to a preset similarity identification result threshold does not exist in the similarity identification result, a corresponding multi-channel radar echo signal is subjected to different matching signal identification, and a first multi-channel different matching signal identification radar echo signal sub-sequence is obtained.

[0031] Step S340: Fuse the radar echo signal sub-sequences identified from the first multi-channel homochromatic signal and the first multi-channel heterochromatic signal in time from front to back, and add the first multi-channel identified radar echo signal sequence obtained by fusion into the multi-channel identified radar echo signal sequence array.

[0032] Specifically, for the multi-channel radar echo signal sequence array, the "first multi-channel radar echo signal sequence" here is not limited in sequence, but refers to any sequence in the array. In processing, according to the pre-defined signal multi-dimensional feature system containing frequency domain features, time domain features, Doppler shift features and spatial direction features, feature extraction operation is carried out on each multi-channel radar echo signal in the selected sequence, Fourier transform is used to analyze the signal spectrum, the time domain signal is converted to the frequency domain, the frequency domain features are extracted, thereby distinguishing the stable and low-frequency dominant homochromatic signal features and the significantly changing and high-frequency prominent heterochromatic signal features; the signal time domain waveform is observed, the time domain features such as the change rate of signal amplitude with time and the amplitude fluctuation are captured, thereby distinguishing the difference between dynamic targets and static background; based on the principle that the relative motion of dynamic targets causes the change of signal frequency, the Doppler shift features are extracted, and the frequency change caused by dynamic targets is identified; combined with the distribution situation of targets in space, including the moving track direction, the position and the pitch angle, etc., the spatial direction features are extracted, and the difference in spatial distribution of dynamic targets and static targets is distinguished. Through the extraction of these dimensional features of each signal in the sequence, the original echo signal is converted into a multi-dimensional feature sequence of multi-channel radar echo signal containing rich information, which provides accurate and comprehensive feature basis for subsequent processes such as distinguishing homochromatic signal and heterochromatic signal, and carrying out similar identification.

[0033] A multi-layer perception same-proto library is acquired, which is a standard feature set generated by training a multi-layer perception model, aggregating features and extracting prototypes from a large amount of background area, i.e. radar echo data of typical scenes of same-proto signals, covering typical feature patterns of background area echo signals in the dimensions of frequency domain, time domain, Doppler shift, spatial direction, etc. Then, for the extracted multi-dimensional feature sequence of the first multi-channel radar echo signal, each multi-dimensional feature of the multi-channel radar echo signal is sequentially taken out and compared with each multi-layer perception same-proto in the multi-layer perception same-proto library for similarity identification, and the matching degree of the two is determined by calculating the distance between the features, such as Euclidean distance or cosine distance. If the similarity identification result of a certain multi-dimensional feature and any one of the multi-layer perception same-protos reaches or exceeds the pre-set similarity identification result threshold, it means that the echo signal has the same-proto signal characteristics of the background area dominated by low-frequency information, and the corresponding multi-channel radar echo signal is marked as a same-proto signal. All the echo signals marked as same-proto signals are arranged in the original sequence order to form a first multi-channel same-proto signal marked radar echo signal sub-sequence, which provides a clear signal classification basis for distinguishing background and target and improving imaging accuracy in subsequent processes.

[0034] For each feature in the first multi-channel radar echo signal multi-dimensional feature sequence, if the similarity identification result of the feature and all multi-layer perception same-protos in the multi-layer perception same-proto library does not reach the pre-set similarity identification result threshold, it means that the multi-channel radar echo signal corresponding to the feature does not have the typical characteristics of the background area same-proto signal, but belongs to the signal type of the target area dominated by high-frequency information. At this time, the signals of this type are marked as different-proto signals, and all the echo signals marked as different-proto signals are integrated according to their time sequence in the original sequence to obtain a first multi-channel different-proto signal marked radar echo signal sub-sequence, which provides clear classification signal data for distinguishing target and background and further imaging processing in subsequent processes.

[0035] The two sub-sequences are spliced and fused according to the time sequence of signal generation to ensure that the original order of same-proto signals and different-proto signals on the time axis is preserved to form a complete first multi-channel marked radar echo signal sequence containing background and target identification. Then, the fused sequence is added to the multi-channel marked radar echo signal sequence array to enrich the data content of the array, provide signal input with clear background-target identification for subsequent applications based on the array, such as synthetic aperture radar imaging, and help improve the accuracy of background and target differentiation in imaging.

[0036] In one possible implementation, step S320 further includes: Step S321: Obtain a plurality of sample same-signal-identified radar echo signal sets and a plurality of sample same-signal-identified radar echo signal multi-dimensional feature sets corresponding to the plurality of sample same-signal-identified radar echo signal sets, and perform same-class aggregation on the plurality of sample same-signal-identified radar echo signal sets according to a preset aggregation scale. According to the same-class aggregation result, perform mapping aggregation on the plurality of sample same-signal-identified radar echo signal multi-dimensional feature sets, to obtain a plurality of aggregated sample same-signal-identified radar echo signal multi-dimensional feature sets.

[0037] Step S322: Perform intra-set prototype extraction on the plurality of aggregated sample same-signal-identified radar echo signal multi-dimensional feature sets respectively, to obtain a plurality of same-signal-identified radar echo signal multi-dimensional feature prototypes.

[0038] Step S323: Aggregate the plurality of same-signal-identified radar echo signal multi-dimensional feature prototypes, to construct the multi-layer perception same-signal prototype library.

[0039] Specifically, first, a large amount of sample data from typical same-signal scenes such as background areas, such as smooth terrain, conventional static ground objects, etc. are collected. These data include a plurality of sample same-signal-identified radar echo signal sets, and a sample same-signal-identified radar echo signal multi-dimensional feature set corresponding to each signal set. The features include frequency domain, time domain, Doppler shift, and spatial direction dimension. Subsequently, according to a preset aggregation scale, which is determined according to the background scene type and the signal feature similarity threshold, for example, the aggregation categories are divided according to the terrain type and the signal frequency interval, the same-class aggregation is performed on the plurality of sample same-signal-identified radar echo signal sets, the signal sets with similar signal features and consistent corresponding background scenes are classified into the same category, and a plurality of same-class signal set groups are formed. On this basis, a mapping relationship between the signal sets and the feature sets is established according to the same-class aggregation result, the sample same-signal-identified radar echo signal multi-dimensional feature set corresponding to each same-class signal set group is mapped and aggregated, the multi-dimensional features in the same group are respectively integrated according to the feature dimensions, the mean, variance, etc. of the same dimension features are calculated, the common information of the same-class features is retained, and finally a plurality of aggregated sample same-signal-identified radar echo signal multi-dimensional feature sets are obtained, which provide regular and representative feature data basis for subsequent extraction of same-signal feature prototypes.

[0040] For each obtained aggregated sample same signal identification radar echo signal multi-dimensional feature set, prototype extraction operation is carried out in the set. In operation, first, all multi-dimensional feature data in a single aggregated feature set is traversed, statistical analysis is carried out according to four dimensions of frequency domain features, time domain features, Doppler shift features and spatial direction features, the mean value of feature data under each dimension is calculated, and the initial feature reference of the aggregated set in each dimension is determined. Then, according to the preset iteration scale, the initial feature mean value is taken as the starting point, and random iteration operation is carried out on the feature data of each dimension in the aggregated feature set, to generate multiple groups of iterated aggregated feature data. Subsequently, the iteration density of the iterated aggregated feature data is compared with the iteration density of the initial feature mean value. If the iteration density after iteration is greater than or equal to the initial density, the iteration process is repeated according to the preset scale, until the preset iteration number is reached, to ensure that the extracted prototype can cover the common information of the same type of features in the aggregated set to the greatest extent. After all iterations are completed, the most representative feature combination is selected from the final obtained iteration feature data, that is, the core data that can reflect the same signal feature of the background scene corresponding to the aggregated set, and the same signal feature prototype of the same signal identification radar echo signal multi-dimensional feature set corresponding to the aggregated sample is determined. By executing the above operation on all aggregated sample same signal identification radar echo signal multi-dimensional feature sets one by one, multiple same signal identification radar echo signal multi-dimensional feature prototypes are finally obtained, laying a foundation for subsequent construction of a multi-layer perception same prototype library.

[0041] The multiple same signal identification radar echo signal multi-dimensional feature prototypes extracted from each aggregated sample same signal identification radar echo signal multi-dimensional feature set are collected and summarized, and the feature dimension information corresponding to each prototype is sorted to ensure that each prototype completely contains four types of core feature data of frequency domain features, time domain features, Doppler shift features and spatial direction features, and the data format is consistent with the calling requirements of the subsequent similar identification link. Subsequently, all multi-dimensional feature prototypes after summarization are classified and regularized, and are secondarily grouped according to the type of the background scene corresponding to the prototype, such as a stable terrain background, a conventional static building background or a feature similarity, so that the internal structure of the prototype library is clearer and subsequent quick retrieval and matching are facilitated. At the same time, an index label is added to each multi-dimensional feature prototype, and the label content includes key information such as the background scene category corresponding to the prototype and the aggregated set number of the feature extraction source, to improve the positioning efficiency during subsequent calling. Finally, all same signal identification radar echo signal multi-dimensional feature prototypes after classification, regularization and addition of indexes are integrated and stored in a preset data storage format, such as a matrix form or a feature vector group form, to construct a multi-layer perception same prototype library that can be directly used for subsequent similar identification operation.

[0042] In one possible implementation manner, step S322 further includes: Step S3221: Traverse the plurality of aggregated sample co-signals identified radar echo signal multi-dimensional feature sets to perform mean value calculation on the plurality of aggregated sample co-signals identified radar echo signal multi-dimensional feature mean values.

[0043] Step S3222: According to the preset iteration scale, randomly iterate the plurality of aggregated sample co-signals identified radar echo signal multi-dimensional feature mean values in the plurality of aggregated sample co-signals identified radar echo signal multi-dimensional feature sets, and obtain a plurality of iteration aggregated sample co-signals identified radar echo signal multi-dimensional features.

[0044] Step S3223: When the iteration density of the plurality of iteration aggregated sample co-signals identified radar echo signal multi-dimensional features is greater than or equal to the iteration density of the plurality of aggregated sample co-signals identified radar echo signal multi-dimensional feature mean values, continue to randomly iterate the plurality of iteration aggregated sample co-signals identified radar echo signal multi-dimensional features according to the preset iteration scale until the preset iteration number is met, and obtain a plurality of co-signals identified radar echo signal multi-dimensional feature prototypes.

[0045] Specifically, the plurality of aggregated sample co-signals identified radar echo signal multi-dimensional feature sets obtained are traversed, and for the multi-dimensional feature data in each set, four dimensions of frequency domain, time domain, Doppler shift and spatial direction are covered, and mean value calculation is performed on the feature dimensions respectively, that is, the frequency domain feature data of all signals in a single aggregated set is averaged to obtain the frequency domain feature mean value of the set; Similarly, the time domain feature mean value, the Doppler shift feature mean value and the spatial direction feature mean value are calculated in turn, and the mean values of the four dimensions are integrated to form the aggregated sample co-signals identified radar echo signal multi-dimensional feature mean value corresponding to each aggregated set, and finally the multi-dimensional feature mean values of all aggregated sets are determined, providing an initial reference for subsequent iteration operations.

[0046] The distribution characteristics of the radar echo signal multi-dimensional features of the same signal identification are based on the aggregated sample. The standard deviation and data range of each dimension feature set the preset iteration scale. For example, the iteration adjustment amplitude of the frequency domain feature is limited to the range of mean ± 5%, the iteration step of the time domain feature is set to 0.01 seconds, the iteration interval of the Doppler shift feature is controlled to mean ± 0.5 Hz, and the iteration angle accuracy of the spatial direction feature is reserved to 0.1°. The iteration ensures that it neither deviates from the commonness range of the same signal features nor covers the reasonable variation interval. Then, the Monte Carlo random sampling algorithm is used. In each aggregated sample same signal identification radar echo signal multi-dimensional feature set, the calculated multi-dimensional feature mean is taken as the center, and 10-20 groups of adjustment parameters are randomly generated according to the preset iteration scale. Each group of parameters corresponds to the adjustment amount of the four dimensions of the frequency domain, time domain, Doppler shift, and spatial direction. Then, each group of adjustment parameters is applied to the initial multi-dimensional feature mean, and the corresponding new feature data is generated through feature dimension mapping operations, such as frequency domain feature superposition adjustment, time domain feature translation correction, Doppler shift feature compensation, and spatial direction feature angle fine tuning. Finally, the effectiveness of the generated new feature data is verified, the abnormal data beyond the aggregated set feature distribution range is removed, and the effective data conforming to the same signal feature rule is retained. Finally, multiple iteration aggregated sample same signal identification radar echo signal multi-dimensional features are obtained, which provide diversified candidate features for subsequent density comparison and prototype screening.

[0047] The iteration density of the multiple iteration aggregated sample same signal identification radar echo signal multi-dimensional features and the iteration density of the multiple aggregated sample same signal identification radar echo signal multi-dimensional feature mean are calculated by using the kernel density estimation algorithm. The iteration density reflects the distribution concentration of the feature data in the multi-dimensional space. The higher the density is, the more representative the group of features is of the same signal commonness of the corresponding aggregated set. Then, the two groups of density data are compared one by one. If the density of an iteration feature is greater than or equal to the density of the initial mean, it means that the iteration feature is more representative, and the random iteration operation needs to be repeated with the iteration feature as the new benchmark according to the preset iteration scale, such as frequency domain ± 5% and time domain 0.01 second step. If the iteration feature density is less than the initial mean density, the iteration direction needs to be adjusted, such as reducing the frequency domain adjustment amplitude and optimizing the time domain sampling range. During the whole process, the feature data and density value of each iteration are continuously recorded until the preset iteration number, such as 10-20 times, is completed. Finally, from the final iteration features of each aggregated set, a group of data with the highest density and conforming to the same signal feature rule of the set is selected and determined as the same signal identification radar echo signal multi-dimensional feature prototype. Multiple same signal identification radar echo signal multi-dimensional feature prototypes belonging to different aggregated sets are finally obtained, which provide core data support for subsequent construction of a multi-layer perception same signal prototype library.

[0048] In a possible implementation manner, step S400 further includes: Step S410: calling a signal processing double-channel middle low-pass filter channel to perform homomorphic low-pass filter processing on the multi-channel identification radar echo signal with the homomorphic signal identification in the multi-channel identification radar echo signal sequence array, and calling a signal processing double-channel middle I-W filter channel to perform heteromorphic difference reservation processing on the multi-channel identification radar echo signal with the heteromorphic signal identification in the multi-channel identification radar echo signal sequence array, to obtain the multi-channel identification filtered radar echo signal sequence array.

[0049] Specifically, a signal classification index table is constructed based on the identification (homomorphic / heteromorphic) of each signal in the multi-channel identification radar echo signal sequence array, the identification type of the signal at each channel and each time node is determined, and data positioning basis is provided for double-channel parallel processing. For the echo signal with the homomorphic signal identification, a Butterworth low-pass filter is used when the low-pass filter channel is called, the cutoff frequency is set to 500 Hz according to the characteristic that the homomorphic signal is mainly of low-frequency background characteristics, can be adjusted according to the actual scene, the low-frequency signal is allowed to pass through by convolution operation, and high-frequency noise is filtered out, and at the same time, zero-phase filter technology is used to avoid signal phase distortion, and the integrity of the background characteristics of the homomorphic signal is ensured. For the echo signal with the heteromorphic signal identification, when the I-W filter channel is enabled, the signal is first decomposed into wavelet coefficients of different scales through wavelet transform, the high-frequency wavelet coefficients reflecting the difference characteristics of the target are assigned a weight coefficient 1.2 to enhance the difference information, the low-frequency coefficients are assigned a weight coefficient 0.8 to suppress background interference, and then the signal is reconstructed through inverse wavelet transform, to realize reservation and enhancement of the difference characteristics of the heteromorphic signal. Finally, the homomorphic signal and the heteromorphic signal processed by the double channels are re-integrated according to the channel attribution and time sequence of the original sequence, to generate the multi-channel identification filtered radar echo signal sequence array, which not only ensures the purity of the homomorphic signal, but also highlights the target characteristics of the heteromorphic signal, and lays a high-quality signal foundation for subsequent imaging.

[0050] In a possible implementation manner, step S500 further includes: Step S510: performing distance compression and azimuth focusing processing on each multi-channel identification filtered radar echo signal sequence in the multi-channel identification filtered radar echo signal sequence array, to obtain a multi-channel identification filtered preprocessed radar echo signal sequence array.

[0051] Step S520: performing intra-sequence signal splicing and fusion on the multi-channel identification filtered preprocessed radar echo signal sequence array, and converting the fused signal into a two-dimensional space image through back projection, to obtain the multi-channel regional SAR image set.

[0052] Specifically, for the obtained multi-channel identification filtering radar echo signal sequence array, distance compression and azimuth focusing processing are respectively performed on each multi-channel identification filtering radar echo signal sequence. The distance compression link adopts a matched filtering algorithm to compress the expanded signal in the distance direction to the target actual distance size, focuses the distance direction energy through the matching operation with the transmitted signal, and greatly improves the distance direction resolution; the azimuth focusing processing is based on the Doppler shift principle, uses the Doppler information generated by the platform motion to perform azimuth phase correction on the signal, eliminates the azimuth signal expansion caused by the platform motion error, and makes the target energy gather in the azimuth direction, further optimizing the imaging clarity. After completing the two processes, each sequence is converted into a signal sequence with better focusing effect, and the overall multi-channel identification filtering pre-processing radar echo signal sequence array is formed.

[0053] According to the association relationship between the time stamp and the spatial coordinates of signal collection, the pre-processed echo signals collected at different time periods and different spatial positions in each sequence are matched segment by segment. First, the time deviation of each signal segment is calibrated through a time synchronization algorithm to ensure that the signals under the same scene are aligned in the time dimension. Then, based on the spatial position information (such as the latitude, longitude, height, and attitude angle data of the radar platform), the signals are spatially registered to eliminate the spatial misalignment of the signals caused by slight fluctuations in the platform motion trajectory. Subsequently, a weighted splicing strategy is adopted to assign weights to the signals in the overlapping area according to the signal-to-noise ratio. The signal segment with a higher signal-to-noise ratio has a higher weight proportion, realizing seamless splicing of the signal segments while retaining the low-frequency stable characteristics of the homologous signals and the high-frequency detailed characteristics of the heterologous signals, and forming a complete sequence fusion signal covering the corresponding area. Then, the fusion signal is converted into a two-dimensional spatial image through back projection, and the specific process is as follows: First, a two-dimensional spatial coordinate system is established, taking the geographic coordinates of the target area, such as the utm coordinate system, as the reference to divide uniform pixel grids and determine the spatial position coordinates of each pixel point. Then, the motion parameters of the radar platform during signal collection are retrieved, including the position, velocity, attitude angle at each sampling time, and radar system parameters such as wavelength, pulse repetition frequency, and antenna gain. The geometric relationships between the radar and each pixel point in the two-dimensional grid at each sampling time, such as slant range and azimuth angle, are calculated. Based on these geometric relationships, the one-dimensional echo signal after fusion is decomposed according to the "range-azimuth" dimension, and the echo signal energy at each sampling time is projected back to the corresponding pixel point. For each pixel point, the integral operation is used to superimpose all the echo signal contribution values that can irradiate the point. Meanwhile, the radar echo amplitude and phase information are combined to calculate the gray value of the pixel point. The amplitude information corresponds to the gray intensity, and the phase information is used to optimize the gray uniformity. During the projection process, distance migration correction is also needed to eliminate the signal shift in the range direction caused by the relative motion of the target and the radar, as well as azimuth ambiguity suppression to avoid signal interference between adjacent pixel points. After completing the back projection operation of all fusion signals, each multi-channel labeled and filtered radar echo signal sequence corresponds to a two-dimensional spatial image. After all these images are collected, a set of multi-channel regional SAR images is obtained.

[0054] In one possible implementation manner, step S600 further includes: Step S610: performing image feature extraction on the set of multi-channel regional SAR images respectively by using an image feature extractor to obtain a set of multi-channel regional SAR image features.

[0055] Step S620: performing two-by-two enumeration combination on the set of multi-channel regional SAR image features to obtain a set of multi-channel regional SAR image feature combinations.

[0056] Step S630: extracting a first multi-channel region SAR image feature combination from the multi-channel region SAR image feature combination set for image feature multi-channel cross enhancement fusion, obtaining a first multi-channel region SAR image cross enhancement fusion feature combination, and adding the first multi-channel region SAR image cross enhancement fusion feature combination into the multi-channel region SAR image cross enhancement fusion feature combination set.

[0057] Step S640: taking the multi-channel region SAR image feature set as an index, performing mean value processing on a plurality of multi-channel region SAR image cross enhancement fusion features corresponding to the same multi-channel region SAR image feature in the multi-channel region SAR image cross enhancement fusion feature combination set, and obtaining a multi-channel region SAR image cross enhancement fusion feature mean value set.

[0058] Step S650: performing inversion image enhancement on the multi-channel region SAR image set based on the multi-channel region SAR image cross enhancement fusion feature mean value set, and obtaining an enhanced multi-channel region SAR image set.

[0059] Step S660: performing image mean value filtering on the enhanced multi-channel region SAR image set, and obtaining the target region SAR image.

[0060] Specifically, a multi-channel regional SAR image set is taken as the processing object, and a convolution kernel-based image feature extractor is used to perform feature extraction. First, according to the characteristics of the multi-channel regional SAR image, such as the difference in ground object scattering, the texture distribution rule and the like, the appropriate convolution kernel parameters are configured, for example, a 3*3 or 5*5 size convolution kernel is selected, a gradient type convolution kernel is set for image edge features, such as a Sobel convolution kernel, a Gaussian type convolution kernel is set for texture features, and a mean type convolution kernel is set for scattering intensity features, to ensure that different convolution kernels can accurately capture feature information of different dimensions of the image. Subsequently, each image in the multi-channel regional SAR image set is sequentially input into the image feature extractor, and feature extraction is realized through sliding convolution operation of the convolution kernel and the image: the convolution kernel slides row by row and column by column on the image pixel matrix according to the preset step size, and each time it slides to a position, it performs product summation operation with the pixels in the corresponding region to generate a feature value reflecting the local features of the region. During the process, different types of convolution kernels act on the image, the gradient type convolution kernel highlights the gray level change of the feature boundary in the image, and the edge feature is strengthened; the Gaussian type convolution kernel smooths the image noise while preserving the detail distribution of the feature texture; the mean type convolution kernel integrates the pixel gray level of the local region, and reflects the scattering intensity mean value feature of the feature. After completing the feature extraction of all convolution kernels for each image, the extracted edge, texture, scattering intensity and other feature information are integrated into a group of feature vectors, which cover the core feature expression of the image in different dimensions. Finally, the feature vectors corresponding to all images in the multi-channel regional SAR image set are summarized to form a multi-channel regional SAR image feature set.

[0061] For all feature vectors in the multi-channel regional SAR image feature set, enumeration operation is performed according to the rule of two-by-two combination, that is, any two feature vectors in the set are selected to form a feature combination, which covers all possible combination forms between different channel image features, and finally a multi-channel regional SAR image feature combination set containing multiple feature combinations is generated, which provides rich feature interaction data for subsequent cross-enhanced fusion.

[0062] From the obtained multi-channel regional SAR image feature combination set, a first set of feature combinations, i.e. a first multi-channel regional SAR image feature combination, is extracted as a processing object. Then, according to the feature fusion specification, three-step core operations are carried out: first, image feature similarity analysis is carried out on the image features of two different channels in the feature combination, by calculating the cosine similarity between feature vectors, the Pearson correlation coefficient and other indicators, the similarity of the two features in the texture, edge, scattering intensity and other dimensions is quantified, and a set of image feature similarity coefficients containing multiple similarity data is generated; second, the obtained image feature similarity coefficient set is subjected to softmax processing, the similarity coefficients are converted into probability distribution form, the feature correlation weight corresponding to the high similarity coefficient is highlighted, and then a cross-enhanced fusion matrix is constructed according to the processed probability distribution, the matrix elements directly reflect the fusion priority and correlation strength between the dimensions of the two features; third, according to the cross-enhanced fusion matrix, the two features in the first multi-channel regional SAR image feature combination are subjected to image feature enhancement, the feature dimensions with high weight in the matrix, such as the edge feature dimension with high similarity, are strengthened, and the redundant dimensions with low weight are appropriately inhibited, and finally a first multi-channel regional SAR image cross-enhanced fusion feature combination with the advantages of the two features and more detailed details is generated. Finally, the fusion feature combination is added to the pre-created multi-channel regional SAR image cross-enhanced fusion feature combination set according to the data storage format requirements, providing high-quality fusion feature data support for subsequent mean value processing and target region image generation.

[0063] With the feature vectors in the multi-channel regional SAR image feature set as the index, all cross-enhanced fusion feature combinations associated with each index feature vector are selected from the multi-channel regional SAR image cross-enhanced fusion feature combination set, i.e. different fusion combinations formed by the same original image feature, and the mean values of these combinations are calculated according to the feature dimensions to eliminate random differences between different combinations, to obtain the cross-enhanced fusion feature mean value corresponding to each index feature vector, and all mean values are integrated to form a multi-channel regional SAR image cross-enhanced fusion feature mean value set.

[0064] A mapping correlation model between feature mean values and image pixels is established. By analyzing the cross-enhanced fusion feature mean value set of each dimension feature, such as texture, edge, and scattering intensity mean value, and the corresponding relationship between the original image pixel gray scale and spatial distribution, the influence weight of the feature mean value on the pixel optimization is determined. For example, the pixel gradient weight of the region corresponding to the high edge feature mean value is increased to strengthen the object contour clarity. Subsequently, for each original image in the multi-channel regional SAR image set, based on the above mapping model, the inversion operation is carried out: taking the pixel coordinates of the image as the index, the cross-enhanced fusion feature mean value at the corresponding position is called to dynamically adjust the gray value of the original pixel. For the region with high scattering intensity mean value, the pixel gray scale is appropriately improved to highlight the strong reflection target. For the region with rich texture feature mean value, the detail texture is retained through the pixel neighborhood enhancement algorithm. For the region with obvious edge feature mean value, the gradient enhancement technology is used to sharpen the object boundary. In the inversion process, image fidelity constraint needs to be introduced to avoid geometric deformation of the object or generation of false features due to feature enhancement. After the inversion and enhancement of all original images are completed, the processed images are arranged according to the channel attribution and the original sequence order to finally form an enhanced multi-channel regional SAR image set with clearer details, better contrast, and higher target recognition, which lays a good image foundation for generating target regional SAR images through mean filtering.

[0065] According to the detail richness and noise distribution of the enhanced image, the neighborhood window parameters of the mean filtering are determined. Usually, a 3*3 or 5*5 square window is selected. The selection of the window size is based on the principle that it can effectively filter local noise without damaging the key details such as weak targets, edges, and textures in the image. Subsequently, each image in the enhanced multi-channel regional SAR image set is input into the filtering processing module in turn, and the filtering operation is carried out in the form of sliding window calculation of mean value: taking each pixel of the image as the center, the gray values of all pixels within the neighborhood window are arithmetically averaged to obtain the filtered gray value of the center pixel, and the new gray value is used to replace the original pixel value. For image edge pixels, if the neighborhood window exceeds the image boundary, boundary extension strategies such as mirror extension and zero padding are used to supplement the missing pixel information within the window, ensuring the integrity and consistency of the filtering effect. In the filtering process, the window sliding step needs to be controlled, which is usually set to 1 to avoid loss of image details or block distortion due to too large step size. After the mean filtering of all enhanced images is completed, the processed images are integrated according to the original channel order and regional correlation to finally generate target regional SAR images with low noise, clear details, and stable overall image quality, meeting the application requirements of subsequent precise observation, identification, or monitoring of the target region.

[0066] In one possible implementation manner, step S630 further includes: Step S631: image feature similarity analysis is performed on the first multi-channel regional sar image feature combination to obtain a set of image feature similarity coefficients.

[0067] Step S632: softmax processing is performed on the set of image feature similarity coefficients, and a cross-enhancement fusion matrix is constructed according to the processing result.

[0068] Step S633: image feature enhancement is performed on the first multi-channel regional sar image feature combination based on the cross-enhancement fusion matrix to obtain a first multi-channel regional sar image cross-enhancement fusion feature combination.

[0069] Specifically, for the image features of two different channels in the first multi-channel regional sar image feature combination, such as the texture features of channel A and the edge features of channel B, a feature similarity calculation algorithm is used for analysis: by calculating the cosine similarity, Euclidean distance or Pearson correlation coefficient between the two sets of feature vectors, the similarity of the two sets of features in each feature dimension, such as gray distribution, texture frequency and edge gradient, is quantified. The similarity coefficient tends to 1, indicating that the two sets of features are highly similar in that dimension, and tends to 0, indicating significant difference. The similarity calculation results of all dimensions are arranged into an ordered data set to form a set of image feature similarity coefficients, which provides a quantitative basis for subsequent fusion weight allocation.

[0070] For the obtained set of image feature similarity coefficients, softmax normalization processing is performed: by means of the softmax function, each similarity coefficient in the set is converted into a probability value with a value range of [0, 1], and the sum of all probability values is 1, so as to eliminate the difference in numerical magnitude between different coefficients and highlight the priority of feature association corresponding to high similarity coefficients. For example, two dimensions with original similarity coefficients of 0.8 and 0.2, after softmax processing, the former probability value will be significantly higher than the latter, more clearly reflecting the difference in feature association strength. Subsequently, a two-dimensional cross-enhancement fusion matrix is created, which is initially empty. The row and column dimensions of the matrix are consistent with the dimensions of the two sets of features in the first multi-channel regional sar image feature combination. According to the correspondence of feature dimensions, the normalized probability values after softmax processing are filled in one by one: the element value of the i-th row and j-th column of the matrix is the fusion weight of the i-th dimension of the first set of features and the j-th dimension of the second set of features. After filling all the normalized values, a cross-enhancement fusion matrix is formed, which can accurately quantify the association strength of each dimension of the two sets of features. This matrix provides a clear weight basis for subsequent feature enhancement operations based on convolutional networks.

[0071] The constructed cross-enhanced fusion matrix is taken as the core parameter of the convolution network, such as the convolution kernel weight, and two groups of features in the first multi-channel regional SAR image feature combination, such as the texture features of channel A and the edge features of channel B, are respectively input into the feature enhancement layer of the convolution network. The convolution network performs dimension-by-dimension enhancement processing on the two groups of features through local connection and parameter sharing mechanism, and the weight values of the elements in the cross-enhanced fusion matrix are used as guidance. For the feature dimensions with high weights in the matrix, such as the texture-edge corresponding dimensions reflecting strong correlation, the network will strengthen the feature response of the dimensions, and highlight the feature details through convolution operation; for the redundant dimensions with low weights, the feature intensity is appropriately inhibited to avoid invalid information interference on the fusion effect. At the same time, the convolution network also performs nonlinear mapping on the enhanced features to further optimize the discrimination and expression ability of the features, so that the two groups of features can not only retain their respective advantages, but also realize efficient integration of complementary information. Finally, after the enhancement processing of the convolution network, the two groups of features form a group of feature data with multi-dimensional advantages, rich details and strong correlation, that is, the first multi-channel regional SAR image cross-enhanced fusion feature combination.

[0072] In the second embodiment, based on the same inventive concept as the micro SAR imaging method in the foregoing embodiments, as shown in FIG. 8, the present application provides a micro SAR imaging system, and the system and method embodiments in the present application are based on the same inventive concept. The system comprises: Figure 2 The array configuration module 10 is configured to configure a multi-channel receiving array for a target micro platform, and select a micro SAR radar sensor for integration.

[0073] The echo signal receiving module 20 is configured to scan a target region according to a preset scanning path by using the integrated micro SAR radar, and receive radar echo signals in parallel through the multi-channel receiving array to obtain a multi-channel radar echo signal sequence array.

[0074] The hetero-separation identification module 30 is configured to perform intra-sequence hetero-separation identification on each multi-channel radar echo signal sequence based on the signal multi-dimensional features of each multi-channel radar echo signal sequence in the multi-channel radar echo signal sequence array, and obtain a multi-channel identified radar echo signal sequence array, wherein each multi-channel identified radar echo signal corresponds to a homo-signal identification and a hetero-signal identification.

[0075] The echo signal sequence array acquisition module 40 is configured to perform homo-low-pass filtering processing and hetero-difference reservation processing on the multi-channel identified radar echo signal sequence array by calling a signal processing double-channel based on the homo-signal identification and the hetero-signal identification, and obtain a multi-channel identified filtered radar echo signal sequence array.

[0076] ​The sar image set acquisition module 50 is configured to traverse the multi-channel identification filtered radar echo signal sequence array by a synthetic aperture radar processing method, perform in-sequence information fusion, and obtain a multi-channel regional sar image set.

[0077] The image fusion module 60 is configured to perform multi-channel cross enhancement fusion of image features on the multi-channel regional sar image set, and obtain a target regional sar image.

[0078] Further, the system is further configured to implement the following functions: According to the preset signal multi-dimensional features, feature extraction is performed on each multi-channel radar echo signal of a first multi-channel radar echo signal sequence in the multi-channel radar echo signal sequence array, to obtain a first multi-channel radar echo signal multi-dimensional feature sequence, wherein the preset signal multi-dimensional features include frequency domain features, time domain features, Doppler shift features, and spatial direction features; a multi-layer perception homoplasmid library is obtained, similarity identification is performed between each multi-channel radar echo signal multi-dimensional feature in the first multi-channel radar echo signal multi-dimensional feature sequence and each multi-layer perception homoplasmid in the multi-layer perception homoplasmid library, when a similarity degree greater than or equal to a preset similarity identification result threshold exists in a similarity identification result, a corresponding multi-channel radar echo signal is identified as a homoplasmic signal, to obtain a first multi-channel homoplasmic signal identification radar echo signal sub-sequence; when a similarity degree greater than or equal to a preset similarity identification result threshold does not exist in a similarity identification result, a corresponding multi-channel radar echo signal is identified as an alloplasmic signal, to obtain a first multi-channel alloplasmic signal identification radar echo signal sub-sequence; the first multi-channel homoplasmic signal identification radar echo signal sub-sequence and the first multi-channel alloplasmic signal identification radar echo signal sub-sequence are fused in a time sequence from front to back, and the first multi-channel identification radar echo signal sequence obtained by the fusion is added to the multi-channel identification radar echo signal sequence array.

[0079] Further, the system is further configured to implement the following functions: A plurality of sample homoplasmic signal identification radar echo signal sets and a plurality of sample homoplasmic signal identification radar echo signal multi-dimensional feature sets corresponding to the plurality of sample homoplasmic signal identification radar echo signal sets are obtained, and the plurality of sample homoplasmic signal identification radar echo signal sets are aggregated according to a preset aggregation scale, the plurality of sample homoplasmic signal identification radar echo signal multi-dimensional feature sets are mapped and aggregated according to the aggregation result, to obtain a plurality of aggregated sample homoplasmic signal identification radar echo signal multi-dimensional feature sets; the plurality of aggregated sample homoplasmic signal identification radar echo signal multi-dimensional feature sets are respectively subjected to in-set prototype extraction, to obtain a plurality of homoplasmic signal identification radar echo signal multi-dimensional feature prototypes; the plurality of homoplasmic signal identification radar echo signal multi-dimensional feature prototypes are summarized, to construct the multi-layer perception homoplasmid library.

[0080] Further, the system is also used to realize the following functions: The multiple aggregated sample co-distribution signal identified radar echo signal multi-dimensional feature sets are traversed to perform mean value calculation on the multiple aggregated sample co-distribution signal identified radar echo signal multi-dimensional feature mean values, and the multiple aggregated sample co-distribution signal identified radar echo signal multi-dimensional feature mean values are randomly iterated in the multiple aggregated sample co-distribution signal identified radar echo signal multi-dimensional feature sets according to a preset iteration scale to obtain multiple iteration aggregated sample co-distribution signal identified radar echo signal multi-dimensional features.

[0081] Further, the system is also used to realize the following functions: The multiple channel identification radar echo signal sequence array is called to perform co-distribution low-pass filtering processing on the multiple channel identification radar echo signals with co-distribution signal identification in the multiple channel identification radar echo signal sequence array by calling a signal processing double-channel low-pass filtering channel, and the multiple channel identification radar echo signals with different distribution signal identification in the multiple channel identification radar echo signal sequence array are processed by calling a signal processing double-channel I-W filtering channel to perform different distribution difference reservation processing, thereby obtaining the multiple channel identification filtered radar echo signal sequence array.

[0082] Further, the system is also used to realize the following functions: The multiple channel identification filtered radar echo signal sequence array is called to perform distance compression and azimuth focusing processing on each multiple channel identification filtered radar echo signal sequence in the multiple channel identification filtered radar echo signal sequence array to obtain a multiple channel identification filtered preprocessed radar echo signal sequence array, the multiple channel identification filtered preprocessed radar echo signal sequence array is subjected to in-sequence signal splicing fusion, and the fused signal is converted into a two-dimensional space image through back projection to obtain the multiple channel regional SAR image set.

[0083] Further, the system is also used to realize the following functions: The image feature extractor is used for image feature extraction on the plurality of sets of multi-channel regional SAR images respectively, to obtain a plurality of sets of multi-channel regional SAR image features; the plurality of sets of multi-channel regional SAR image features are enumerated and combined two by two, to obtain a set of multi-channel regional SAR image feature combinations; a first multi-channel regional SAR image feature combination is extracted from the set of multi-channel regional SAR image feature combinations for image feature multi-channel cross-enhancement fusion, to obtain a first multi-channel regional SAR image cross-enhancement fusion feature combination, and the first multi-channel regional SAR image cross-enhancement fusion feature combination is added to a set of multi-channel regional SAR image cross-enhancement fusion feature combinations; the plurality of sets of multi-channel regional SAR image cross-enhancement fusion features corresponding to the same multi-channel regional SAR image feature in the set of multi-channel regional SAR image cross-enhancement fusion feature combinations are subjected to mean value processing with the plurality of sets of multi-channel regional SAR image features as indexes, to obtain a set of multi-channel regional SAR image cross-enhancement fusion feature mean values; the plurality of sets of multi-channel regional SAR images are subjected to inversion image enhancement based on the set of multi-channel regional SAR image cross-enhancement fusion feature mean values, to obtain a plurality of sets of enhanced multi-channel regional SAR images; the plurality of sets of enhanced multi-channel regional SAR images are subjected to image mean value filtering, to obtain the target regional SAR image.

[0084] Further, the system is further used to implement the following functions: The image feature similarity analysis is performed on the first multi-channel regional SAR image feature combination, to obtain a set of image feature similarity coefficients; the set of image feature similarity coefficients is subjected to softmax processing, and a cross-enhancement fusion matrix is constructed according to the processing result; the first multi-channel regional SAR image feature combination is subjected to image feature enhancement based on the cross-enhancement fusion matrix, to obtain a first multi-channel regional SAR image cross-enhancement fusion feature combination.

[0085] Embodiment three, based on the same inventive concept as the micro SAR imaging method in the foregoing embodiments, this embodiment provides a computer readable storage medium, which can be used for storing software programs, computer executable programs and modules, such as the program instructions / modules corresponding to the micro SAR imaging method in the embodiments of the present application. The processor executes the software programs, instructions and modules stored in the memory, thereby performing various functional applications and data processing of the computer device, i.e. implementing the above-mentioned micro SAR imaging method.

[0086] It should be noted that the above-mentioned embodiment sequences of the present application are merely for description only, but not for representing the advantages and disadvantages of the embodiments. And the above-mentioned embodiments of the present specification have been described. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0087] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0088] The specification and drawings are merely exemplary of the present application, and any and all modifications, variations, combinations or equivalents that are within the scope of the present application should be included. Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A micro-SAR imaging method, characterized in that: The method comprises: Configure a multi-channel receiving array on the target micro platform and select a micro SAR radar sensor for integration; The integrated micro-SAR radar scans the target area according to the preset scanning path, and receives the radar echo signals in parallel through the multi-channel receiving array to obtain a multi-channel radar echo signal sequence array; performing, based on the multidimensional signal characteristics of each multi-channel radar echo signal sequence in the multi-channel radar echo signal sequence array, an intra-sequence heterogeneous separation identification of each multi-channel radar echo signal sequence to obtain a multi-channel identification radar echo signal sequence array, wherein the identification corresponding to each multi-channel identification radar echo signal is a homogeneous signal identification and a heterogeneous signal identification; Based on the same-match signal identifier and the different-match signal identifier, calling the signal processing dual channel to perform the same-match low-pass filtering processing and the different-match difference retention processing on the multi-channel identification radar echo signal sequence array to obtain the multi-channel identification filtered radar echo signal sequence array; By using a synthetic aperture radar processing method, the multi-channel identification filter radar echo signal sequence array is traversed to perform intra-sequence information fusion to obtain a multi-channel regional SAR image set; The multi-channel regional SAR image set is subjected to image feature multi-channel cross-enhancement fusion to obtain the target region SAR image.

2. A micro-SAR imaging method as claimed in claim 1, characterized in that, According to the signal multidimensional characteristics of each multi-channel radar echo signal sequence in the multi-channel radar echo signal sequence array, an intra-sequence heterogeneous separation identification is performed on each multi-channel radar echo signal sequence to obtain a multi-channel identification radar echo signal sequence array, wherein the identification corresponding to each multi-channel identification radar echo signal is a homogeneous signal identification and a heterogeneous signal identification, including: performing feature extraction on each multi-channel radar echo signal of a first multi-channel radar echo signal sequence in the multi-channel radar echo signal sequence array according to preset signal multi-dimensional features to obtain a first multi-channel radar echo signal multi-dimensional feature sequence, wherein the preset signal multi-dimensional features include frequency domain features, time domain features, Doppler frequency shift features, and spatial direction features; Obtain a multi-layer perception same-matching prototype library, perform similarity identification on the multi-dimensional features of each multi-channel radar echo signal in the first multi-channel radar echo signal multi-dimensional feature sequence and each multi-layer perception same-matching prototype in the multi-layer perception same-matching prototype library, and when the similarity of the similarity identification results is greater than or equal to a preset similarity identification result threshold, perform same-matching signal identification on the corresponding multi-channel radar echo signal to obtain a first multi-channel same-matching signal identified radar echo signal subsequence; When the similarity of the similar recognition results does not have a similarity greater than or equal to the preset similar recognition result threshold, the corresponding multi-channel radar echo signal is marked as an alien signal to obtain a first multi-channel alien signal marked radar echo signal subsequence; The first multi-channel identical signal identification radar echo signal subsequence and the first multi-channel incompatible signal identification radar echo signal subsequence are fused in chronological order, and the fused first multi-channel identification radar echo signal sequence is added to the multi-channel identification radar echo signal sequence array.

3. A micro-SAR imaging method as claimed in claim 2, characterized in that, Get the multi-layer perception matching prototype library, including Acquire multiple sample co-matched signal identification radar echo signal sets and corresponding multiple sample co-matched signal identification radar echo signal multidimensional feature sets, perform homogeneous aggregation on the multiple sample co-matched signal identification radar echo signal sets according to a preset aggregation scale, and perform mapping aggregation on the multiple sample co-matched signal identification radar echo signal multidimensional feature sets based on the homogeneous aggregation results to obtain multiple aggregated sample co-matched signal identification radar echo signal multidimensional feature sets; Performing prototype extraction on each of the plurality of aggregated samples with the same signal identification radar echo signal multidimensional feature sets to obtain a plurality of multidimensional feature prototypes of the same signal identification radar echo signal; The multi-dimensional feature prototypes of the multiple co-matched signal identification radar echo signals are aggregated to construct the multi-layer perception co-matched prototype library.

4. A micro-SAR imaging method as claimed in claim 3, characterized in that, Performing prototype extraction on each of the plurality of aggregated samples and the multidimensional feature sets of the radar echo signals with the same signal identification to obtain a plurality of multidimensional feature prototypes of the radar echo signals with the same signal identification, including: Traversing the plurality of aggregated samples with the same signal identifier radar echo signal multidimensional feature set to perform mean calculation, and determining the mean of the plurality of aggregated samples with the same signal identifier radar echo signal multidimensional feature set; According to a preset iteration scale, randomly iterating the mean values ​​of the multidimensional features of the radar echo signals with the same signal identifications of the multiple aggregated samples in the multidimensional feature set of the radar echo signals with the same signal identifications of the multiple aggregated samples, to obtain a plurality of iterative multidimensional features of the radar echo signals with the same signal identifications of the multiple aggregated samples; When the iteration density of the multidimensional features of the radar echo signals with the same signal identification of the multiple iterative aggregated samples is greater than or equal to the iteration density of the mean of the multidimensional features of the radar echo signals with the same signal identification of the multiple aggregated samples, continue to randomly iterate the multidimensional features of the radar echo signals with the same signal identification of the multiple iterative aggregated samples according to the preset iteration scale until the preset number of iterations is met, and obtain multiple multidimensional feature prototypes of the radar echo signals with the same signal identification.

5. A micro-SAR imaging method as claimed in claim 1, characterized in that: The low-pass filter channel in the signal processing dual-channel is called to perform same-matching low-pass filtering processing on the multi-channel identification radar echo signals with same-matching signal identifiers in the multi-channel identification radar echo signal sequence array, and the IW filter channel in the signal processing dual-channel is called to perform different-matching difference retention processing on the multi-channel identification radar echo signals with different-matching signal identifiers in the multi-channel identification radar echo signal sequence array to obtain the multi-channel identification filtered radar echo signal sequence array.

6. A micro-SAR imaging method as claimed in claim 1, characterized in that: By using a synthetic aperture radar processing method, the multi-channel identification filter radar echo signal sequence array is traversed to perform intra-sequence information fusion to obtain a multi-channel regional SAR image set, including: performing range compression and azimuth focusing processing on each multi-channel identification filtering radar echo signal sequence in the multi-channel identification filtering radar echo signal sequence array to obtain a multi-channel identification filtering pre-processed radar echo signal sequence array; The multi-channel identification filter preprocessing radar echo signal sequence array is subjected to intra-sequence signal splicing and fusion, and the fused signal is converted into a two-dimensional spatial image by a back projection method to obtain the multi-channel regional SAR image set.

7. A micro-SAR imaging method as claimed in claim 1, characterized in that: Performing image feature multi-channel cross-enhancement fusion on the multi-channel regional SAR image set to obtain a target region SAR image, including: Using an image feature extractor to extract image features from the multi-channel regional SAR image set to obtain a multi-channel regional SAR image feature set; Enumerate and combine the multi-channel regional SAR image feature sets in pairs to obtain a multi-channel regional SAR image feature combination set; Extracting a first multi-channel regional SAR image feature combination from the multi-channel regional SAR image feature combination set to perform multi-channel cross-enhancement fusion of image features to obtain a first multi-channel regional SAR image cross-enhancement fusion feature combination, and adding the first multi-channel regional SAR image cross-enhancement fusion feature combination to the multi-channel regional SAR image cross-enhancement fusion feature combination set; Taking the multi-channel region SAR image feature set as an index, performing mean processing on multiple multi-channel region SAR image cross-enhancement fusion features corresponding to the same multi-channel region SAR image feature in the multi-channel region SAR image cross-enhancement fusion feature combination set to obtain a multi-channel region SAR image cross-enhancement fusion feature mean set; Performing inversion image enhancement on the multi-channel regional SAR image set based on the multi-channel regional SAR image cross-enhancement fusion feature mean set to obtain an enhanced multi-channel regional SAR image set; Perform image mean filtering on the enhanced multi-channel regional SAR image set to obtain the target region SAR image.

8. A micro-SAR imaging method as claimed in claim 7, characterized in that: Extracting a first multi-channel regional SAR image feature combination from the multi-channel regional SAR image feature combination set and performing multi-channel cross-enhancement fusion of image features to obtain a first multi-channel regional SAR image cross-enhancement fusion feature combination, including: performing image feature similarity analysis on the first multi-channel regional SAR image feature combination to obtain an image feature similarity coefficient set; Performing softmax processing on the image feature similarity coefficient set, and constructing a cross-enhancement fusion matrix based on the processing results; Based on the cross-enhancement fusion matrix, image feature enhancement is performed on the first multi-channel region SAR image feature combination to obtain the first multi-channel region SAR image cross-enhancement fusion feature combination.

9. A micro-SAR imaging system, characterized in that: The system is used to implement a micro-SAR imaging method according to any one of claims 1 to 8, and the system comprises: The array configuration module is used to configure a multi-channel receiving array for the target micro-platform and select the micro-SAR radar sensor for integration; The echo signal receiving module is used to scan the target area according to the preset scanning path through the integrated micro-SAR radar, and receive the radar echo signals in parallel through the multi-channel receiving array to obtain a multi-channel radar echo signal sequence array; a heterogeneous separation identification module, configured to perform intra-sequence heterogeneous separation identification of each multi-channel radar echo signal sequence in the multi-channel radar echo signal sequence array according to the multi-dimensional signal characteristics of each multi-channel radar echo signal sequence, thereby obtaining a multi-channel identification radar echo signal sequence array, wherein the identification corresponding to each multi-channel identification radar echo signal is a homogeneous signal identification and a heterogeneous signal identification; An echo signal sequence array acquisition module is used to call a signal processing dual channel to perform a same-match low-pass filtering process and a different-match difference retention process on the multi-channel identification radar echo signal sequence array based on the same-match signal identifier and the different-match signal identifier, so as to obtain a multi-channel identification filtered radar echo signal sequence array; A SAR image set acquisition module is used to traverse the multi-channel identification filter radar echo signal sequence array through a synthetic aperture radar processing method to perform intra-sequence information fusion to obtain a multi-channel regional SAR image set; The image fusion module is used to perform image feature multi-channel cross-enhancement fusion on the multi-channel regional SAR image set to obtain the target region SAR image.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, a micro-SAR imaging method according to any one of claims 1 to 8 is implemented.

Images