A SAR image target detection system and method based on artificial intelligence

Abstract

The invention discloses a SAR image target detection system and method based on artificial intelligence, belonging to the technical field of SAR image target detection. The system and method collect the distance of the vertical projection of a synthetic aperture radar device on the ground, attach a sampling period label, set sampling points in different directions with the projection point of the synthetic aperture radar device on the ground as the center, set different emission times, perform image fusion on the collected images, calculate the number of image noise points, calculate the average value of the noise points, preset a noise point number threshold, analyze and output the optimal distance range and the range of emission times. The invention processes the images collected for multiple times, identifies the number of image noise points by second-order derivative calculation, evaluates the number of image noise points in different sampling periods, and screens out high-quality relationship pairs, thereby effectively reducing the number of noise points in the SAR image, improving the image quality and the accuracy of target detection, and realizing efficient unmanned aerial vehicle SAR image target detection.

Description

Technical Field

[0001] The present invention relates to the technical field of SAR image target detection, and in particular to an artificial intelligence-based SAR image target detection system and method. Background Art

[0002] Synthetic aperture radar (SAR) technology has been widely used in recent years, especially in the fields of terrain mapping, military reconnaissance and disaster monitoring. SAR generates high-resolution two-dimensional or three-dimensional images by repeatedly emitting electromagnetic waves to the target area and receiving echoes. With the development of drone technology, the application of combining SAR devices with drones has gradually emerged. This combination not only improves the monitoring capability of complex environments, but also greatly expands the applicable scenarios of SAR technology. However, due to factors such as the flight altitude of drones, terrain changes and environmental interference, SAR images are often affected by noise and distortion. How to optimize image quality has become the focus of current research. In addition, the progress of artificial intelligence (AI) technology in image processing and data fusion has also provided a new path for improving the quality of SAR images. For example, AI can intelligently identify and eliminate noise in SAR images through deep learning algorithms, thereby improving the accuracy and efficiency of target detection.

[0003] However, the existing SAR image processing methods still have some shortcomings; first, traditional noise elimination technology mainly relies on image preprocessing or post-processing, lacks a dynamic adjustment mechanism, and cannot cope with the noise interference caused by environmental changes during the flight of the UAV in real time; second, most existing technologies do not fully consider the impact of the UAV altitude and the number of launches on the SAR image quality, which makes the image quality generated under different flight conditions unstable, and prone to excessive noise or blurred targets; in addition, although the existing SAR image fusion algorithm can improve the image quality to a certain extent, its computational efficiency is often low and cannot meet the needs of real-time target detection; therefore, there is an urgent need for a SAR image processing method that can combine UAV flight parameters, adjust the launch settings in real time, and optimize the image quality through intelligent algorithms. Summary of the invention

[0004] The purpose of the present invention is to provide a SAR image target detection system and method based on artificial intelligence to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] A SAR image target detection method based on artificial intelligence comprises the following steps: installing a synthetic aperture radar device on an unmanned aerial vehicle, collecting the distance of the vertical projection of the synthetic aperture radar device on the ground; setting a sampling period, when the sampling period ends, the unmanned aerial vehicle will move in the vertical distance, and construct a sampling period set; adding a sampling period label to the projection distance set, recording it as a sampling period distance, setting sampling points in different directions with the projection point of the synthetic aperture radar device on the ground as the center, setting different transmission times in each sampling period, and constructing a transmission time set; obtaining all SAR images generated by the synthetic aperture radar device in the sampling period, and performing image fusion; calculating the number of noise points of the fused image; generating a distance-transmission time correspondence pair based on the sampling period distance and the transmission time, and calculating the average value of the noise points; presetting a noise point number threshold, analyzing and outputting an optimal distance range and transmission time range.

[0007] As a preferred solution of the SAR image target detection method based on artificial intelligence described in the present invention, a synthetic aperture radar device is installed on the unmanned aerial vehicle, and one synthetic aperture radar device is installed on one unmanned aerial vehicle, and the synthetic aperture radar devices are uniformly numbered to construct a synthetic aperture radar device set. ,in, represents the synthetic aperture radar device installed on the a-th UAV, Represents the total number of drones.

[0008] Acquisition Synthetic Aperture Radar Device The distance projected vertically on the ground, constructing a projection distance set ,in, represents the synthetic aperture radar device installed on the ath UAV The distance projected vertically on the ground; set the sampling period. When the sampling period ends, the drone will move in the vertical distance to build a sampling period set ,in, represents the bth sampling period, and B represents the total number of sampling periods.

[0009] As a preferred solution of the SAR image target detection method based on artificial intelligence described in the present invention, a sampling period label is added to the projection distance set, which is recorded as the sampling period distance , the sampling period distance Indicates that during the sampling period Synthetic Aperture Radar Distance projected vertically onto the ground; measured using a synthetic aperture radar device With the projection point on the ground as the center, set sampling points in different directions to construct a sampling point set ,in, represents the sampling point in the hth direction, H represents the total number of directions, and the direction refers to the center position pointing to the sampling point position.

[0010] In each sampling period Set different launch times and build a launch time set ,in, Indicates the sampling period The number of launches within .

[0011] As a preferred solution of the SAR image target detection method based on artificial intelligence described in the present invention, in the sampling period Synthetic aperture radar device For each sampling point, The signal is received and the echo is generated to generate the SAR image and calculate the total number of SAR images. .

[0012] All generated SAR images are fused, and the fusion formula is as follows:

[0013] ;

[0014] in, represents the fused image, and n represents the number of launches The nth time in , and , Synthetic Aperture Radar In the sampling period Internal sampling point The image generated by the nth signal emission.

[0015] In the present invention, this formula represents the conversion of all generated images By fusion through weighted averaging, a comprehensive image is obtained. ,Multiple images of each sampling point in different directions are synthesized into a high-quality image through weighted averaging, which reduces noise and improves resolution.

[0016] Based on the fused image , calculate the number of noise points in the image, the calculation formula is as follows:

[0017] ;

[0018] ;

[0019] in, Indicates the sampling period The number of noise points in the image, (x, y) represents the pixel coordinates of the image, x is the horizontal pixel position of the image, y is the vertical pixel position; X and Y represent the width and height of the image, that is, the total number of pixels; and Represents the second-order derivative of an image in the x and y directions, and is usually used to detect edges and noise in an image; Indicates pixel weight, which represents the weight of each pixel in noise calculation and can be set according to the specific needs of the image; represents the sampling period distance function; Indicates the preset coefficient; Indicates the preset growth factor.

[0020] In the present invention, Represents the fused image The second-order rate of change of the brightness value at a specific coordinate point (x, y). The second-order derivative is used to describe the "acceleration" of the image brightness, that is, the rate of change of the brightness change. This rate of change can be used to detect edges or noise in the image; and Respectively represent the second-order position change of the image in the x and y directions, indicating the change of pixel coordinates; and The combination of is used to represent the rate of change of brightness of the image in the horizontal and vertical directions. If the change is too large, it means that the point is likely to be a noise point, because noise points usually appear as abnormal changes in local brightness; pixel weight Used to adjust the contribution of each pixel to the total amount of noise. If we care about the noise in certain specific areas (such as the edges or high-frequency areas of the image), we can assign higher weights to these areas; by introducing the function ,The impact of different altitudes of the UAV on the image quality is considered. Generally speaking, the higher the UAV flies, the lower the image resolution and the more noise points are present. Therefore, the altitude will have a significant impact on noise point detection. The brightness change of the image is combined with the altitude effect. If the brightness change is too large and the impact of the altitude function is large enough, it is judged as noise.

[0021] As a preferred solution of the SAR image target detection method based on artificial intelligence described in the present invention, based on the sampling period distance and the sampling period Number of launches within , generate the distance-number of launches corresponding relationship pair, recorded as .

[0022] Based on the relationship , calculate each relationship pair The corresponding number of noise points , and calculate the average noise value of all sampling periods. The calculation formula is as follows:

[0023] ;

[0024] in, Indicates the average value of noise over all sampling periods.

[0025] Preset noise threshold .

[0026] For the relationship , if the relationship is The corresponding number of noise points , then it is recorded as a high-quality relationship pair , and obtain all high-quality relationship pairs distance range and range of launch times.

[0027] In the present invention, by calculating the average noise point value of all sampling periods, the overall level of the influence of different altitudes and different transmission times on the resolution of SAR images can be obtained. If the average noise point value is lower than the noise point number threshold, it can be concluded that different altitudes and different transmission times in all sampling periods can obtain SAR images with higher resolution. Then, by comparing the number of noise points and the average noise value of SAR images collected by the altitude and transmission times in a single sampling period, high-quality relationship pairs can be screened out, so that the best altitude range and transmission number range can be obtained based on all high-quality relationship pairs.

[0028] A SAR image target detection system based on artificial intelligence comprises: a synthetic aperture radar configuration module, a sampling period management module, a SAR image generation and noise detection module and a distance-transmission number association module.

[0029] The synthetic aperture radar configuration module installs a synthetic aperture radar device on the UAV to collect the distance of the vertical projection of the synthetic aperture radar device on the ground; sets a sampling period, and when the sampling period ends, the UAV will move in the vertical distance to construct a sampling period set.

[0030] The sampling period management module adds a sampling period label to the projection distance set, records it as the sampling period distance, sets sampling points in different directions with the projection point of the synthetic aperture radar device on the ground as the center, sets different transmission times in each sampling period, and constructs a transmission time set.

[0031] The SAR image generation and processing module acquires all SAR images generated by the synthetic aperture radar device within a sampling period, performs image fusion, and calculates the number of noise points of the fused image.

[0032] The distance-transmission number association module generates a distance-transmission number correspondence pair based on the sampling period distance and the transmission number, calculates the noise point average value, presets the noise point number threshold, analyzes and outputs the optimal distance range and transmission number range.

[0033] Furthermore, the synthetic aperture radar configuration module also includes a radar numbering unit and a projection distance unit.

[0034] The radar numbering unit installs a synthetic aperture radar device on the UAV, and one synthetic aperture radar device is installed on one UAV, and the synthetic aperture radar devices are uniformly numbered to construct a synthetic aperture radar device set. ,in, represents the synthetic aperture radar device installed on the a-th UAV, Represents the total number of drones.

[0035] The projection distance unit collects the synthetic aperture radar device The distance projected vertically on the ground, constructing a projection distance set ,in, represents the synthetic aperture radar device installed on the ath UAV The distance projected vertically on the ground; set the sampling period. When the sampling period ends, the drone will move in the vertical distance to build a sampling period set ,in, represents the bth sampling period, and B represents the total number of sampling periods.

[0036] Furthermore, the sampling period management module also includes a sampling period marking and sampling point setting unit and a transmission times setting unit.

[0037] The sampling period labeling and sampling point setting unit adds a sampling period label to the projection distance set, which is recorded as the sampling period distance , the sampling period distance Indicates that during the sampling period Synthetic Aperture Radar Distance projected vertically onto the ground; measured using a synthetic aperture radar device With the projection point on the ground as the center, set sampling points in different directions to construct a sampling point set ,in, represents the sampling point in the hth direction, H represents the total number of directions, and the direction refers to the center position pointing to the sampling point position.

[0038] The transmission frequency setting unit, in each sampling period Set different launch times and build a launch time set ,in, Indicates the sampling period The number of launches within .

[0039] Furthermore, the SAR image generation and noise detection module also includes a SAR image fusion unit and a noise calculation unit.

[0040] The SAR image fusion unit, during the sampling period Synthetic aperture radar device For each sampling point, The signal is received and the echo is generated to generate the SAR image and calculate the total number of SAR images. ; Fuse all generated SAR images.

[0041] The noise calculation unit is based on the fused image , calculate the number of noise points in the image.

[0042] Furthermore, the distance-shot number association module also includes a relationship pair generation unit and a noise point number analysis unit.

[0043] The relationship pair generating unit is based on the sampling period distance and the sampling period Number of launches within , generate the distance-number of launches corresponding relationship pair, recorded as .

[0044] The noise point number analysis unit is based on the relationship , calculate each relationship pair The corresponding number of noise points , and calculate the average noise value of all sampling periods ; Preset noise threshold For the relationship , if the relationship is The corresponding number of noise points , then it is recorded as a high-quality relationship pair , and obtain all high-quality relationship pairs distance range and range of launch times.

[0045] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: by installing a synthetic aperture radar device on a UAV, collecting projection distances at different heights and sampling periods, and constructing a projection distance set, accurate scanning of the ground at different heights and time points is achieved. By setting the sampling period distance and multi-directional sampling points, and adjusting the number of transmissions, a SAR image with multiple transmissions is generated. On this basis, an image fusion algorithm is used to process the images collected multiple times, and the number of image noise points is identified by second-order derivative calculation. By generating a corresponding relationship between distance and number of transmissions, the number of noise points at different sampling periods is evaluated, and high-quality relationship pairs are screened out, thereby optimizing the combination of distance and number of transmissions. Finally, the present invention effectively reduces the number of noise points in SAR images, improves image quality and the accuracy of target detection, and realizes efficient UAV SAR image target detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0047] Figure 1 A schematic diagram of a flow chart of a SAR image target detection method based on artificial intelligence according to an embodiment of the present invention;

[0048] Figure 2 The present invention is a schematic structural diagram of a SAR image target detection system based on artificial intelligence. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] See also Figure 1-Figure 2 , the present invention provides a technical solution:

[0051] See also Figure 1 In the first embodiment, a SAR image target detection method based on artificial intelligence is provided, and the method comprises the following steps:

[0052] Step S1: Install a synthetic aperture radar device on the UAV to collect the distance of the vertical projection of the synthetic aperture radar device on the ground; set a sampling period, and when the sampling period ends, the UAV will move in the vertical distance to construct a sampling period set.

[0053] Install a synthetic aperture radar device on the UAV, and install one synthetic aperture radar device on each UAV, uniformly number the synthetic aperture radar devices, and construct a synthetic aperture radar device collection ,in, represents the synthetic aperture radar device installed on the a-th UAV, Represents the total number of drones.

[0054] Acquisition Synthetic Aperture Radar Device The distance projected vertically on the ground, constructing a projection distance set ,in, represents the synthetic aperture radar device installed on the ath UAV The distance projected vertically on the ground; set the sampling period. When the sampling period ends, the drone will move in the vertical distance to build a sampling period set ,in, represents the bth sampling period, and B represents the total number of sampling periods.

[0055] Step S2: Add a sampling period label to the projection distance set, record it as the sampling period distance, set sampling points in different directions with the projection point of the synthetic aperture radar device on the ground as the center, set different transmission times in each sampling period, and construct a transmission time set.

[0056] The sampling period label is added to the projection distance set, which is recorded as the sampling period distance , the sampling period distance Indicates that during the sampling period Synthetic Aperture Radar Distance projected vertically onto the ground; measured using a synthetic aperture radar device With the projection point on the ground as the center, set sampling points in different directions to construct a sampling point set ,in, represents the sampling point in the hth direction, H represents the total number of directions, and the direction refers to the center position pointing to the sampling point position.

[0057] In each sampling period Set different launch times and build a launch time set ,in, Indicates the sampling period The number of launches within .

[0058] Step S3: Acquire all SAR images generated by the synthetic aperture radar device within the sampling period, and perform image fusion; calculate the number of noise points of the fused image.

[0059] In the sampling period Synthetic aperture radar device For each sampling point, The signal is received and the echo is generated to generate the SAR image and calculate the total number of SAR images. .

[0060] All generated SAR images are fused, and the fusion formula is as follows:

[0061] ;

[0062] in, represents the fused image, and n represents the number of launches The nth time in , and , Synthetic Aperture Radar In the sampling period Internal sampling point The image generated by the nth signal emission.

[0063] Based on the fused image , calculate the number of noise points in the image, the calculation formula is as follows:

[0064] ;

[0065] ;

[0066] in, Indicates the sampling period The number of noise points in the image, (x, y) represents the pixel coordinates of the image, x is the horizontal pixel position of the image, y is the vertical pixel position; X and Y represent the width and height of the image, that is, the total number of pixels; and Represents the second-order derivative of an image in the x and y directions, and is usually used to detect edges and noise in an image; Indicates pixel weight, which represents the weight of each pixel in noise calculation and can be set according to the specific needs of the image; represents the sampling period distance function; Indicates the preset coefficient; Indicates the preset growth factor.

[0067] Step S4: Based on the sampling period distance and the number of transmissions, a distance-transmission number correspondence pair is generated, and the average noise point value is calculated; a noise point number threshold is preset, and the optimal distance range and transmission number range are analyzed and output.

[0068] Based on the sampling period distance and the sampling period Number of launches within , generate the distance-number of launches corresponding relationship pair, recorded as .

[0069] Based on the relationship , calculate each relationship pair The corresponding number of noise points , and calculate the average noise value of all sampling periods. The calculation formula is as follows:

[0070] ;

[0071] in, Indicates the average value of noise over all sampling periods.

[0072] Preset noise threshold .

[0073] For the relationship , if the relationship is The corresponding number of noise points , then it is recorded as a high-quality relationship pair , and obtain all high-quality relationship pairs distance range and range of launch times.

[0074] See also Figure 2 In the second embodiment of the present invention, a SAR image target detection system based on artificial intelligence is provided, which includes: a synthetic aperture radar configuration module, a sampling period management module, a SAR image generation and noise detection module and a distance-transmission number association module.

[0075] The synthetic aperture radar configuration module installs a synthetic aperture radar device on the UAV to collect the distance of the vertical projection of the synthetic aperture radar device on the ground; sets a sampling period, and when the sampling period ends, the UAV will move in the vertical distance to construct a sampling period set.

[0076] The sampling period management module adds a sampling period label to the projection distance set, records it as the sampling period distance, sets sampling points in different directions with the projection point of the synthetic aperture radar device on the ground as the center, sets different transmission times in each sampling period, and constructs a transmission time set.

[0077] The SAR image generation and processing module acquires all SAR images generated by the synthetic aperture radar device within a sampling period, performs image fusion, and calculates the number of noise points of the fused image.

[0078] The distance-transmission number association module generates a distance-transmission number correspondence pair based on the sampling period distance and the transmission number, calculates the average value of the noise point, presets the noise point number threshold, analyzes and outputs the optimal distance range and transmission number range.

[0079] Furthermore, the synthetic aperture radar configuration module also includes a radar numbering unit and a projection distance unit.

[0080] The radar numbering unit installs a synthetic aperture radar device on the UAV, and one synthetic aperture radar device is installed on one UAV, and the synthetic aperture radar devices are uniformly numbered to construct a synthetic aperture radar device set. ,in, represents the synthetic aperture radar device installed on the a-th UAV, Represents the total number of drones.

[0081] The projection distance unit collects the synthetic aperture radar device The distance projected vertically on the ground, constructing a projection distance set ,in, represents the synthetic aperture radar device installed on the ath UAV The distance projected vertically on the ground; set the sampling period. When the sampling period ends, the drone will move in the vertical distance to build a sampling period set ,in, represents the bth sampling period, and B represents the total number of sampling periods.

[0082] Furthermore, the sampling period management module also includes a sampling period marking and sampling point setting unit and a transmission times setting unit.

[0083] The sampling period labeling and sampling point setting unit adds a sampling period label to the projection distance set, which is recorded as the sampling period distance , the sampling period distance Indicates that during the sampling period Synthetic Aperture Radar Distance projected vertically onto the ground; measured using a synthetic aperture radar device With the projection point on the ground as the center, set sampling points in different directions to construct a sampling point set ,in, represents the sampling point in the hth direction, H represents the total number of directions, and the direction refers to the center position pointing to the sampling point position.

[0084] The transmission frequency setting unit, in each sampling period Set different launch times and build a launch time set ,in, Indicates the sampling period The number of launches within .

[0085] Furthermore, the SAR image generation and noise detection module also includes a SAR image fusion unit and a noise calculation unit.

[0086] The SAR image fusion unit, during the sampling period Synthetic aperture radar device For each sampling point, The signal is received and the echo is generated to generate the SAR image and calculate the total number of SAR images. ; Fuse all generated SAR images.

[0087] The noise calculation unit is based on the fused image , calculate the number of noise points in the image.

[0088] Furthermore, the distance-shot number association module also includes a relationship pair generation unit and a noise point number analysis unit.

[0089] The relationship pair generating unit is based on the sampling period distance and the sampling period Number of launches within , generate the distance-number of launches corresponding relationship pair, recorded as .

[0090] The noise point number analysis unit is based on the relationship , calculate each relationship pair The corresponding number of noise points , and calculate the average noise value of all sampling periods ; Preset noise threshold For the relationship , if the relationship is The corresponding number of noise points , then it is recorded as a high-quality relationship pair , and obtain all high-quality relationship pairs distance range and range of launch times.

[0091] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0092] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A SAR image target detection method based on artificial intelligence, characterized in that: The method comprises the following steps: Step S1: Install a synthetic aperture radar device on the UAV to collect the distance of the vertical projection of the synthetic aperture radar device on the ground; set a sampling period, and when the sampling period ends, the UAV will move in the vertical distance to construct a sampling period set; Step S2: adding a sampling period label to the projection distance set, recording it as the sampling period distance, setting sampling points in different directions with the projection point of the synthetic aperture radar device on the ground as the center, setting different transmission times in each sampling period, and constructing a transmission time set; Step S3: acquiring all SAR images generated by the synthetic aperture radar device within the sampling period, and performing image fusion; calculating the number of noise points of the fused image; Step S4: Based on the sampling period distance and the number of transmissions, a distance-transmission number correspondence pair is generated, and the average noise point value is calculated; a noise point number threshold is preset, and the optimal distance range and transmission number range are analyzed and output.

2. The SAR image target detection method based on artificial intelligence according to claim 1, characterized in that: The step S1 comprises the following steps: Install a synthetic aperture radar device on the UAV, and install one synthetic aperture radar device on each UAV, uniformly number the synthetic aperture radar devices, and construct a synthetic aperture radar device collection ,in, represents the synthetic aperture radar device installed on the a-th UAV, represents the total number of drones; Acquisition Synthetic Aperture Radar Device The distance projected vertically on the ground, constructing a projection distance set ,in, represents the synthetic aperture radar device installed on the ath UAV The distance projected vertically on the ground; set the sampling period. When the sampling period ends, the drone will move in the vertical distance to build a sampling period set ,in, represents the bth sampling period, and B represents the total number of sampling periods.

3. The SAR image target detection method based on artificial intelligence according to claim 2 is characterized in that: The step S2 comprises the following steps: The sampling period label is added to the projection distance set, which is recorded as the sampling period distance , the sampling period distance Indicates that during the sampling period Synthetic Aperture Radar Distance projected vertically onto the ground; measured using a synthetic aperture radar device With the projection point on the ground as the center, set sampling points in different directions to construct a sampling point set ,in, represents the sampling point in the hth direction, H represents the total number of directions, and the direction refers to the center position pointing to the sampling point position; In each sampling period Set different launch times and build a launch time set ,in, Indicates the sampling period The number of launches within .

4. The SAR image target detection method based on artificial intelligence according to claim 3 is characterized in that: The step S3 comprises the following steps: In the sampling period Synthetic aperture radar device For each sampling point, The signal is received and the echo is generated to generate the SAR image and calculate the total number of SAR images. ; All generated SAR images are fused, and the fusion formula is as follows: ； in, represents the fused image, and n represents the number of launches The nth time in , and , Synthetic Aperture Radar In the sampling period Internal sampling point The image generated by the nth signal emitted; Based on the fused image , calculate the number of noise points in the image, the calculation formula is as follows: ； ； in, Indicates the sampling period The number of noise points in the image, (x, y) represents the pixel coordinates of the image, x is the horizontal pixel position of the image, y is the vertical pixel position; X and Y represent the width and height of the image, that is, the total number of pixels; and Represents the second-order derivative of the image in the x and y directions; represents pixel weight; represents the sampling period distance function; Indicates the preset coefficient; Indicates the preset growth factor.

5. The SAR image target detection method based on artificial intelligence according to claim 4 is characterized in that: The step S4 comprises the following steps: Based on the sampling period distance and the sampling period Number of launches within , generate the distance-number of launches corresponding relationship pair, recorded as ; Based on the relationship , calculate each relationship pair The corresponding number of noise points , and calculate the average noise value of all sampling periods. The calculation formula is as follows: ； in, Represents the average value of noise in all sampling periods; Preset noise threshold ; For the relationship , if the relationship is The corresponding number of noise points , then it is recorded as a high-quality relationship pair , and obtain all high-quality relationship pairs distance range and range of launch times.

6. A SAR image target detection system based on artificial intelligence, characterized in that: The system includes: a synthetic aperture radar configuration module, a sampling period management module, a SAR image generation and noise detection module, and a distance-transmission number association module; The synthetic aperture radar configuration module installs a synthetic aperture radar device on the UAV to collect the distance of the vertical projection of the synthetic aperture radar device on the ground; sets a sampling period, and when the sampling period ends, the UAV will move in the vertical distance to construct a sampling period set; The sampling period management module adds a sampling period tag to the projection distance set, records it as the sampling period distance, sets sampling points in different directions with the projection point of the synthetic aperture radar device on the ground as the center, sets different transmission times in each sampling period, and constructs a transmission time set; The SAR image generation and processing module acquires all SAR images generated by the synthetic aperture radar device within a sampling period, and performs image fusion; calculates the number of noise points in the fused image; The distance-transmission number association module generates a distance-transmission number correspondence pair based on the sampling period distance and the transmission number, calculates the noise point average value, presets the noise point number threshold, analyzes and outputs the optimal distance range and transmission number range.

7. The SAR image target detection system based on artificial intelligence according to claim 6, characterized in that: The synthetic aperture radar configuration module also includes a radar numbering unit and a projection distance unit; The radar numbering unit installs a synthetic aperture radar device on the UAV, and one synthetic aperture radar device is installed on one UAV, and the synthetic aperture radar devices are uniformly numbered to construct a synthetic aperture radar device set. ,in, represents the synthetic aperture radar device installed on the a-th UAV, represents the total number of drones; The projection distance unit collects the synthetic aperture radar device The distance projected vertically on the ground, constructing a projection distance set ,in, represents the synthetic aperture radar device installed on the ath UAV The distance projected vertically on the ground; set the sampling period. When the sampling period ends, the drone will move in the vertical distance to build a sampling period set ,in, represents the bth sampling period, and B represents the total number of sampling periods.

8. The SAR image target detection system based on artificial intelligence according to claim 7, characterized in that: The sampling period management module also includes a sampling period marking and sampling point setting unit and a transmission number setting unit; The sampling period labeling and sampling point setting unit adds a sampling period label to the projection distance set, which is recorded as the sampling period distance , the sampling period distance Indicates that during the sampling period Synthetic Aperture Radar Distance projected vertically onto the ground; measured using a synthetic aperture radar device With the projection point on the ground as the center, set sampling points in different directions to construct a sampling point set ,in, represents the sampling point in the hth direction, H represents the total number of directions, and the direction refers to the center position pointing to the sampling point position; The transmission frequency setting unit, in each sampling period Set different launch times and build a launch time set ,in, Indicates the sampling period The number of launches within .

9. The SAR image target detection system based on artificial intelligence according to claim 8, characterized in that: The SAR image generation and noise detection module also includes a SAR image fusion unit and a noise calculation unit; The SAR image fusion unit, during the sampling period Synthetic aperture radar device For each sampling point, The signal is received and the echo is generated to generate the SAR image and calculate the total number of SAR images. ; Fuse all generated SAR images; The noise calculation unit is based on the fused image , calculate the number of noise points in the image.

10. The SAR image target detection system based on artificial intelligence according to claim 9, characterized in that: The distance-shot number association module also includes a relationship pair generation unit and a noise point number analysis unit; The relationship pair generating unit is based on the sampling period distance and the sampling period Number of launches within , generate the distance-number of launches corresponding relationship pair, recorded as ; The noise point number analysis unit is based on the relationship , calculate each relationship pair The corresponding number of noise points , and calculate the average noise value of all sampling periods ; Preset noise threshold For the relationship , if the relationship is The corresponding number of noise points , then it is recorded as a high-quality relationship pair , and obtain all high-quality relationship pairs distance range and range of launch times.

Images