CN113450308A - Radar rainfall detection method and device, computer equipment and storage medium - Google Patents

Radar rainfall detection method and device, computer equipment and storage medium Download PDF

Info

Publication number
CN113450308A
CN113450308A CN202110520287.0A CN202110520287A CN113450308A CN 113450308 A CN113450308 A CN 113450308A CN 202110520287 A CN202110520287 A CN 202110520287A CN 113450308 A CN113450308 A CN 113450308A
Authority
CN
China
Prior art keywords
radar
rain
radar image
image
clustering
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110520287.0A
Other languages
Chinese (zh)
Other versions
CN113450308B (en
Inventor
卫延波
周涛
张莹文
陈菲
许应豪
宋会丽
王培鑫
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Luoyang Normal University
Original Assignee
Luoyang Normal University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Luoyang Normal University filed Critical Luoyang Normal University
Priority to CN202110520287.0A priority Critical patent/CN113450308B/en
Publication of CN113450308A publication Critical patent/CN113450308A/en
Application granted granted Critical
Publication of CN113450308B publication Critical patent/CN113450308B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10032—Satellite or aerial image; Remote sensing
    • G06T2207/10044—Radar image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Physics & Mathematics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Probability & Statistics with Applications (AREA)
  • Quality & Reliability (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention relates to the technical field of remote sensing, in particular to a radar rainfall detection method, a radar rainfall detection device, computer equipment and a storage medium. A radar rainfall detection method comprises the following steps: acquiring a radar image and preprocessing the radar image; extracting features of the radar image; calculating the distance between the features of the radar images and a plurality of preset clustering centers, wherein the clustering centers comprise the clustering center of the rain radar images and the clustering center of the rain-free radar images; judging whether the clustering center with the minimum distance between the features of the radar images and the clustering center is the clustering center of the radar image with rain; if so, the obtained radar image is a radar image with rain, otherwise, the radar image without rain is obtained. The embodiment of the invention provides a rainfall detection method based on the combination of correlation coefficient vector characteristics and unsupervised K-means clustering, and the problem of low rainfall detection precision caused by a fixed threshold is solved.

Description

Radar rainfall detection method and device, computer equipment and storage medium
Technical Field
The invention relates to the technical field of remote sensing, in particular to a radar rainfall detection method, a radar rainfall detection device, computer equipment and a storage medium.
Background
The X-band navigation radar image is generally used for ship navigation and inversion of sea wave parameters, and due to the influence of the marine environment, the navigation radar image generally contains some non-sea wave information, such as rainfall interference and the like.
The X-band marine radar is very sensitive to rain interference, and raindrops can change the roughness of the sea surface and affect the propagation of electromagnetic waves when the navigation radar is used to telemeter the sea surface. Meanwhile, the echo intensity of the radar image increases with the roughness of the sea surface. In general, an inversion method for extracting sea wave parameters and sea surface wind field information is based on radar images without rainfall, and the existence of rainfall interference changes texture characteristics of the radar images, so that the quality of the radar images is influenced, and the reliability of the extracted information is reduced.
In order to control the image quality of the X-band navigation radar and improve the inversion accuracy of the sea wave parameters, it is necessary to detect whether the radar image contains rainfall interference before inverting the sea wave parameters.
Disclosure of Invention
In view of the above, it is necessary to provide a radar rainfall detection method, apparatus, computer device and storage medium.
The embodiment of the invention is realized in such a way that the radar rainfall detection method comprises the following steps:
acquiring a radar image and preprocessing the radar image;
extracting features of the radar image;
calculating the distance between the features of the radar images and a plurality of preset clustering centers, wherein the clustering centers comprise a clustering center of a rain radar image and a clustering center of a rain-free radar image;
judging whether the cluster center with the minimum distance between the features of the radar image and the cluster center is the cluster center of the radar image with rain; if so, the obtained radar image is a radar image with rain, otherwise, the radar image without rain is obtained.
In one embodiment, the present invention provides a radar rainfall detection device comprising:
the acquisition module is used for acquiring a radar image and preprocessing the radar image;
the characteristic extraction module is used for extracting the characteristics of the radar image;
the computing module is used for computing the distance between the features of the radar images and a plurality of preset clustering centers, wherein the clustering centers comprise the clustering center of the rain radar images and the clustering center of the rain-free radar images;
the judging module is used for judging whether the clustering center with the minimum distance between the features of the radar images and the clustering center is the clustering center with the rain radar images; if so, the obtained radar image is a radar image with rain, otherwise, the radar image without rain is obtained.
In one embodiment, the invention provides a computer device comprising a memory and a processor, the memory having stored therein a computer program which, when executed by the processor, causes the processor to perform the steps of the above-described radar rain detection method.
In one embodiment, the present invention provides a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and when executed by a processor, causes the processor to execute the steps of the radar rainfall detection method.
The radar rainfall detection method provided by the embodiment of the invention is based on the combination of the relevance coefficient vector characteristics and unsupervised K-means clustering, the problem of low rainfall detection precision caused by a fixed threshold value is avoided, the rainfall detection task can be completed according to the relevance coefficient vector characteristics and the extracted clustering class center, the precision requirement of the engineering on radar image rainfall detection can be met, and the radar rainfall detection method can be applied to the engineering in a programmed manner. The advantages of the invention at least include: a threshold value of rainfall detection does not need to be set, the task of rainfall detection can be automatically completed under different rainfall conditions, and the detection precision is high; the difference of the correlation coefficients of the rain-free radar image and the rain radar image in the azimuth direction is fully utilized to extract the vector characteristics of the correlation coefficients, the problem that the detection position is accurately selected in a lagging azimuth angle is solved, and the precision of rainfall detection is improved; after the cluster centroids are extracted offline, the vector features extracted from the radar images are compared with the distance between the centroids to complete the rainfall detection task, the method is low in complexity and small in computation amount, and rainfall detection can be performed online in real time; by setting the clustering class center number of the characteristics, the radar images without rain, light rain, medium rain and heavy rain can be distinguished, so that the detection precision and the reliability of the method are improved; the invention utilizes unsupervised K-means clustering machine learning algorithm, the precision of the detection method is related to the radar sample data, and the reliability and precision of the method can be further improved in the later period on the basis of large data accumulation.
Drawings
FIG. 1 is a diagram of an application environment of a radar rainfall detection method provided in one embodiment;
FIG. 2 is a flow diagram of a radar rain detection method in one embodiment;
FIG. 3 is a logic diagram of a radar rain detection method in one embodiment;
FIG. 4 is a radar image and selected analysis area before preprocessing according to the present invention;
FIG. 5 is an autocorrelation coefficient of a radar image at a lag azimuth;
FIG. 6 is a diagram illustrating cluster class centroid extracted by using correlation coefficient vector features according to the present invention;
FIG. 7 is a diagram illustrating the distance between the radar image features and the cluster centroid in accordance with the present invention;
FIG. 8 is a graph of rainfall intensity recorded simultaneously during the experiment;
FIG. 9 is a diagram illustrating a detection result of the rainfall detection method proposed in the present invention;
FIG. 10 is a graph of the percent zero intensity based on radar images;
FIG. 11 is a graph of the results of the zero intensity percentage method;
FIG. 12 is a block diagram showing the structure of a radar rainfall detection device in one embodiment;
FIG. 13 is a block diagram showing an internal configuration of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
It will be understood that, as used herein, the terms "first," "second," and the like may be used herein to describe various elements, but these elements are not limited by these terms unless otherwise specified. These terms are only used to distinguish one element from another. For example, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script, without departing from the scope of the present disclosure.
Fig. 1 is a diagram of an application environment of a radar rainfall detection method provided in an embodiment, as shown in fig. 1, in the application environment, including a radar 100 and a computer device 200.
In the embodiment of the present invention, the radar 100 is mainly an X-band navigation radar, and the present invention is proposed based on ship navigation and inversion of sea wave problem, but the application thereof is not limited to this specific application environment. The type of radar 100 is not necessarily limited to the X-band radar, only because the X-band radar is more significantly affected by rainfall, and the effect of applying the method is more significant, and the method provided by the invention can be applied to other types of radars as well.
In the embodiment of the present invention, the computer device 200 may be an independent physical server or terminal, may also be a server cluster formed by a plurality of physical servers, and may be a cloud server providing basic cloud computing services such as a cloud server, a cloud database, a cloud storage, and a CDN. The computer 200 may acquire the detection image of the radar 100 on line, or may perform the detection on line in an off-line manner, that is, the image acquired by the radar 100 is not transmitted to the computer device 200 directly or indirectly in a real-time transmission manner for processing, and both the two manners may implement the method of the present invention, which is not limited in this embodiment of the present invention.
As shown in fig. 2 and 3, in an embodiment, a radar rainfall detection method is provided, and this embodiment is mainly illustrated by applying the method to the computer device 200 in fig. 1. The radar rainfall detection comprises the following steps:
and step S202, acquiring a radar image and preprocessing the radar image.
In the embodiment of the invention, the radar image can be acquired on line, or can be acquired from the radar and then transferred to computer equipment for processing through a storage medium and the like. The image is preprocessed, mainly image noise is filtered, and the influence of same frequency interference on rainfall detection is restrained.
And step S204, extracting the features of the radar image.
In the embodiment of the invention, sea clutters in the navigation radar image are distributed in stripes with alternate light and shade, and have strong space-time correlation characteristics, but when high-frequency rainfall interference is introduced into the collected radar image, the radar echo intensity is increased along with the increase of the rainfall intensity, so that the correlation characteristics of the echoes in the radar image are influenced. The radar image identification method and the radar image identification device can accurately identify whether the radar image is rainy or rainless by extracting the features of the radar image and identifying the radar image by using the features.
And step S206, calculating the distance between the characteristics of the radar images and a plurality of preset cluster centers, wherein the cluster centers comprise the cluster center of the rain radar image and the cluster center of the rain-free radar image.
In the embodiment of the invention, the clustering centroid is determined by a training data set, and the clustering centroid comprises the clustering centroid with the rain radar image and the clustering centroid without the rain radar image; further, by setting the number of cluster centroids, heavy rain, medium rain, light rain, and the like can be distinguished.
Step S208, judging whether the clustering center with the minimum distance between the features of the radar images and the clustering center is the clustering center of the radar images with rain; if so, the obtained radar image is a radar image with rain, otherwise, the radar image without rain is obtained.
In the embodiment of the invention, if the cluster centroid closest to the characteristic distance of the radar image is the cluster centroid of the rain image, the radar image can be judged to be the rain image, otherwise, the radar image is the rain-free image; if the image is a rain image, the image can be divided into heavy rain, medium rain, light rain and the like, so that the identification is more accurate.
The radar rainfall detection method provided by the embodiment of the invention is based on the combination of the relevance coefficient vector characteristics and unsupervised K-means clustering, the problem of low rainfall detection precision caused by a fixed threshold value is avoided, the rainfall detection task can be completed according to the relevance coefficient vector characteristics and the extracted clustering class center, the precision requirement of the engineering on radar image rainfall detection can be met, and the radar rainfall detection method can be applied to the engineering in a programmed manner. The advantages of the invention at least include: a threshold value of rainfall detection does not need to be set, the task of rainfall detection can be automatically completed under different rainfall conditions, and the detection precision is high; the difference of the correlation coefficients of the rain-free radar image and the rain radar image in the azimuth direction is fully utilized to extract the vector characteristics of the correlation coefficients, the problem that the detection position is accurately selected in a lagging azimuth angle is solved, and the precision of rainfall detection is improved; after the cluster centroids are extracted offline, the vector features extracted from the radar images are compared with the distance between the centroids to complete the rainfall detection task, the method is low in complexity and small in computation amount, and rainfall detection can be performed online in real time; by setting the clustering class center number of the characteristics, the radar images without rain, light rain, medium rain and heavy rain can be distinguished, so that the detection precision and the reliability of the method are improved; the invention utilizes unsupervised K-means clustering machine learning algorithm, the precision of the detection method is related to the radar sample data, and the reliability and precision of the method can be further improved in the later period on the basis of large data accumulation.
In one embodiment of the present invention, the preprocessing of the acquired radar image comprises the steps of:
carrying out median filtering on the polar coordinate radar image by using a preset template so as to inhibit co-channel interference;
and selecting an effective observation area as an analysis area.
In the embodiment of the invention, the image is preprocessed mainly by filtering image noise and inhibiting the influence of same frequency interference on rainfall detection. As a specific embodiment, the preset template is a 3 × 3 2-D template:
f'(r,θ)＝median{f(r,θ)}
wherein f (r, theta) is the radar image echo strength value at the polar coordinate position (r, theta); f' (r, theta) is an echo intensity value after radar image filtering;
sorting 9 pixel point values in the radar image corresponding to the 3 x 3 median filter template, and assigning the selected middle pixel point value to a radar image coordinate position coincident with the center of the filter template;
traversing the median filter template in the radar image to obtain a radar image after median filtering.
In the embodiment of the invention, the images acquired by the radar can not be used completely, and because the radar scanning range is generally wider, in order to improve the accuracy of image identification, an area which is more effective for observation in the radar scanning image can be selected for identification processing. This is an optional specific implementation and is not used to limit the implementation of the embodiments of the present invention. The effectiveness of the method is verified by data collected by a 760B marine navigation radar in 2013, 12, 14-18 days, and the detection performance of the method is analyzed. FIG. 4 is an original navigation radar image acquired at 21 minutes in 12 months, 15 days and 15 days, wherein the instantaneous rainfall is 2mm/5min, the wind direction and the wind speed are respectively 35 degrees and 13.7m/s, and the effective wave height of sea waves is 2.1 m. A sector area formed by grey thick solid lines in the graph is an effective sea observation area of the X-band navigation radar in the experimental site, and the echo intensity in the sector area is mainly the back scattering of sea clutter and raindrops. Considering the attenuation of radar signals in the distance direction, the texture characteristics of sea clutter, the influence of a coastline and submarine topography, radar images in the areas with the distance direction of 900-1800 m and the azimuth direction of 125-190 degrees are selected for analysis, and the area formed by black bold lines is a radar image analysis area used for calculating a correlation coefficient.
In an embodiment of the present invention, the extracting the features of the radar image includes:
acquiring an autocorrelation coefficient of the radar image in the direction;
and selecting a plurality of autocorrelation coefficients meeting preset conditions, extracting and constructing into correlation coefficient vector characteristics.
In the embodiment of the invention, sea clutters in the navigation radar image are distributed in stripes with alternate light and shade and have strong space-time correlation characteristics, but when high-frequency rainfall interference is introduced into the collected radar image, the radar echo intensity is increased along with the increase of the rainfall intensity, so that the correlation characteristics of the echoes in the radar image are influenced. The autocorrelation coefficient of the radar image echo in the azimuth is as follows:
Figure RE-GDA0003230454850000081
where E (-) is the expected value of the echo intensity in the azimuth direction, τ is the lag azimuth angle in the azimuth direction, θ is the azimuth angle, and x (θ) is the backscatter echo intensity in the azimuth direction. When the correlation coefficient is close to 0, the correlation characteristic of the radar echo signal is weak, and vice versa. Since the correlation coefficients of the radar image with rain and the radar image without rain are different in the azimuth direction, in order to fully utilize the characteristic difference of the characteristic coefficient, the correlation coefficient with a large difference in the azimuth direction is extracted and constructed into the correlation coefficient vector characteristic. Correlation coefficients can be obtained by calculating the echo intensity in the azimuth direction, 120 groups of correlation coefficients can be obtained for the selected analysis area due to the radar image distance resolution rate of 7.5m, then the 120 groups of correlation coefficients are subjected to averaging processing, and the obtained average correlation coefficients are used for subsequent feature extraction. The autocorrelation coefficients of the acquired radar images at the lag azimuth are shown in fig. 5. The solid line with a circle indicates the correlation coefficient of the no-rain radar image, and the dotted line with an asterisk indicates the correlation coefficient of the rain radar image.
In addition, the correlation coefficient of the radar image with rainfall interference and the correlation coefficient of the radar image without rainfall have larger difference at the horizontal beam width, and the correlation coefficients near the horizontal beam width are selected to construct vector characteristics. The correlation coefficients of 0.3 to 1.2 are selected in the azimuthal direction to construct the feature vector. Considering the horizontal beam width and the number of lines of the radar image, the characteristic vector of the selected correlation number in the invention comprises 8 units.
In one embodiment of the present invention, the cluster centroid is determined by:
obtaining a feature data set S ═ S1,s2,…,sNAnd setting a classification number K; wherein N is the number of data in the data set;
arbitrarily selecting m from the feature data set S1(0),m2(0),…,mk(0) As C1，C2，…，CkThe initial class core of the class;
the feature data set S is classified.
In the embodiment of the invention, in order to perform rainfall detection on the radar image, the characteristic class centers of the radar image with rain and the radar image without rain are acquired. And clustering the features extracted from the navigation radar images according to an unsupervised machine learning K-means clustering algorithm and extracting class centers. The K-means clustering algorithm mainly divides feature set data into the categories of rain-free images and rain images through an iterative process. For a given feature data set S, the clustering process seeks the optimal cluster center miTo minimize the cost function:
Figure RE-GDA0003230454850000091
where K is the number of classes into which the data features are divided, miIs the class core of the ith class, where i ═ 1,2, …, K, CiDenotes the ith class, sjIndicates belongings to C in the iterative ProcessiData of a class.
In an embodiment of the present invention, the classifying the feature data set S includes the following steps:
sequentially calculating the Euclidean distance from each feature data to each clustering center;
according to a minimum distance principle, sequentially dividing each feature data into classes with the nearest class centers;
Ciclass center m of class after j iterationsi(j) Comprises the following steps:
Figure RE-GDA0003230454850000092
wherein n isiIs CiThe number of data in the class;
if mi(j)＝mi(j-1), converging the K-means clustering algorithm, and finishing the classification to obtain a clustering center; otherwise, repeating the above steps.
In the embodiment of the invention, K clustering centers can be obtained based on the correlation coefficient vector characteristics in the training set, and the correlation coefficient vector characteristics s in the data set are measuredmAnd the clustering center miThe euclidean distance between them is:
di(sm,mi)＝||sm-mi||
in the formula, let m beiAnd mjRespectively, the cluster center of the image features with the rain radar and the cluster center of the image features without the rain radar. For example, a rainfall detection task with a cluster center number K-3 may be described as:
H：d1＜d2∩d1＜d3
wherein d isiAnd expressing the Euclidean distance between the correlation coefficient vector feature and the clustering center i, wherein i is 1,2 and 3. Phase of phaseWhen the distance between the relevance coefficient vector characteristics and the rainfall-free radar image clustering center 1 is minimum, it is indicated that the radar image does not contain rainfall interference, and then the task of detecting whether rainfall exists or not by utilizing the radar image is accurately completed.
In the embodiment of the invention, the radar image rainfall detection is a method based on correlation coefficient vector characteristics and a K-means clustering technology. When the distance between the vector characteristic of the correlation coefficient in the test data set and the centroid of the image without the rain radar is smaller than the distance between the vector characteristic and the centroid of the image with the rain radar, the secondary radar image is considered to be free of rainfall interference; and vice versa.
The cluster centroids extracted based on the vector feature data of the training set are shown in fig. 6. The solid line with a cross is a cluster centroid 1, which represents the centroid obtained based on the image features of the rain-free radar. The dotted line with circles is the cluster centroid 2, which represents the centroid of the radar image with light rain. The dotted line with the triangle is a cluster centroid 3, which represents the centroid of the radar image containing medium rain and heavy rain.
The distance between the correlation coefficient vector feature and the class center can be calculated based on the extracted cluster class center, as shown in fig. 7. The abscissa represents the number of experiments, and the ordinate is the euclidean distance calculated based on the present invention. The dashed and dotted lines indicate the distance between the correlation coefficient vector features of the radar image and the cluster centroid of the rain radar image. The solid line refers to the distance between the radar image features and the cluster centroid of the radar image without rain under the condition of no rainfall interference. When the extracted correlation coefficient vector features are closest to a certain class center, the acquired radar image belongs to the class.
The technical effect of the present invention is illustrated in a specific embodiment as follows:
the experimental site is selected from a place, and the average water depth in the effective observation area of the radar is about 25 m. The horizontal beam width Δ θ of the radar antenna in the present invention is about 0.9 °, and one radar image generally includes about 3600 data lines in azimuth, so that the data line in the horizontal beam width Δ θ is 9. 1149 images are selected from images acquired by the navigation radar for the ship based on 760B in 2013, 12, 14-18 months, and are used for verifying the effectiveness of the rainfall detection method provided by the invention and analyzing the detection performance of the rainfall detection method. The rainfall intensity recorded synchronously by the rain gauge during the experiment was taken as a reference value, as shown in fig. 8.
The number of the cluster centers is 3, and the radar image without rain can be detected by judging whether the distance from the correlation coefficient vector characteristic in the test data set to the cluster center 1 is minimum or not. When there is a rain disturbance in the detected radar image, the radar image is marked as 1, otherwise, the radar image is marked as 0. The detection results of the method proposed by the present invention are shown in fig. 9. Compared with the rainfall intensity synchronously recorded in fig. 8, it can be seen that the method provided by the invention can complete the rainfall detection task, and can better distinguish the radar image with rain from the radar image without rain under the condition of no rainfall interference.
To further verify the effectiveness of the present invention, the detection result of the present invention is compared with the existing mainstream rainfall detection method, i.e. the zero intensity percentage method, and the zero intensity percentage and the detection result calculated by the zero intensity percentage method are shown in fig. 10 and fig. 11.
Fig. 10 shows the percentage of zero intensity calculated using the radar image, with the threshold value being the dashed line. And when the calculated zero intensity percentage is less than the threshold value, the secondary radar image is a rain image. As can be seen from fig. 10, the percentage of zero intensity is relatively small when the radar image contains rain disturbances. The zero intensity percentage of the radar image decreases with increasing rainfall intensity. The percentage of zero intensity containing the rain fall radar image fluctuates around the threshold.
FIG. 11 shows the results of the zero intensity percentage method. When the radar image is detected to contain rainfall, marking the image as 1; otherwise the flag is 0. As can be seen by comparing with fig. 9, there is a large difference in the detection result of the zero intensity percentage method near the 800 th radar image.
The invention provides a radar rainfall detection method based on the combination of navigation radar image correlation coefficient vector characteristics and K mean value clustering, and firstly, aiming at the difference of correlation coefficients between a rain radar image and a rain-free radar image, the correlation coefficient vector characteristics are extracted from a radar image analysis area; secondly, clustering the extracted correlation coefficient vector characteristics by using a K-means clustering algorithm to obtain clustering class centers of the radar images without rain, light rain, medium rain and heavy rain; and finally, measuring and comparing the distance between the radar data of the test set and the class centers of the radar images of no rain, light rain, medium rain and heavy rain, and finishing the task of rainfall detection by judging whether the vector characteristics of the radar images of no rain are nearest or not. The radar image rainfall detection precision statistical results of the method and the zero intensity percentage method provided by the invention are shown in table 1. The rainfall detection success rate of the invention for training set and test set with rain data respectively reaches 89.6% and 93.2%. The method provided by the invention does not need to set a threshold value, the precision is improved compared with the existing zero-intensity-percentage rainfall detection method, the rainfall detection task can be automatically completed under different rainfall conditions, and the detection precision completely meets the engineering application requirements.
Table 1: radar image rainfall detection statistical result
Figure BDA0003063660160000101
As shown in fig. 12, in an embodiment, a radar rainfall detection device is provided, which may be integrated in the computer device 200, and specifically may include:
an obtaining module 1201, configured to obtain a radar image and perform preprocessing;
a feature extraction module 1202, configured to perform feature extraction on the radar image;
a calculating module 1203, configured to calculate distances between the features of the radar image and a plurality of preset cluster centroids, where the cluster centroids include a cluster centroid of a rain radar image and a cluster centroid of a rain-free radar image;
a judging module 1204, configured to judge whether the cluster centroid with the smallest distance between the feature of the radar image and the cluster centroid is the cluster centroid of the radar image with rain; and if so, the acquired radar image is a radar image with rain, otherwise, the radar image without rain is acquired.
In the embodiment of the present invention, please refer to the contents of the method part of the present invention for the explanation of the steps executed by the above modules, which is not described again in the embodiment of the present invention.
FIG. 13 is a diagram illustrating an internal structure of a computer device in one embodiment. The computer device may specifically be the computer device 200 in fig. 1. As shown in fig. 13, the computer apparatus includes a processor, a memory, a network interface, an input device, and a display screen connected through a system bus. Wherein the memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system and may also store a computer program, and when the computer program is executed by a processor, the computer program may enable the processor to implement the radar rainfall detection method provided by the embodiment of the present invention. The internal memory may also store a computer program, and when the computer program is executed by the processor, the processor may execute the radar rainfall detection method provided by the embodiment of the present invention. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
Those skilled in the art will appreciate that the architecture shown in fig. 13 is merely a block diagram of some of the structures associated with the inventive arrangements and is not intended to limit the computing devices to which the inventive arrangements may be applied, as a particular computing device may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, the radar rainfall detection apparatus provided by the embodiment of the present invention can be implemented in the form of a computer program, and the computer program can be run on a computer device as shown in fig. 13. The memory of the computer device may store various program modules constituting the radar rainfall detection apparatus, such as an acquisition module, a feature extraction module, a calculation module, and a judgment module shown in fig. 12. The program modules constitute computer programs that cause the processor to execute the steps of the radar rainfall detection method of the various embodiments of the present invention described in this specification.
For example, the computer device shown in fig. 13 may perform step S202 by the acquisition module in the radar rainfall detection apparatus shown in fig. 12; the computer device may perform step S204 through the feature extraction module; the computer device may perform step S206 by the computing module; the computer device may perform step S208 through the determination module.
In one embodiment, a computer device is proposed, the computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the following steps when executing the computer program:
acquiring a radar image and preprocessing the radar image;
extracting features of the radar image;
calculating the distance between the features of the radar images and a plurality of preset clustering centers, wherein the clustering centers comprise a clustering center of a rain radar image and a clustering center of a rain-free radar image;
judging whether the cluster center with the minimum distance between the features of the radar image and the cluster center is the cluster center of the radar image with rain; if so, the obtained radar image is a radar image with rain, otherwise, the radar image without rain is obtained.
In one embodiment, a computer readable storage medium is provided, having a computer program stored thereon, which, when executed by a processor, causes the processor to perform the steps of:
acquiring a radar image and preprocessing the radar image;
extracting features of the radar image;
calculating the distance between the features of the radar images and a plurality of preset clustering centers, wherein the clustering centers comprise a clustering center of a rain radar image and a clustering center of a rain-free radar image;
judging whether the cluster center with the minimum distance between the features of the radar image and the cluster center is the cluster center of the radar image with rain; if so, the obtained radar image is a radar image with rain, otherwise, the radar image without rain is obtained.
It should be understood that, although the steps in the flowcharts of the embodiments of the present invention are shown in sequence as indicated by the arrows, the steps are not necessarily performed in sequence as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a portion of the steps in various embodiments may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performing the sub-steps or stages is not necessarily sequential, but may be performed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program that can be stored in a non-volatile computer-readable storage medium and can include the processes of the embodiments of the methods described above when executed. Any reference to memory, storage, databases, or other media used in embodiments provided herein may include non-volatile and/or volatile memory. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims (10)

1. A radar rainfall detection method is characterized by comprising the following steps:
acquiring a radar image and preprocessing the radar image;
extracting features of the radar image;
calculating the distance between the features of the radar images and a plurality of preset clustering centers, wherein the clustering centers comprise the clustering center of the rain radar images and the clustering center of the rain-free radar images;
judging whether the clustering center with the minimum distance between the features of the radar images and the clustering center is the clustering center of the radar image with rain; if so, the obtained radar image is a radar image with rain, otherwise, the radar image without rain is obtained.
2. The radar rainfall detection method of claim 1, wherein the preprocessing of the acquired radar image comprises the steps of:
carrying out median filtering on the polar coordinate radar image by using a preset template so as to inhibit co-channel interference;
and selecting an effective observation area as an analysis area.
3. The radar rainfall detection method of claim 1, wherein the preset template is a 3 x 3 2-D template:
f'(r,θ)＝median{f(r,θ)}
wherein f (r, theta) is the radar image echo strength value at the polar coordinate position (r, theta); f' (r, theta) is an echo intensity value after radar image filtering;
sorting 9 pixel point values in the radar image corresponding to the 3 x 3 median filter template, selecting a middle pixel point value and assigning the middle pixel point value to a radar image coordinate position coincident with the center of the filter template;
traversing the median filter template in the radar image to obtain a radar image after median filtering.
4. The radar rainfall detection method of claim 1, wherein said feature extracting said radar image comprises the steps of:
acquiring an autocorrelation coefficient of the radar image in the direction;
and selecting a plurality of autocorrelation coefficients meeting preset conditions, extracting and constructing into correlation coefficient vector characteristics.
5. The radar rainfall detection method of claim 1, wherein the cluster centroid is determined by:
obtaining a feature data set S ═ S1,s2,…,sNAnd setting a classification number K; wherein N is the number of data in the data set;
arbitrarily selecting m from the feature data set S1(0),m2(0),…,mk(0) As C1，C2，…，CkAn initial class core of the class;
the feature data set S is classified.
6. The radar rainfall detection method of claim 5 wherein said classifying of said feature data set S comprises the steps of:
sequentially calculating the Euclidean distance from each feature data to each clustering center;
according to a minimum distance principle, sequentially dividing each feature data into classes with the nearest class centers;
Ciclass center m of class after j iterationsi(j) Comprises the following steps:
Figure RE-FDA0003230454840000021
wherein n isiIs CiThe number of data in the class;
if mi(j)＝mi(j-1), converging the K-means clustering algorithm, and finishing the classification to obtain a clustering center; otherwise, repeating the above steps.
7. The radar rain detection method of claim 6, wherein correlation coefficient vector feature s in a metric datasetmAnd the clustering center miThe euclidean distance between them is:
di(sm,mi)＝||sm-mi||
in the formula, miAnd mjRespectively, a cluster center of the image features of the rain radar and a cluster center of the image features of the rain radar.
8. A radar rainfall detection device, characterized in that, radar rainfall detection device includes:
the acquisition module is used for acquiring a radar image and preprocessing the radar image;
the characteristic extraction module is used for extracting the characteristics of the radar image;
the computing module is used for computing the distance between the features of the radar images and a plurality of preset clustering centers, wherein the clustering centers comprise the clustering center of a rain radar image and the clustering center of a rain-free radar image;
the judging module is used for judging whether the clustering center with the minimum distance between the features of the radar images and the clustering center is the clustering center with the rain radar images; if so, the obtained radar image is a radar image with rain, otherwise, the radar image without rain is obtained.
9. A computer arrangement, comprising a memory and a processor, the memory having stored therein a computer program that, when executed by the processor, causes the processor to carry out the steps of the radar rain detection method according to any one of claims 1 to 7.
10. A computer-readable storage medium, having a computer program stored thereon, which, when executed by a processor, causes the processor to perform the steps of the radar rain detection method according to any one of claims 1 to 7.
CN202110520287.0A 2021-05-13 2021-05-13 Radar rainfall detection method and device, computer equipment and storage medium Active CN113450308B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110520287.0A CN113450308B (en) 2021-05-13 2021-05-13 Radar rainfall detection method and device, computer equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110520287.0A CN113450308B (en) 2021-05-13 2021-05-13 Radar rainfall detection method and device, computer equipment and storage medium

Publications (2)

Publication Number Publication Date
CN113450308A true CN113450308A (en) 2021-09-28
CN113450308B CN113450308B (en) 2022-10-28

Family

ID=77809724

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110520287.0A Active CN113450308B (en) 2021-05-13 2021-05-13 Radar rainfall detection method and device, computer equipment and storage medium

Country Status (1)

Country Link
CN (1) CN113450308B (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150145717A1 (en) * 2013-11-26 2015-05-28 Korea Institute Of Construction Technology Integrated rainfall estimation method using x-band dual-polarimetric radar measurement data
US20170371074A1 (en) * 2016-06-24 2017-12-28 Climacell Inc. Real-Time Precipitation Forecasting System
CN109116359A (en) * 2018-09-28 2019-01-01 西北工业大学 A kind of estimation method of airborne radar low-level wind shear wind field echo wind speed
CN111830506A (en) * 2020-07-22 2020-10-27 江苏科技大学 K-means clustering algorithm-based sea surface wind speed method
CN112633595A (en) * 2020-12-31 2021-04-09 南京师范大学 Rainfall station observation network design method based on radar rainfall data mining

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150145717A1 (en) * 2013-11-26 2015-05-28 Korea Institute Of Construction Technology Integrated rainfall estimation method using x-band dual-polarimetric radar measurement data
US20170371074A1 (en) * 2016-06-24 2017-12-28 Climacell Inc. Real-Time Precipitation Forecasting System
CN109116359A (en) * 2018-09-28 2019-01-01 西北工业大学 A kind of estimation method of airborne radar low-level wind shear wind field echo wind speed
CN111830506A (en) * 2020-07-22 2020-10-27 江苏科技大学 K-means clustering algorithm-based sea surface wind speed method
CN112633595A (en) * 2020-12-31 2021-04-09 南京师范大学 Rainfall station observation network design method based on radar rainfall data mining

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
XINWEI CHEN等: "Identification of Rain and Low-Backscatter Regions in X-Band Marine Radar Images: An Unsupervised Approach", 《IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING》 *
YANG HONG等: "Precipitation Estimation from Remotely Sensed Imagery Using an Artificial Neural Network Cloud Classification System", 《JOURNAL OF APPLIED METEOROLOGY》 *
卫延波: "X波段导航雷达图像反演海面高程技术研究", 《中国优秀博士学位论文全文数据库信息科技辑》 *
吕博群: "基于航海雷达图像反演的降雨检测技术研究", 《万方学位论文全文数据库》 *
沈继红等: "X-band雷达图像中降雨干扰的识别与抑制", 《光学精密工程》 *

Also Published As

Publication number Publication date
CN113450308B (en) 2022-10-28

Similar Documents

Publication Publication Date Title
Kang et al. A modified faster R-CNN based on CFAR algorithm for SAR ship detection
CN107341488B (en) SAR image target detection and identification integrated method
Guo et al. Scattering enhanced attention pyramid network for aircraft detection in SAR images
Zheng et al. Development of a gray-level co-occurrence matrix-based texture orientation estimation method and its application in sea surface wind direction retrieval from SAR imagery
Graziano et al. Ship heading and velocity analysis by wake detection in SAR images
CN108171193B (en) Polarized SAR (synthetic aperture radar) ship target detection method based on super-pixel local information measurement
CN110208806B (en) Marine radar image rainfall identification method
Silva et al. Computer-based identification and tracking of Antarctic icebergs in SAR images
CN111239704B (en) Atmosphere detection radar signal identification method, device, equipment and medium
Huang et al. Wave height estimation from X-band nautical radar images using temporal convolutional network
CN115184915A (en) Sea clutter suppression method and system based on random clutter loitering behavior
Ahmadvand et al. 1D-CNNs for autonomous defect detection in bridge decks using ground penetrating radar
Huang et al. Change detection method based on fractal model and wavelet transform for multitemporal SAR images
CN113450308B (en) Radar rainfall detection method and device, computer equipment and storage medium
CN107729903A (en) SAR image object detection method based on area probability statistics and significance analysis
CN113822361B (en) SAR image similarity measurement method and system based on Hamming distance
Kvasnov et al. A classification technique of civil objects by artificial neural networks using estimation of entropy on synthetic aperture radar images
Arivazhagan et al. Optimal Gabor sub-band-based spectral kurtosis and Teager Kaiser energy for maritime target detection in SAR images
Yuan et al. A Novel Method for Pavement Transverse Crack Detection Based on 2D Reconstruction of Vehicle Vibration Signal
Zhang et al. Marine radar monitoring IoT system and case study of target detection based on PPI images
CN114724036A (en) Ice crack detection method based on multi-scale learning
Lu et al. Research on rainfall identification based on the echo differential value from X-band navigation radar image
CN113298759A (en) Water area detection method and device, electronic equipment and storage medium
CN113869119A (en) Multi-temporal SAR ship target tracking method, system, equipment and medium
CN112686222A (en) Method and system for detecting ship target by satellite-borne visible light detector

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant