Meteorological feature image clustering and fuzzy fusion non-meteorological target identification method

Abstract

The invention discloses a non-meteorological target identification method based on meteorological feature image clustering and fuzzy fusion. Firstly, millimeter wave cloud radar meteorological data are preprocessed; and then generating a meteorological data parameter feature image, and converting the two-dimensional meteorological data array into a two-dimensional pseudo-color image. According to the method, distribution characteristics of meteorological targets and non-meteorological targets on a data image are analyzed, a meteorological data parameter characteristic image is divided into different potential target areas by using an image segmentation large model, and parameter characteristics in the potential target areas are extracted; and calculating the average radial texture and inter-library change degree of the reflectivity factor, the regional average value and variance of the radial speed, and the regional average value parameter characteristic statistical information of the speed spectrum width. A fuzzy logic fusion algorithm is adopted to determine a fuzzy membership function of statistical information, a fuzzy rule is set to calculate a fuzzy membership degree of each data point, and a weighted summation method is adopted to perform defuzzification to obtain a judgment value compared with a threshold value. According to the invention, the accuracy of identifying the non-meteorological target is improved.

Description

Technical Field

[0001] The present invention belongs to the field of computer vision and artificial intelligence, and relates to a method for identifying non-meteorological targets using a millimeter-wave cloud radar, and in particular to a method for identifying non-meteorological targets using meteorological feature image clustering and fuzzy fusion. Background Art

[0002] Meteorology plays an important role in disaster prevention and mitigation, agricultural production, transportation, energy supply and allocation, urban construction and planning, public services and life convenience, sustainable research and technological innovation.

[0003] With the steady development of economy and science and technology, meteorological radars are becoming more and more advanced, and meteorological observation stations have been built in various locations. In practical applications, radar observations not only include meteorological targets (such as clouds, precipitation, etc.), but also detect a large number of non-meteorological targets (such as bird flocks, insects, ground clutter, etc.), which will interfere with the accuracy of radar data. Intelligent meteorological target recognition methods have become a major trend, and there is a strong demand in the field of meteorological development. Classifying and labeling large-scale meteorological data is conducive to improving the accuracy of meteorological observations, supporting refined weather forecasts and disaster warnings, assisting atmospheric and environmental science research, expanding the application scenarios of millimeter-wave radars, and promoting the integrated development of intelligent recognition technology and radar technology. How to identify non-meteorological targets from massive and complex meteorological data has become an important research direction.

[0004] Millimeter-wave cloud radar is a radar system that uses electromagnetic waves in the millimeter wave band (30-300GHz) for detection. It is widely used in meteorological observations, especially for detecting clouds and precipitation. Compared with traditional S-band or C-band radars, millimeter-wave cloud radars have higher spatial resolution and sensitivity, and can detect meteorological elements such as tiny water droplets and ice crystals. Millimeter-wave cloud radar is not only suitable for different weather conditions such as clear sky, cloud layer and precipitation, but can also penetrate thinner cloud layer and light precipitation. It can provide effective data under clear sky and light precipitation conditions, especially in light rain or foggy weather. It can still provide stable observation data. Its beam is narrow and the detection height is low. It can simultaneously obtain multiple parameters such as reflectivity factor, speed, spectrum width, etc., and combined with polarization technology, it can further obtain information such as the shape, size and speed of the target. These characteristics make millimeter-wave cloud radar an important tool for refined meteorological observation and cloud physical property research, and it also shows potential application value in low-altitude aircraft monitoring, ecological research and other fields.

[0005] The research on intelligent recognition of non-meteorological targets by millimeter-wave cloud radar focuses on the challenge of accurately identifying and processing non-meteorological targets in complex environments. Millimeter-wave cloud radar generates a large amount of echo data. The echo signals of non-meteorological targets (such as insects, birds, buildings, etc.) may overlap with those of meteorological targets in terms of shape and intensity, making it difficult for the radar system to effectively distinguish them during detection, interfering with the detection of meteorological targets and reducing data quality and analysis accuracy. Therefore, real-time processing of these echo data and rapid identification and classification are the key to ensuring the accuracy of meteorological observations.

[0006] Machine learning and deep learning methods require a large amount of labeled data for training, but in actual operation, it is difficult to obtain and label large-scale, comprehensive non-meteorological target data, especially rare target types. Without sufficient data support, it is difficult for the model to learn stable features, resulting in poor recognition accuracy and generalization ability. Traditional algorithms are often inefficient when processing these large-scale data and cannot meet real-time requirements, resulting in an inability to respond in a timely manner, affecting real-time monitoring and analysis in practical applications. When faced with different geographical regions, climatic conditions and detection environments, the characteristics of radar echoes will change, which means that after the model is trained in a certain scenario, it may not be able to adapt to data in different scenarios. In order to improve recognition accuracy, it is often necessary to combine with other sensors (such as lidar, cameras, etc.), but due to differences in data types, resolutions and sampling rates between different sensors, data fusion faces technical challenges.

[0007] Using millimeter-wave cloud radar meteorological data to identify non-meteorological targets combines the distribution characteristics of meteorological targets and non-meteorological targets. Non-meteorological targets can be identified based on the distribution characteristics, but when the echo characteristics of non-meteorological targets and meteorological targets are similar, the segmentation results may be inaccurate, increasing the misidentification rate, especially in the distinction between insects and small cloud droplets. At the same time, noise data is easily misclassified as an independent non-meteorological target class, resulting in increased misclassification and interference, thus affecting the overall recognition accuracy.

[0008] Therefore, improving the accuracy of non-meteorological target recognition in meteorological data remains a major challenge. Summary of the invention

[0009] In view of the shortcomings of the prior art, the present invention proposes a non-meteorological target recognition method based on meteorological feature image clustering and fuzzy fusion. For the multiple parameter feature data of millimeter wave cloud radar meteorological data, the image segmentation model is used to realize the feature clustering of multiple parameter feature images, and then the fuzzy logic fusion algorithm is used to realize the fusion judgment of the multi-feature clustering results, and the high-precision recognition of non-meteorological targets is completed. This method can fully utilize the global feature aggregation capability of the image segmentation model and the local rule judgment of fuzzy logic, thereby optimizing the recognition accuracy of non-meteorological targets.

[0010] A non-meteorological target recognition method based on meteorological feature image clustering and fuzzy fusion, the specific steps are as follows:

[0011] Step 1: Preprocessing of millimeter wave cloud radar meteorological data

[0012] Collect the raw meteorological echo data of a certain site in millimeter-wave cloud radar THI mode throughout the day.

[0013] Since the scattered point clutter and radial thin line clutter in the original meteorological echo data are mostly Gaussian noise clutter caused by birds in the air and electromagnetic interference, they will have an adverse effect on the subsequent non-meteorological target identification. These clutters in the original meteorological echo data are removed to obtain filtered meteorological echo data. The meteorological data parameter characteristics of the filtered meteorological echo data are extracted, including reflectivity factor Z, velocity V, spectrum width W, signal-to-noise ratio SNR and other parameter characteristics, and normalized so that different parameter characteristics can be clustered under a unified dimension.

[0014] Step 2: Generation of millimeter wave cloud radar meteorological data parameter feature image

[0015] The pseudo-color processing method is used to correspond the normalized meteorological data parameter characteristics (reflectivity factor Z, velocity V, spectral width W, signal-to-noise ratio SNR, etc.) to different pseudo-colors according to the numerical values ​​to produce characteristic images of various meteorological data parameters; image segmentation and clustering methods are used to analyze the characteristics of meteorological targets and non-meteorological targets on the characteristic images of various meteorological data parameters.

[0016] Step 3: Clustering of potential targets in the parameter feature image of millimeter-wave cloud radar meteorological data

[0017] The Segment Anything (SAM) image segmentation model is fine-tuned in a one-shot manner, and unsupervised clustering is performed on the parameter feature images of each meteorological data. According to the continuity and consistency of the parameter features in the time dimension, each meteorological data parameter feature image is divided into different regions. Each region represents different echo features, such as background area, meteorological echo area and non-meteorological echo area. The meteorological echo area and non-meteorological echo area are regarded as potential target areas.

[0018] Step 4: Extract parameter feature statistics of potential target areas

[0019] According to the potential target areas divided in step 3, the statistical information of the meteorological data parameter characteristics in each area is extracted, including the average radial texture of the reflectivity factor, the radial variation of the echo intensity between libraries, the regional average of the radial velocity, the regional variance of the radial velocity and the regional average of the velocity spectrum width, and used as the input of the subsequent fuzzy logic fusion algorithm.

[0020] Step 5: Non-meteorological target recognition using fuzzy logic fusion algorithm

[0021] According to the parameter characteristic statistical information of each potential target area obtained in step 4, the fuzzy membership function of the parameter characteristic statistical information of each meteorological data is determined; according to the typical parameter characteristics of non-meteorological targets, fuzzy logic rules are set to calculate the non-meteorological target fuzzy membership of different parameter characteristic statistical information of each data point in the potential target area, and the weighted sum of these fuzzy memberships is used to obtain the non-meteorological target possibility score of each data point; according to the pre-set threshold, each data point in the potential target area is determined as a non-meteorological target or a meteorological target. The threshold value is adjusted to make the fuzzy boundaries of the potential target area and the points that are difficult to clearly classify in the area clearer and clearer, thereby improving the accuracy of non-meteorological target recognition.

[0022] The present invention has the following beneficial effects:

[0023] This method combines the global data parameter feature clustering capability of the large image segmentation model and the local rule judgment of the fuzzy logic fusion algorithm, making the identification of non-meteorological targets more accurate. By using multiple parameter features such as reflectivity factor, speed, spectral width and signal-to-noise ratio of millimeter-wave cloud radar meteorological data and then filtering noise, the impact of noise on the overall non-meteorological target recognition system can be reduced, and complex echo scenes under different meteorological conditions can be effectively dealt with, and the robustness of the model in a changing environment can be enhanced. The large image segmentation model can divide potential target areas according to the parameter characteristics of meteorological data. The fuzzy logic fusion algorithm allows the weights of different parameter feature statistical information to be flexibly adjusted, and fuzzy logic rules can also be defined according to the specific characteristics of non-meteorological echoes. This flexibility can adapt to different observation conditions or the characteristics of radar equipment, and can generate a comprehensive non-meteorological target possibility score for each data point in each potential target area, making subsequent processing and analysis more convincing. After the fusion of the large image segmentation model and the fuzzy logic algorithm, it can better handle the fuzzy area or transition area of ​​the cluster boundary. Especially in the transition area between meteorological echoes and non-meteorological echoes, the model can perform more detailed classification through fuzzy membership judgment, significantly improving the accuracy of identifying non-meteorological targets in the original millimeter-wave cloud radar meteorological data. By fusing multiple parameter feature statistical information, it can more comprehensively capture the differences between non-meteorological target and meteorological target echoes and reduce the misjudgment rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a flow chart of the non-meteorological target recognition method based on meteorological feature image clustering and fuzzy fusion;

[0025] Figure 2 A flowchart of the method for fine-tuning the Segment Anything model;

[0026] Figure 3 It is the algorithm flow chart of fuzzy logic algorithm;

[0027] Figure 4 The result map obtained by directly using image segmentation for the meteorological map measured at a certain site 1;

[0028] Figure 5 The result map obtained by directly using image segmentation for the meteorological map measured at a certain site 2;

[0029] Figure 6 The result map obtained by directly using image segmentation for the meteorological map measured at a certain site 3;

[0030] Figure 7 The result map obtained by directly using image segmentation for the meteorological map measured at a certain site 4;

[0031] Figure 8 This is the result of the clustering and fuzzy fusion algorithm of the meteorological characteristic image of a certain site 1;

[0032] Fig. 9 This is the result of clustering and fuzzy fusion algorithm of meteorological characteristic images of a certain site 2;

[0033] Fig.10 This is the result of clustering and fuzzy fusion algorithm of meteorological characteristic images of a certain site 3;

[0034] Fig.11 This is the result obtained by clustering and fuzzy fusion algorithm of 4 meteorological characteristic images at a certain station. DETAILED DESCRIPTION

[0035] The method of the present invention is further explained below with reference to the accompanying drawings. Figure 1 As shown in FIG, a non-meteorological target recognition method based on meteorological feature image clustering and fuzzy fusion is described. The specific steps are as follows:

[0036] Step 1: Preprocessing of millimeter wave cloud radar meteorological data

[0037] Import the original meteorological echo data of the millimeter-wave cloud radar in THI mode at a certain station, and save the four parameter features of the meteorological echo data, namely, reflectivity factor Z, radial velocity V, spectrum width W, and signal-to-noise ratio SNR, in the form of array matrix. THI mode is a time-altitude indication mode, which is usually used for vertical profile scanning. In this mode, the millimeter-wave cloud radar can vertically scan the atmosphere at a fixed azimuth to capture meteorological target information at different altitudes. In the two-dimensional matrix data of the parameter features of the original meteorological echo data, each array dimension is 1440×500, where the longitudinal dimension is the number of bins, with a maximum observation of 500 distance bins, and 1 distance bin is 30 meters, that is, the maximum observation height is 15 kilometers; the horizontal dimension is time, and the time resolution is 1 minute. The dimension 1440×500 is the meteorological echo data from 0 to 15 kilometers observed within 24 hours, ensuring that each parameter feature matrix data is in the same dimension.

[0038] Noise filtering: Near-ground, discrete and isolated noise signals often appear in meteorological radar data. The purpose of filtering the noise is to reduce the adverse effects of noise clutter in the subsequent image segmentation model division and fuzzy logic fusion algorithm analysis. The formula is:

[0039]

[0040] Where k represents the distance library of the reflectivity factor selected in the meteorological radar data; N represents the number of distance libraries of valid values ​​in the sliding window (N is 3×3); N t is the total number of all values ​​contained in the selected sliding window; P k It indicates the percentage of the effective echo values ​​in the selected sliding window to the total library number. k If the value is less than a certain threshold (initial setting is 75%), the numerical point can be determined to be isolated and discrete, identified as noise clutter and removed (set to NaN value). The reflectivity factor Z, velocity V, spectrum width W, signal-to-noise ratio SNR and other parameter features of the filtered meteorological echo data are normalized and mapped to between 0 and 1, so that different parameter features can be clustered under the same dimension.

[0041] Step 2: Generation of millimeter wave cloud radar meteorological data parameter feature image

[0042] The parameter characteristics (reflectivity factor Z, velocity V, spectral width W, signal-to-noise ratio SNR) of the filtered meteorological echo data are mapped to different pseudo colors according to the numerical values ​​to generate parameter characteristic images of each meteorological data, ensuring that the parameter characteristic image format of the millimeter-wave cloud radar meteorological data is suitable for the input of the large image segmentation model, facilitating the analysis of the distribution characteristics of meteorological targets and non-meteorological targets on the pseudo-color image, so that the model can better utilize the image characteristics to divide the area.

[0043] When performing pseudo-color mapping, a Cartesian coordinate system is used, with the horizontal axis set as the time axis, with each scale being 1 minute and containing 1440 points, corresponding to 1440 minutes, or 24 hours of observation data; the vertical axis is set as the altitude axis, with each scale being 30 meters and containing 500 points, corresponding to 15 kilometers.

[0044] In terms of color mapping selection, a color mapping scheme is selected for each parameter feature. Reflectivity factor Z: the commonly used "viridis" color band is easy to identify different echo intensities and clearly present the strong and weak contrast; velocity V: the "coolwarm" color band is used to distinguish between positive and negative velocity values; spectrum width W: the "plasma" color band is used to highlight the areas with different velocity spectrum widths; signal-to-noise ratio SNR: the "YlGnBu" color band is used to show the signal-to-noise ratio. According to the above color mapping schemes, a two-dimensional color map of each parameter feature is generated. These color mapping schemes can form a significant visual difference between different parameters, so as to more effectively parse and compare millimeter-wave cloud radar meteorological echo data.

[0045] Step 3: Clustering of potential targets in the parameter feature image of millimeter-wave cloud radar meteorological data

[0046] For each parameter feature image generated in step 2, analyze the distribution characteristics of non-meteorological targets on different parameter feature images one by one. Non-meteorological targets, such as clear sky echoes, have a low reflectivity factor, a wide distribution range, are strip-shaped, and are generally low in height (1-4 kilometers); the radial velocity is relatively smooth or close to 0 and evenly distributed, which is a low momentum feature; the spectral width is small; and the signal-to-noise ratio is also low. For example, ground clutter has a high reflectivity factor; ground clutter is generally static, with a velocity close to zero; the spectral width is small because its velocity changes little; and the signal-to-noise ratio is high.

[0047] The Segment Anything (SAM) model is used as an image segmentation model to process the millimeter wave cloud radar meteorological data parameter feature image. It is a segmentation model based on deep learning. Its main task is to classify each pixel in the parameter feature image into different regions, such as background region, meteorological echo region and non-meteorological echo region, so as to achieve the segmentation of the content of the parameter feature image. The SAM image segmentation model is fine-tuned in a one-shot manner to make it suitable for the regional clustering task of meteorological parameter feature images. The fine-tuning method flow is as follows: Figure 2 As shown, specifically:

[0048] First, a reference meteorological feature image I is given R (such as reflectivity factor map) and non-meteorological target area mask image Then, the image encoder ImEncoder in the trained SAM image segmentation model is used to extract the reference meteorological feature image I R The pixel value and position distribution of each pixel point in the meteorological feature image I to be measured are extracted as the semantic feature vector. The process is expressed as follows:

[0049] F I =ImEncoder(I)

[0050] F R =ImEncoder(I R )

[0051] in Respectively represent the reference meteorological characteristic image I R and the semantic feature vector of the meteorological feature image I to be tested, h represents the height of the image, w represents the width of the image, and c represents the number of channels of the image.

[0052] Then, according to the non-meteorological target area mask image M of the reference meteorological feature image R , calculate the non-meteorological target characteristics T p , the formula is:

[0053]

[0054] in, Represents the matrix dot product.

[0055] Next, calculate T P With F I The cosine similarity between them is used to obtain the confidence probability map S of the meteorological feature image I to be tested. p , using the statistical average confidence probability map S p Aggregate into an overall confidence map

[0056] Select the two pixels with the highest and lowest confidence values ​​in the overall confidence map S, denoted as P h and P l .P h Indicates the most likely center pixel of the non-meteorological target area, P l P is the pixel farthest from the non-meteorological target. h and P l They are regarded as positive and negative cue points and input into the cue encoder in the SAM image segmentation model to guide the SAM image segmentation model to complete the regional clustering of non-meteorological targets. In this way, the SAM model tends to complete the clustering of continuous regions around the positive cue points, and can segment out different potential target regions, such as background regions with no data, meteorological target regions, and non-meteorological targets.

[0057] Step 4: Extract parameter feature statistics of potential target areas

[0058] Through step 3, we can get each potential target area. Different clusters in the area may represent different meteorological targets. We can get the corresponding values ​​in the area according to the pixels in the regional image. One pixel corresponds to one value. We can extract the corresponding statistical information of these regional values, including the reflectivity factor and the average radial texture (T DBZ ), the radial variation of echo intensity between libraries (S PIN ), the regional average of radial velocity (M DVE ), the regional variance of radial velocity (S DVE ) and the regional average of the velocity spectrum width (M DSW ), the formula is as follows:

[0059]

[0060] Among them, N A 、N R Indicates the definition range in the azimuth and distance directions, N A ×N R is the size of the window; Z i,j is the reflection factor of any point, T DBZ Mainly reflects the local change of echo intensity; Z thresh The threshold of echo intensity change between libraries is usually set at 2-5dB. PIN It reflects the consistency of echo intensity variation along the radial direction; Indicates intermediate parameters; Indicates the radial velocity value of a point in the velocity graph within the sliding window range. represents the average value within the sliding window size; S DVE represents the regional variance of radial velocity; M DSW represents the average value of the velocity spectrum width within the sliding window; W i,j represents the spectral width at any point, V i,j Indicates the velocity magnitude at any point; for the characteristic quantity related to the echo intensity, set N A =3,N R =3; for the characteristic quantities related to radial velocity and velocity spectrum width, set N A =3,N R = 9. The above five statistical information are used as inputs for the subsequent fuzzy logic fusion algorithm.

[0061] Step 5: Non-meteorological target recognition using fuzzy logic fusion algorithm

[0062] After the image segmentation model divides the potential target area, there will be fuzzy boundaries and points in the area that are difficult to clearly classify. The fuzzy logic fusion algorithm can further subdivide the non-meteorological echo class and the meteorological echo class to help identify subtle differences and distinguish the echo type more accurately. The process is as follows: Figure 3 shown.

[0063] According to the five parameter feature statistical information extracted in step 4, a fuzzy membership function is defined to represent the different value ranges of the parameter feature statistical information. The membership function of the fuzzy logic algorithm is defined: the output value range of the membership function is 0 to 1, and the larger the value, the higher the possibility that the point is a non-meteorological target. According to the probability distribution of the five parameter feature statistical information, a trapezoidal broken line is used to represent the membership function of each parameter feature statistical information to achieve the fuzzification of the parameter feature statistical information, and the range of the trapezoidal membership function points and the trapezoidal inflection points of each parameter feature statistical information are roughly determined to ensure that the value range is between 0 and 1.

[0064] Fuzzy rules are designed to describe non-meteorological targets. The basis for judging non-meteorological targets is that clear sky echoes are at low altitudes, with small radial velocities and spectral widths, and small reflectivity factors; ground object echoes also mainly appear in the lower layers, with radial velocities and velocity spectral widths close to 0, no obvious movement, but large reflectivity factors. For meteorological targets such as precipitation echoes, they are judged from aspects such as echo shape, vertical structure, evolution, and movement.

[0065] Rule 1: If the average radial texture of the reflectivity factor is lower than the set minimum threshold, and the radial variation of the echo intensity is within the set range, and the regional average of the radial velocity is less than the set minimum threshold, and the regional variance of the radial velocity is less than the set minimum variance threshold, and the regional average of the velocity spectrum width is less than the set minimum spectrum width setting threshold, then it is judged as a clear sky echo among non-meteorological targets.

[0066] Rule 2: If the average radial texture of the reflectivity factor is high, the echo intensity varies little along the radial direction, the average value of the radial velocity region is close to 0, the velocity variance is low, and the spectral width value is low, it is judged to be a ground object echo among non-meteorological targets.

[0067] Substitute the parameter feature statistical information of each data point in the potential target area into the defined membership function to calculate the membership of each data point. Defuzzify the obtained membership, that is, convert the membership into a specific numerical output, and use the weighted summation method to calculate the defuzzification result. Different weight coefficients are set for each parameter feature statistical information according to the degree of influence, and the sum of all weight coefficients is 1 to obtain the final judgment value of each data point. Set a judgment threshold (set to 0.6 for clear sky echo and 0.4 for ground object echo). When the judgment value of the defuzzification result is higher than the judgment threshold, the data point is classified as a non-meteorological target echo; when it is lower than the judgment threshold, it is classified as a meteorological target echo. Adjust the threshold size and the weight coefficient size of the characteristic parameter statistical information to make the fuzzy boundaries of the potential target area and the points that are difficult to clearly classify in the area clearer and clearer, thereby improving the accuracy of non-meteorological target recognition.

[0068] Taking the meteorological data collected by stations 1, 2, 3, and 4 as an example, the results obtained by directly using the fine-tuning SAM image segmentation model are as follows: Figure 4 , Figure 5 , Figure 6 , Figure 7 As shown in the figure, the boundary area of ​​the non-meteorological target marked by the dotted rectangle is fuzzy, that is, the low-altitude non-meteorological target is not accurately identified and is mistakenly identified as a meteorological target. However, the meteorological data collected by a certain station 1, 2, 3, and 4 are used through the meteorological feature image clustering and fuzzy fusion non-meteorological target recognition method to obtain the following results Figure 8 , 9 As shown in , 10 and 11, the fuzzy boundary of the non-meteorological target area is optimized, and the non-meteorological targets can be accurately identified.

Claims

1. A non-meteorological target recognition method based on meteorological feature image clustering and fuzzy fusion, characterized in that: For the various parameter feature data of millimeter-wave cloud radar meteorological data, a large image segmentation model is used to realize the feature clustering of various parameter feature images, and then the fuzzy logic fusion algorithm is used to realize the fusion judgment of the multi-feature clustering results, and complete the high-precision recognition of non-meteorological targets.

2. The non-meteorological target recognition method based on meteorological characteristic image clustering and fuzzy fusion according to claim 1 is characterized in that: The method specifically comprises the following steps: Step 1: Preprocessing of millimeter wave cloud radar meteorological data; 1-1. Collect the original meteorological echo data of a certain site in millimeter-wave cloud radar THI mode throughout the day; 1-2. Filter the original meteorological echo data, extract the meteorological data parameter characteristics of the filtered meteorological echo data, including four parameter characteristics of reflectivity factor Z, radial velocity V, spectrum width W, and signal-to-noise ratio SNR, and save them in array matrix form after normalization processing; Step 2: Generation of millimeter wave cloud radar meteorological data parameter feature image; 2-1. Using pseudo color processing, normalized meteorological data parameter features are matched to different pseudo colors according to the numerical values, and characteristic images of each meteorological data parameter are produced; 2-2. Use image segmentation and clustering methods to analyze the characteristics of meteorological targets and non-meteorological targets on the characteristic images of various meteorological data parameters; Step 3: Clustering of potential targets in the parameter feature image of millimeter-wave cloud radar meteorological data 3-1. Fine-tune the SAM image segmentation model in a one-shot manner; 3-2. Perform unsupervised clustering on each meteorological data parameter feature image. According to the continuity and consistency of the parameter features in the time dimension, each meteorological data parameter feature image is divided into different regions. Each region represents different echo features, including background region, meteorological echo region and non-meteorological echo region. The meteorological echo region and non-meteorological echo region are regarded as potential target regions. Step 4: Extract parameter feature statistics of potential target areas According to the divided potential target areas, the statistical information of the meteorological data parameter characteristics in each target area is extracted, including the average radial texture of the reflectivity factor, the radial variation of the echo intensity between libraries, the regional average of the radial velocity, the regional variance of the radial velocity and the regional average of the velocity spectrum width, and used as the input of the subsequent fuzzy logic fusion algorithm; Step 5: Non-meteorological target recognition using fuzzy logic fusion algorithm 5-1. Determine the fuzzy membership function of the characteristic statistical information of each meteorological data parameter based on the obtained statistical information of each potential target area; 5-2. According to the typical parameter characteristics of non-meteorological targets, fuzzy logic rules are set to calculate the non-meteorological target fuzzy membership of different parameter characteristic statistical information of each data point in the potential target area, and these fuzzy memberships are weighted and summed to obtain the non-meteorological target possibility score of each data point; 5-3. According to the pre-set threshold, each data point in the potential target area is determined as a non-meteorological target or a meteorological target. The threshold value is adjusted to make the fuzzy boundaries of the potential target area and the points in the area that are difficult to clearly classify clearer and clearer, thereby improving the accuracy of non-meteorological target identification.

3. The non-meteorological target recognition method of meteorological characteristic image clustering and fuzzy fusion according to claim 2 is characterized in that: The specific process of step 3-1 fine-tuning is as follows: First, a reference meteorological feature image I is given R and non-meteorological target area mask images Then, the image encoder ImEncoder in the trained SAM image segmentation model is used to extract the reference meteorological feature image I R The pixel value and position distribution of each pixel point in the meteorological feature image I to be measured are extracted as the semantic feature vector. The process is expressed as follows: F I =ImEncoder(I) F R =ImEncoder(I R ) in Respectively represent the reference meteorological characteristic image I R and the semantic feature vector of the meteorological feature image I to be tested, h represents the height of the image, w represents the width of the image, and c represents the number of channels of the image; Then, according to the non-meteorological target area mask image M of the reference meteorological feature image R , calculate the non-meteorological target characteristics T p , the formula is: in, represents matrix dot product; Next, calculate T P With F I The cosine similarity between them is used to obtain the confidence probability map S of the meteorological feature image I to be tested. p , using the statistical average confidence probability map S p Aggregate into an overall confidence map 4. The non-meteorological target recognition method of meteorological characteristic image clustering and fuzzy fusion according to claim 3 is characterized in that: Step 3-2 is implemented as follows: Select the two pixels with the highest and lowest confidence values ​​in the overall confidence map S, denoted as P h and P l ;P h Indicates the most likely center pixel of the non-meteorological target area, P l P is the pixel farthest from the non-meteorological target. h and P l They are regarded as positive cue points and negative cue points and input into the cue encoder in the SAM image segmentation model to guide the SAM image segmentation model to complete the regional clustering of non-meteorological targets. In this way, the SAM model will complete the clustering of continuous regions around the positive cue points and segment out different potential target areas.

5. The non-meteorological target recognition method of meteorological characteristic image clustering and fuzzy fusion according to claim 4 is characterized in that: Step 4 is implemented as follows: Through step 3, we can get each potential target area. Different clusters in the area may represent different meteorological targets. We can get the corresponding values ​​in the area according to the pixels in the regional image. One pixel corresponds to one value. We can extract the corresponding statistical information of these regional values, including reflectivity factor, average radial texture T DBZ , the radial variation of echo intensity between libraries S PIN , the regional average of radial velocity M DVE , the regional variance of radial velocity S DVE and the regional average of the velocity spectrum width M DSW , the formula is as follows: Among them, N A 、N R Indicates the definition range in the azimuth and distance directions, N A ×N R is the size of the window; Z i,j is the reflection factor of any point, T DBZ Mainly reflects the local change of echo intensity; Z thresh is the threshold of echo intensity change between bins, S PIN It reflects the consistency of echo intensity variation along the radial direction; Indicates intermediate parameters; Indicates the radial velocity value of a point in the velocity graph within the sliding window range. represents the average value within the sliding window size; S DVE represents the regional variance of radial velocity; M DSW represents the average value of the velocity spectrum width within the sliding window; W i,j represents the spectral width at any point, V i,j Represents the velocity at any point.

Images