A method for achieving self-focusing in airborne synthetic aperture radar imaging

Abstract

This application provides a method for achieving autofocus in airborne synthetic aperture radar (SAR) imaging, belonging to the technical field of airborne SAR imaging. Specifically, it includes the following steps: calling azimuth autofocus parameters, running the autofocus processing module, outputting the range image domain and azimuth image domain, and storing them according to the azimuth direction; calling azimuth block overlap autofocus parameters, repeatedly calling the autofocus processing module, outputting the range image domain and azimuth image domain, and storing them according to the azimuth direction; running azimuth FFT, converting the azimuth frequency domain to the azimuth image domain, storing it according to the range direction; calling the range autofocus parameters, running the autofocus processing module, outputting the range image domain and azimuth image domain, and storing them according to the range direction. This processing scheme improves the autofocus effect of the image.

Description

Technical Field

[0001] This application relates to the field of airborne synthetic aperture radar imaging, and in particular to a method for achieving self-focusing in airborne synthetic aperture radar imaging. Background Technology

[0002] Currently, the mainstream chips used in radar signal processing software development are multi-core processors, one of the representatives being the TMS320C6678. This DSP (Digital Signal Processor) integrates eight C66x cores, offering excellent computational and storage performance. The overall performance of the TMS320C6678 is several times that of the previously widely used TS201. OpenMP is a multi-processor, multi-threaded parallel programming protocol for shared memory, offering good portability and applicable to multi-core parallel development of the TMS320C6678. When airborne synthetic aperture radar performs long-range, high-resolution, large-scene imaging under conditions of large oblique angles, it generally employs a PGA (Phase Gradient Aperture) algorithm based on maximum contrast for autofocus processing. SAR is an abbreviation for Synthetic Aperture Radar.

[0003] A master's thesis titled "Implementation of Real-Time SAR Imaging Algorithm Based on TMS320C6678" was found on CNKI (China National Knowledge Infrastructure). The thesis describes the implementation of the PGA algorithm based on the C6678, noting that the PGA algorithm has a large computational load, and the processing time increases significantly with the number of image points. Therefore, it fully utilizes the multi-core advantage of the C6678 to perform parallel processing. The thesis employs a multi-core navigator to achieve communication and data transmission between the cores, with core0 acting as the master control core and cores 1 through 7 as slave cores. Synchronization and communication between the cores, as well as the updating of shared data, are achieved through IPC interrupts and semaphore modules. However, its software architecture is quite complex.

[0004] In the traditional autofocusing processing based on the maximum contrast PGA algorithm, the computational load of PGA is relatively large, and the parallel processing design is complex, resulting in a low degree of parallelization and a long autofocusing operation time. Generally, autofocusing processing is only performed once in the azimuth direction. Summary of the Invention

[0005] In view of this, this application provides a method for achieving autofocus in airborne synthetic aperture radar imaging, which solves the problems in the prior art and improves the autofocus effect of the image.

[0006] The self-focusing method for airborne synthetic aperture radar imaging provided in this application adopts the following technical solution:

[0007] A method for achieving self-focusing in airborne synthetic aperture radar imaging includes the following steps:

[0008] Step 1: Input the distance image domain and the azimuth image domain, and store them according to the azimuth direction;

[0009] Step 2: Call the azimuth autofocus parameter setting, run the autofocus processing module, and output the distance image domain and azimuth image domain, storing them according to the azimuth direction;

[0010] Step 3: Call the azimuth direction block overlap autofocus parameter setting, call the autofocus processing module in a loop, and output the distance image domain and azimuth image domain, and store them according to the azimuth direction.

[0011] Optionally, it also includes: running an azimuth FFT, converting the azimuth frequency domain to the azimuth image domain, storing it by range, calling the range autofocus parameters, running the autofocus processing module, and outputting the range image domain and azimuth image domain, storing them by range.

[0012] Optionally, the self-focusing process is run in parallel using OpenMP multi-core processors.

[0013] Optionally, the self-focusing algorithm process is streamlined, and modules with computational loads exceeding preset values ​​and no strong correlation between data are identified. When designing and implementing these modules, the OpenMP parallel protocol directive #pragma ompparallel is used to automatically allocate them to the multi-core TMSC6678 for parallel processing based on the data block size.

[0014] Optionally, during autofocus processing, the data input address is first configured. The output address of the range autofocus is configured as the input address of the azimuth-direction whole-block autofocus, and the output address of the azimuth-direction whole-block autofocus is configured as the input address of the azimuth-direction segmented autofocus.

[0015] Optionally, the configuration method for azimuth block overlap autofocus during parameter configuration is as follows: the blocks are divided into a data processing area and a compensation output area in two dimensions. The output area is divided equally in two dimensions. The processing area is four times the size of the output area. The output area is located inside the processing area. When the output area is located at the edge of the image, the processing area expands with the point of the output area closest to the center of the image as the center. When the output area is located in a non-edge area, the processing area expands with the center of the output area as the center. The autofocus module is executed cyclically for each block data processing area. Only the block compensation output area is stored during output.

[0016] Optionally, the output area can be divided into (8-32)*(8-32).

[0017] In summary, this application includes the following beneficial technical effects:

[0018] This application employs range autofocus, azimuth autofocus, and block-overlapping autofocus in autofocus processing, significantly improving the focusing effect of the range and azimuth dimensions of SAR images under large-scene, high-sampling-rate conditions, resulting in higher resolution. Simultaneously, based on the OpenMP multi-core parallel processing technology of the TMS320C6678, modular development is adopted, simplifying the development and implementation process of the signal processing software. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a flowchart of the self-focusing algorithm in this application.

[0021] Figure 2 This is a diagram illustrating the directional self-focusing data segmentation method used in this application.

[0022] Figure 3 This application presents a flowchart of the self-focusing signal processing implementation.

[0023] Figure 4 This is a comparison chart of the self-focusing effect of a single self-focusing test and that of this application. Detailed Implementation

[0024] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0025] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this application, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0027] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. The drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0028] Furthermore, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.

[0029] This application provides a method for achieving self-focusing in airborne synthetic aperture radar imaging.

[0030] A method for achieving self-focusing in airborne synthetic aperture radar imaging includes the following steps.

[0031] Step 1: Input the distance image domain and the orientation image domain, and store them according to the orientation.

[0032] Step 2: Call the azimuth autofocus parameter setting, run the autofocus processing module, and output the distance image domain and azimuth image domain, storing them according to the azimuth direction.

[0033] Step 3: Call the azimuth-oriented block overlap autofocus parameter setting, and iteratively call the autofocus processing module to output the distance image domain and the azimuth image domain, storing them according to the azimuth direction. The configuration method for azimuth-oriented block overlap autofocus is as follows: the blocks are divided into a data processing area and a compensation output area in two dimensions. The output area is divided equally in two dimensions, using a (8-32)*(8-32) division, which can be 8*8, 32*32, or 16*16. The processing area is four times the size of the output area, and the output area is located inside the processing area. When the output area is located at the image edge, the processing area expands from the point closest to the image center of the output area; when the output area is located in a non-edge area, the processing area expands from the center of the output area. The autofocus module is executed iteratively on each block data processing area, and only the block compensation output area is stored during output.

[0034] Run the azimuth FFT, convert the azimuth frequency domain to the azimuth image domain, store it according to the range direction, call the range autofocus parameter setting, run the autofocus processing module, and output the range image domain and azimuth image domain, and store them according to the range direction.

[0035] During autofocus processing, the data input address is first configured. The output address of the range autofocus is configured as the input address of the azimuth-directed block autofocus, and the output address of the azimuth-directed block autofocus is configured as the input address of the azimuth-directed segmented autofocus.

[0036] The self-focusing processing utilizes OpenMP multi-core parallel operation. The self-focusing algorithm flow was streamlined, identifying modules with computational demands exceeding preset values ​​and no strong correlation between data segments. During the design and implementation of these modules, the OpenMP parallel protocol directive `#pragma omp parallel` was used to automatically allocate data to multiple cores of the TMS320C6678 for parallel processing based on the data block size. Using OpenMP significantly improves upon this, enabling flexible and convenient multi-core parallel development. For the 8-core parallel development of the TMS320C6678, OpenMP makes multi-core parallel development equivalent to designing and implementing on the master core (single core), eliminating the need to concern oneself with slave core communication and shared data updates, thus maximizing the computing power of a single TMS320C6678 chip.

[0037] The OpenMP multi-core parallel implementation of airborne synthetic aperture radar signal processing on the TMS320C6678 significantly reduces the complexity of program design logic, making the process clearer and easier to implement. Simultaneously, it fully utilizes the multi-core computing power of the TMS320C6678, effectively reducing software processing time and greatly improving the software's data processing capabilities. This application describes signal processing software running on the TMS320C6678. After azimuth interpolation is completed in the SAR operation, autofocus processing is designed and implemented. The autofocus processing employs the PGA algorithm based on maximum contrast. First, the autofocus function is generalized according to the algorithm, generating an autofocus processing module. For processes such as azimuth FFT and contrast calculation, which involve no iterative calculations and a large amount of data computation, the OpenMP multi-core parallel protocol is used. Then, parameter setting modules are configured to set parameters for range autofocus, azimuth autofocus, and azimuth block overlap autofocus. Finally, each autofocus parameter setting module and autofocus processing module is called sequentially to complete the focusing processing of the SAR image.

[0038] Compared to traditional autofocus that only processes the entire image in the azimuth direction, this application adds range autofocus processing and azimuth-oriented block overlapping autofocus processing, which greatly improves the image focusing quality under large scenes and high sampling rates. When using block autofocus, if no overlapping area is set between blocks, there will be obvious block effects at the block edges. In order to eliminate the block effects, an overlapping block area method is adopted in practical applications.

[0039] This application can be applied to autofocus processing in airborne synthetic aperture radar imaging. The processor used in this application is the TMS320C6678, but it can also be applied to other processors that support OpenMP or multi-core parallel development.

[0040] In one embodiment, a method for achieving self-focusing in airborne synthetic aperture radar imaging includes:

[0041] See appendix Figure 1This paper designs and implements signal processing software based on the self-focusing algorithm, using an OpenMP parallel architecture. First, the algorithm flow is analyzed, identifying modules with computational costs exceeding preset values ​​and no strong correlation between data points. Modules such as contrast calculation in the image domain, contrast sorting, self-focusing compensation in the frequency domain, and FFT to image domain conversion all meet this requirement. While the iterative process, due to data correlation, does not conform to OpenMP parallel design principles, its internal modules such as FFT with shifting and windowing, phase gradient calculation, and phase error compensation still comply with these principles. In designing and implementing these modules, the OpenMP parallel protocol directive `#pragma omp parallel` is used to automatically allocate data to the TMS320C6678's multi-core processors for parallel processing based on data block size. Throughout the design process, only flow control needs to be considered for the master core; slave cores only participate in computation. This results in a concise and clear design that fully utilizes the multi-core performance of the TMS320C6678.

[0042] Secondly, to adapt to range-direction whole-block autofocus, azimuth-direction whole-block autofocus, and azimuth-direction segmented overlapping autofocus, a configuration module before autofocus is added. The configuration mainly sets the data input address and output storage address. First, the data input address is configured so that the output address of range-direction autofocus is the same as the input address of azimuth-direction whole-block autofocus, and the output address of azimuth-direction whole-block autofocus is the same as the input address of azimuth-direction segmented autofocus. Because autofocus in large scenes with high sampling rates can reach 8K*8K points in the range and azimuth directions, requiring 512MB of space, two spaces are allocated in DDR and processed and stored using a ping-pong method.

[0043] For the configuration of azimuth-oriented block overlap autofocus, see Appendix. Figure 2 The blocks are divided equally in two dimensions.

[0044] The data processing area is divided into two dimensions: a data processing area and a compensation output area. The output area is also divided into two dimensions. The processing area is four times the size of the output area. The processing area includes overlapping areas. The output area is located inside the processing area and does not include overlapping areas. When the output area is located at the edge of the image, the processing area expands from the point closest to the center of the image. When the output area is located in a non-edge area, the processing area expands from the center of the output area. The self-focusing module is used to process each data processing area in a loop. Only the compensation output area of ​​the data processing area is stored at the output.

[0045] See appendix Figure 3 After the autofocus module algorithm is designed and implemented, the parallel design and implementation method for autofocus is started in SAR, with the number of output points designed as 8K*8K, including the following steps:

[0046] Step 1: Azimuth FFT and Transpose with Windowing: After azimuth interpolation, the data is stored in external memory according to the azimuth direction, with the range direction in the frequency domain and the azimuth direction in the image domain. First, OpenMP multi-core parallel processing is used to perform FFT processing on the azimuth direction and the data is stored in external memory according to the range direction. Then, OpenMP multi-core parallel processing is used to window the range direction and store it in external memory. At this point, the data is stored in external memory according to the range direction, with the range direction in the frequency domain and the azimuth direction in the image domain.

[0047] Step 2, Range-oriented whole PGA: Call the range-oriented autofocus parameter matching module and run the autofocus processing module to complete the range-oriented autofocus. After autofocus, the data is stored in external storage, stored according to the range direction. The range direction is the image domain and the azimuth direction is the image domain.

[0048] Step 3, Azimuth IFFT: Using OpenMP multi-core parallel processing, the azimuth data is first extracted, then azimuth IFFT processing is performed, and finally the results are stored in external storage. The final data is stored in external storage according to the azimuth direction; the range direction is in the image domain, and the azimuth direction is in the frequency domain.

[0049] Step 4, Azimuth-oriented whole-block autofocus processing: Call the azimuth-oriented autofocus parameter matching module and run the autofocus processing module to complete the azimuth-oriented whole-block autofocus. After autofocusing, the data is stored in external storage, stored according to the azimuth direction. The distance direction is the image domain and the azimuth direction is the image domain.

[0050] Step 5, Azimuth-direction block overlap autofocus processing: Call the azimuth-direction block overlap autofocus parameter configuration module, run the autofocus processing module in a loop, complete the azimuth-direction block overlap autofocus, and store the autofocused data to external storage, storing it according to the azimuth direction, with the distance direction as the image domain and the azimuth direction as the image domain, thus completing the entire autofocus processing.

[0051] This application employs range autofocus, azimuth autofocus, and block-overlapping autofocus in autofocus processing, significantly improving the focusing effect of the range and azimuth dimensions of SAR images under large-scene, high-sampling-rate conditions, resulting in higher resolution. Simultaneously, based on the OpenMP multi-core parallel processing technology of the TMS320C6678, modular development is adopted, simplifying the development and implementation process of the signal processing software.

[0052] Figure 4 These are angular reflection images of SAR imaging under different autofocusing processes. Figure 4 In the image, 'a' represents the effect of a single self-focusing operation. Figure 4 Figure b shows the self-focusing effect of this application. It can be seen that compared with single self-focusing, this application has a better focusing effect and stronger resolution.

[0053] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for achieving self-focusing in airborne synthetic aperture radar imaging, characterized in that, Includes the following steps: Step 1: Input the distance image domain and the azimuth image domain, and store them according to the azimuth direction; Step 2: Call the azimuth autofocus parameter setting, run the autofocus processing module, and output the range image domain and azimuth image domain, which are stored according to the azimuth direction. The autofocus processing adopts OpenMP multi-core parallel operation. The autofocus algorithm process is sorted out, and modules with a computational load greater than the preset value and no strong correlation between the data are identified. When designing and implementing the identified modules, the OpenMP parallel protocol directive #pragma omp parallel is used to automatically allocate the data to the TMSC6678 multi-core for parallel processing according to the data block size. Step 3: Call the azimuth-oriented block overlap autofocus parameter setting, and call the autofocus processing module in a loop. Output the distance image domain and the azimuth image domain, and store them according to the azimuth direction. The configuration method for the azimuth-oriented block overlap autofocus parameter setting is as follows: the block is divided into a data processing area and a compensation output area in two dimensions. The output area is divided into two equal parts in two dimensions. The processing area is four times the size of the output area. The output area is located inside the processing area. When the output area is located at the edge of the image, the processing area expands from the point of the output area closest to the center of the image. When the output area is located in a non-edge area, the processing area expands from the center of the output area. The autofocus module is executed on each block data processing area in a loop. Only the block compensation output area is stored when outputting.

2. The method for achieving self-focusing in airborne synthetic aperture radar imaging according to claim 1, characterized in that, Also includes: Run the azimuth FFT, convert the azimuth frequency domain to the azimuth image domain, store it according to the range direction, call the range autofocus parameter setting, run the autofocus processing module, and output the range image domain and azimuth image domain, and store them according to the range direction.

3. The method for achieving self-focusing in airborne synthetic aperture radar imaging according to claim 2, characterized in that, During autofocus processing, the data input address is first configured. The output address of the range autofocus is configured as the input address of the azimuth-directed block autofocus, and the output address of the azimuth-directed block autofocus is configured as the input address of the azimuth-directed segmented autofocus.

4. The method for achieving self-focusing in airborne synthetic aperture radar imaging according to claim 1, characterized in that, The output area follows (8-32). (8-32) Division.

Images