CN104237890A - Recognition and forecast method for rainstorm caused by train effect - Google Patents
Recognition and forecast method for rainstorm caused by train effect Download PDFInfo
- Publication number
- CN104237890A CN104237890A CN201410446007.6A CN201410446007A CN104237890A CN 104237890 A CN104237890 A CN 104237890A CN 201410446007 A CN201410446007 A CN 201410446007A CN 104237890 A CN104237890 A CN 104237890A
- Authority
- CN
- China
- Prior art keywords
- train effect
- echo
- monomer
- train
- monomers
- 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
Links
- 230000000694 effects Effects 0.000 title claims abstract description 136
- 238000000034 method Methods 0.000 title claims abstract description 62
- 238000002592 echocardiography Methods 0.000 claims abstract description 15
- 238000013213 extrapolation Methods 0.000 claims abstract description 13
- 238000012216 screening Methods 0.000 claims abstract description 5
- 239000000178 monomer Substances 0.000 claims description 151
- 238000002310 reflectometry Methods 0.000 claims description 20
- 238000007781 pre-processing Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 238000011160 research Methods 0.000 claims description 4
- 230000002093 peripheral effect Effects 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 3
- 238000013479 data entry Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008033 biological extinction Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000029087 digestion Effects 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000013277 forecasting method Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 230000004660 morphological change Effects 0.000 description 1
- 238000001556 precipitation Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/88—Radar or analogous systems specially adapted for specific applications
- G01S13/95—Radar or analogous systems specially adapted for specific applications for meteorological use
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S7/415—Identification of targets based on measurements of movement associated with the target
-
- 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
- G06F18/211—Selection of the most significant subset of features
- G06F18/2111—Selection of the most significant subset of features by using evolutionary computational techniques, e.g. genetic algorithms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/66—Analysis of geometric attributes of image moments or centre of gravity
-
- 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)
- Physics & Mathematics (AREA)
- Remote Sensing (AREA)
- Radar, Positioning & Navigation (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Networks & Wireless Communication (AREA)
- Health & Medical Sciences (AREA)
- Physiology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Electromagnetism (AREA)
- Geometry (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention discloses a recognition and forecast method for a rainstorm caused by the train effect. The method includes the steps that real-time radar data are preprocessed; single bodies subjected to preliminary area screening are subjected to preliminary fitting, interference is eliminated, and a suspected train effect area is searched for; after the suspected train effect area is determined, the train effect can be automatically recognized; through the moving distance of the center of mass of banded echoes of the suspected area and the rotating angle of the axis of the banded echoes of the suspected area, banded echo tracing is performed, and therefore the train effect can be traced; according to the moving inertia of the single bodies, the moving direction, the moving speed and the shape change of the suspected train effect area are extrapolated. The method has the advantages that when the train effect weather occurs, the train effect can be recognized, an alarm is given to a forecaster, extrapolation can be performed according to the train effect weather conditions at various moments, important reference is provided for the forecaster to perform subsequent forecast, and accuracy and rapidity are high.
Description
Technical Field
The invention relates to the field of weather, in particular to a method for identifying and forecasting rainstorm by train effect.
Background
In cities, the strong convection weather of small and medium sizes often causes violent weather phenomena, such as local heavy storm, strong wind, hail and other disastrous weather, and seriously threatens the life and property safety of people. The weather radar is a main means for detecting a precipitation system and is a main tool for monitoring and early warning strong convection weather. The Doppler weather radar has all-weather detection capability and very abundant dynamic information reflecting the atmospheric layer cloud and rain digestion evolution process, greatly enhances the detection and early warning capability of medium and small scale weather systems, and lays a solid foundation for monitoring and forecasting short-time disastrous weather.
The 'train effect' means that a certain area frequently and continuously generates convection monomers with small space scale within a period of time, each generated convection monomer moves along a certain direction, then new convection monomers are generated at the same place and continuously move along the same direction, so that an arrangement similar to a 'train' is formed by arranging a series of convection monomers, and the 'train' can continuously and continuously affect a certain area at the downstream of the 'train', so that strong rainfall is caused, and the 'train effect' is a main echo characteristic of disasters such as flood and the like caused by extreme rainfall.
In recent years, along with the gradual establishment of Doppler weather radar networks in China, the monitoring and early warning functions of Doppler radars on disastrous weather are increasingly highlighted, so that the accuracy of the disastrous weather forecast in China is improved on the original basis, but if the functions of the Doppler radar in the weather monitoring and early warning are to be fully exerted, corresponding storm identification, tracking and forecasting algorithms based on Doppler weather radar data are matched with the storm identification, tracking and forecasting algorithms. At present, the method can be applied to a recognition tracking algorithm of train effect: thunderstorm identification tracking analysis forecast (TITAN), storm identification tracking algorithm (SCIT). The TITAN algorithm is used for identifying the strong convection storm by using a single reflectivity factor threshold, is suitable for the whole storm zone, but has less intensity threshold, and cannot well extract parameters in the storm and distinguish the fine structure of a storm cluster. The SCIT algorithm adopts seven broad values to identify the mass center of the storm, so that the storm monomers in the storm cluster can be well identified, but only the identification result with a high threshold value is retained, the identification result with a low threshold value is abandoned, and a large amount of storm body structure information is lost.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a rainstorm forecasting method which can not only identify the train effect to give an alarm to a forecaster when the train effect phenomenon occurs in real-time weather monitoring and forecasting, but also can extrapolate according to the weather conditions of the train effect at a plurality of moments.
The technical scheme adopted by the invention is as follows: a method for identifying and forecasting rainstorm caused by train effect comprises the following steps: step S1, preprocessing the real-time radar data: eliminating super refraction, fine monomer and false merging; step S2, carrying out primary fitting on the monomer subjected to the area primary screening: eliminating interference, clustering the result, performing secondary fitting on the clustering result, and searching a suspected area of the train effect; step S3, after determining the suspected area of "train effect": automatically recognizing the train effect, if so, judging the train effect as the train effect, performing S4, otherwise, skipping to the radar data processing at the next moment, namely, starting from S1; step S4, tracking the strip echo using the centroid movement distance and the axis rotation angle of the suspected area strip echo: after the position of the whole strip echo is determined, monomer tracking in a suspected area is carried out by using the Hu moment and the form change, and the tracking of a train effect is realized; and step S5, extrapolating the movement direction, the movement speed and the shape change of the suspected area of the train effect according to the movement inertia of the single body.
The step of preprocessing the real-time radar data in S1 is: s101: by raising the elevation angle of the radar echo data, super-refraction echoes with strong interference on an automatic identification stage are filtered; s102: deleting the fine monomers which do not meet the threshold condition by setting an area threshold; s103: judging whether the reflectivity of the target connected region is greater than 40dBZ, if so, corroding one pixel point, and corroding a strong echo region less; if the detected signal is less than 40dBZ, corroding two pixel points, corroding a weak echo region a little more, then improving a first-level threshold value, judging whether a monomer with a higher first-level threshold value can be detected, if the monomer can be detected, expanding, repeating the process, and if the monomer cannot be detected, stopping.
The step S2 specifically includes the following steps: s201: firstly, a centroid method is used for obtaining representative points of each monomer, then straight line fitting is carried out on the points to obtain a fitting straight line of the train effect, then the monomer which deviates from the fitting straight line to a large extent is deleted, and interference of peripheral monomers irrelevant to the train effect is eliminated; s202: obtaining monomers meeting requirements after preliminary fitting, extracting the mass centers of the monomers, automatically clustering by using a neighbor propagation method, then respectively carrying out straight line fitting on each type of point set again, and deleting the monomers with larger deviation from the fitting straight line degree to obtain a suspected area of the train effect.
The step S3 specifically includes the following steps: s301: extracting information of the whole strip-shaped echo in the suspected area, judging whether the length-width ratio of the minimum external rectangle of the whole strip-shaped echo, the average reflectivity and the average speed of the whole strip-shaped echo meet the standard of the train effect, if the length-width ratio of the minimum external rectangle of the whole strip-shaped echo, the average reflectivity and the average speed meet the standard of the train effect, continuing the following judging process, if the length-width ratio does not meet the standard of the train effect, judging that the length-width ratio is not the train effect, then discarding the length-width ratio; s302: extracting various characteristics of all monomers in the suspected area, putting the characteristics into a rule base, judging whether the monomer echoes meet a 'train effect' monomer echo condition or not by using a standard voting method, counting the proportion of the monomers meeting the 'train effect' monomer echo condition in all the monomers after judging and identifying all the monomers in the suspected area, and judging the 'train effect' if the conditions meet a threshold value.
The step S4 specifically includes the following steps: s401: when the train effect is tracked, the attributes of the train effect integral strip echo and the single echo are fused; s402: calculating the centroid position and the long axis direction of the whole strip echo by taking the whole strip echo with the train effect as a research object, finding out all possible motion path combinations of the whole strip echo between two adjacent moments, calculating the moving distance of the centroid of the strip echo and the rotating angle of the long axis under each path combination, and selecting the shortest path combination as the moving path of the echo band in the path combinations meeting the conditions; s403: after the position of the whole band echo is determined, calculating the area difference between each monomer at the current moment and each monomer at the previous moment aiming at the same whole band echo, if the area difference is smaller than a threshold value, storing the monomers at the previous moment as possible matching with the current monomer, then calculating the similarity degree of the current monomer and the outlines of all the possible matched monomers through Hu moment, and considering the combination with the maximum similarity degree as the same monomer, and marking the same ID (identity) with the same monomer; s404: based on the recognition and judgment results of the train effect of the primary pair, the area range is defined, the secondary operation time is shortened, and the accuracy is improved; s405: the information between the integral belt with the maximum relevance and each monomer in the belt is obtained by comparing various characteristics of the front time sequence monomer and the rear time sequence monomer, comparing integral belt characteristics, comparing the long axis direction of the integral belt unique to the train effect with the moving direction, and comparing the speed with the moving direction, so that a one-to-one correspondence relationship is formed.
The step S5 specifically includes the following steps: s501: continuously inputting 'train effect' data at three moments, acquiring relevant characteristics, constructing a time sequence of the continuous moments, establishing a monomer family relation, checking and comparing the similarity of adjacent time monomers, and identifying a motion track of the monomer which develops and disappears through calculation so as to predict the motion trend of the monomer; s502: firstly, extrapolating the integral band echo, respectively determining the moving speed, the moving direction and the long axis direction of the integral band at the next moment according to the average speed, the moving direction and the long axis direction of the integral band among three recorded continuous moments according to the inertia of the movement of the single body, and integrating the three information to finish the tracking of the integral band echo with the train effect; s503: on the basis of extrapolating the position of the whole strip echo, the position of each single echo is extrapolated by the same method.
The three times were extrapolated for 6 minutes, 12 minutes, and 18 minutes, respectively.
The invention has the beneficial effects that: when the weather of the train effect phenomenon occurs, the train effect can be identified to give an alarm to a forecaster, extrapolation can be performed according to the weather condition of the train effect at multiple moments, important reference is provided for the forecaster to perform subsequent forecasting, and accuracy and rapidity are high.
Drawings
FIG. 1 is a graph a showing the original reflectivity, and a graph b showing the reflectivity after the super-refraction is filtered;
FIG. 2 is a preliminary fit plot of the residual monomer after area prescreening;
in FIG. 3, the graphs a and b are the original contour and its corresponding representative points, respectively;
FIG. 4 is the auto-clustering map of FIG. 3;
in fig. 5, the images a and b are respectively the images after quadratic fitting of each local region;
FIG. 6 is a typical "train effect" diagram where there are two processes;
in FIG. 7, the graphs a, b and c are respectively the comparison graph of the real graph of the "train effect" at 6 minutes, 12 minutes and 18 minutes and the extrapolated graph,
the images are compared with real images and external images, wherein a1, b1, c1 and a3, b3 and c3 are real reflectivity maps of 6, 12 and 18 minutes 'train effect' and partial enlarged images thereof, and a2, b2, c2 and a4, b4 and c4 are extrapolation maps of 6, 12 and 18 minutes 'train effect' and partial enlarged images thereof.
Detailed Description
The invention is described in further detail below with reference to the following figures and detailed description:
a method for identifying and forecasting rainstorm caused by train effect comprises the following steps:
step S1, preprocessing the real-time radar data: eliminating super refraction, fine monomer and false merging; s101: by raising the elevation angle of the radar echo data, super-refraction echoes with strong interference on an automatic identification stage are filtered; s102: deleting the fine monomers which do not meet the threshold condition by setting an area threshold; s103: judging whether the reflectivity of the target connected region is greater than 40dBZ, if so, corroding one pixel point, and corroding a strong echo region less; if the detected signal is less than 40dBZ, corroding two pixel points, corroding a weak echo region a little more, then improving a first-level threshold value, judging whether a monomer with a higher first-level threshold value can be detected, if the monomer can be detected, expanding, repeating the process, and if the monomer cannot be detected, stopping.
Step S2, carrying out primary fitting on the monomer subjected to the area primary screening: eliminating interference, clustering the result, performing secondary fitting on the clustering result, and searching a suspected area of the train effect; s201: firstly, a centroid method is used for obtaining representative points of each monomer, then straight line fitting is carried out on the points to obtain a fitting straight line of the train effect, then the monomer which deviates from the fitting straight line to a large extent is deleted, and interference of peripheral monomers irrelevant to the train effect is eliminated; s202: obtaining monomers meeting requirements after preliminary fitting, extracting the mass centers of the monomers, automatically clustering by using a neighbor propagation method, then respectively carrying out straight line fitting on each type of point set again, and deleting the monomers with larger deviation from the fitting straight line degree to obtain a suspected area of the train effect.
Step S3, after determining the suspected area of "train effect": automatically identifying the train effect, if so, judging the train effect to be the train effect, performing S4, and if not, skipping to the radar data processing at the next moment, namely, starting from S1; s301: extracting information of the whole strip-shaped echo in the suspected area, judging whether the length-width ratio of the minimum external rectangle of the whole strip-shaped echo, the average reflectivity and the average speed of the whole strip-shaped echo meet the standard of the train effect, if the length-width ratio of the minimum external rectangle of the whole strip-shaped echo, the average reflectivity and the average speed meet the standard of the train effect, continuing the following judging process, if the length-width ratio does not meet the standard of the train effect, judging that the length-width ratio is not the train effect, then discarding the length-width ratio; s302: extracting various characteristics of all monomers in the suspected area, putting the characteristics into a rule base, judging whether the monomer echoes meet a 'train effect' monomer echo condition or not by using a standard voting method, counting the proportion of the monomers meeting the 'train effect' monomer echo condition in all the monomers after judging and identifying all the monomers in the suspected area, and judging the 'train effect' if the conditions meet a threshold value.
Step S4, tracking the strip echo using the centroid movement distance and the axis rotation angle of the suspected area strip echo: after the position of the whole strip echo is determined, monomer tracking in a suspected area is carried out by using the Hu moment and the form change, and the tracking of a train effect is realized; s401: when the train effect is tracked, the attributes of the train effect integral strip echo and the single echo are fused; s402: calculating the centroid position and the long axis direction of the whole strip echo by taking the whole strip echo with the train effect as a research object, finding out all possible motion path combinations of the whole strip echo between two adjacent moments, calculating the moving distance of the centroid of the strip echo and the rotating angle of the long axis under each path combination, and selecting the shortest path combination as the moving path of the echo band in the path combinations meeting the conditions; s403: after the position of the whole band echo is determined, calculating the area difference between each monomer at the current moment and each monomer at the previous moment aiming at the same whole band echo, if the area difference is smaller than a threshold value, storing the monomers at the previous moment as possible matching with the current monomer, then calculating the similarity degree of the current monomer and the outlines of all the possible matched monomers through Hu moment, and considering the combination with the maximum similarity degree as the same monomer, and marking the same ID (identity) with the same monomer; s404: based on the recognition and judgment results of the train effect of the primary pair, the area range is defined, the secondary operation time is shortened, and the accuracy is improved; s405: the information between the integral belt with the maximum relevance and each monomer in the belt is obtained by comparing various characteristics of the front time sequence monomer and the rear time sequence monomer, comparing integral belt characteristics, comparing the long axis direction of the integral belt unique to the train effect with the moving direction, and comparing the speed with the moving direction, so that a one-to-one correspondence relationship is formed.
And step S5, extrapolating the movement direction, the movement speed and the shape change of the suspected area of the train effect according to the movement inertia of the single body. S501: continuously inputting 'train effect' data at three moments, acquiring relevant characteristics, constructing a time sequence of the continuous moments, establishing a monomer family relation, checking and comparing the similarity of adjacent time monomers, and identifying a motion track of the monomer which develops and disappears through calculation so as to predict the motion trend of the monomer; s502: firstly, extrapolating the integral band echo, respectively determining the moving speed, the moving direction and the long axis direction of the integral band at the next moment according to the average speed, the moving direction and the long axis direction of the integral band among three recorded continuous moments according to the inertia of the movement of the single body, and integrating the three information to finish the tracking of the integral band echo with the train effect; s503: on the basis of extrapolating the position of the whole strip echo, extrapolating the position of each single echo by the same method; the three times were extrapolated for 6 minutes, 12 minutes, and 18 minutes, respectively.
Example (b):
as shown in the process of fig. 1, panels a-b, the real-time radar data is preprocessed to eliminate super-refraction, subtle singles, and spurious mergers. By raising the elevation angle of the radar echo data, super-refraction echoes with strong interference on an automatic identification stage are filtered; deleting the fine monomers which do not meet the threshold condition by setting an area threshold; and judging whether the reflectivity of the target connected region is greater than 40dBZ, corroding one pixel point if the reflectivity of the target connected region is greater than 40dBZ, and corroding two pixel points if the reflectivity of the target connected region is less than 40dBZ so as to corrode a strong echo region less and a weak echo region more. And then, identifying by using a higher threshold, judging whether the monomer with the higher threshold can be detected or not, if so, expanding, repeating the process, and if not, stopping.
As shown in fig. 2 to 5, the monomers subjected to area preliminary screening are preliminarily fitted to eliminate interference, the results are clustered by using a neighbor propagation method, and then the clustering results are secondarily fitted to search for a suspected train effect area. The specific process is as follows: calculate twoThe centroid of each monomer profile after being valued is determined according to Q ═ Sigma (y)i-a-bxi)2Where Q is the sum of the residual squares of the fitted straight lines formed, a represents the intercept of the fitted straight line, <math>
<mrow>
<mi>a</mi>
<mo>=</mo>
<mover>
<mi>y</mi>
<mo>‾</mo>
</mover>
<mo>-</mo>
<mi>b</mi>
<mover>
<mi>x</mi>
<mo>‾</mo>
</mover>
<mo>;</mo>
</mrow>
</math> b is the fit and slope of the line <math>
<mrow>
<mi>b</mi>
<mo>=</mo>
<mi>Σ</mi>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mover>
<mi>x</mi>
<mo>‾</mo>
</mover>
<mo>)</mo>
</mrow>
<mrow>
<mo>(</mo>
<msub>
<mi>y</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mover>
<mi>y</mi>
<mo>‾</mo>
</mover>
<mo>)</mo>
</mrow>
<mo>/</mo>
<mi>Σ</mi>
<msup>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mover>
<mi>x</mi>
<mo>‾</mo>
</mover>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
<mo>,</mo>
</mrow>
</math> In order to minimize the target residual sum of squares, the partial derivative can be calculated by:and obtaining a fitting straight line, and obtaining the fitting straight line and two end points. The thick line in fig. 2 is a straight line obtained by fitting each centroid of the single body; and calculating the slope formed by each centroid and the two end points, comparing the slope with a fitting straight line, and deleting the profile corresponding to the centroid of which the difference is greater than the set threshold value. Aiming at each clustered monomerThe formed blind processes are respectively fitted to obtain different fitted straight lines, fig. 3 to 4 are a clustering process, the centroid of each monomer is obtained through primary fitting in fig. 3, fig. 4 is a clustering result, and fig. 5 is a result of secondary fitting after classification. Through the processing of the process, the interference factors in the train effect identification process can be greatly eliminated, and the suspected area of the train effect is obtained.
In the process, a neighbor propagation method is adopted for the selection of the clustering algorithm. The theoretical basis is a similarity matrix of data points. For a large-scale data set, the neighbor propagation method can perform clustering quickly and efficiently, and the result is ideal. The core of the method is to find an optimal representative point of a target area. This class represents points such that the sum of the similarities to all data in this region is maximized, they can be grouped into a class. Unlike other clustering algorithms, the neighbor propagation method assumes all data points as class representative points when clustering is started, so that the situations of receiving initial value setting and artificial participation in clustering results cannot occur in the subsequent clustering process, and the neighbor propagation method is an important link for realizing automatic identification and tracking of the train effect. The formula is as follows:
where s (i, k) represents data point xkTo what extent data x is suitable foriThe class of (1) represents a point; representative matrix R ═ R (i, k)]n×nAnd the fitting matrix a ═ a (i, k)]n×nAnd r (i, k) represents the data point xkAs data point xiA (i, k) represents the data point xiSelecting a data point xkAs the suitability degree of its representative point, the sum of these two degrees needs to be calculated:
by solving the optimal representative point of the class, the automatic clustering can be carried out without manually indicating the number of the required clusters at the initial clustering stage. And extracting the central point data obtained by the outline through the reflectivity image, enabling the information of the adjacent points to form mutual influence, and automatically clustering in real time so as to meet the current situation of a plurality of 'train effect' processes. The references are: xiaoyu, yu, semi-supervised clustering based on neighbor propagation algorithm [ J ]. software science, 2008, 19 (11): 2803-2813. the clustering process of fig. 3 to 4 is implemented by using the algorithm described above.
As shown in fig. 6, after the suspected area is determined, the train effect can be automatically identified. After each suspected train effect area is obtained, extracting the information of the whole strip-shaped echo in the suspected area, and judging whether the aspect ratio of the minimum external rectangle of the whole strip-shaped echo is between 1.5:1 and 3: 1; whether the average reflectivity of the whole strip echo is between 30dbz and 45 dbz; whether the moving speed is less than 1.5 km/min or not, if so, extracting all the characteristics of all the monomers in the suspected area, putting the monomers into a rule base for judgment by using a standard voting method, seeing whether the echoes of the monomers meet the train effect monomer echo condition or not, counting the proportion of the monomers meeting the rule base after judging and identifying all the monomers in the suspected area, and if the proportion of the monomers meeting the condition to the total number of the monomers exceeds 2/3, judging the train effect; if these 3 criteria are not satisfied, it is determined that the train effect is not present. The result of the automatic identification is shown in region I, II of fig. 6, which is a typical "train effect" of two processes.
Respectively extracting the whole strip echo information and the monomer characteristics in the suspected area, combining the tracking of the strip echo with the tracking of the monomer, and tracking the train effect: firstly, the centroid moving distance and the axis rotation angle of the suspected area strip echo are utilized to track the strip echo. After the position of the whole strip echo is determined, monomer tracking in the suspected area is carried out by using the Hu moment and the morphological change. And completing one-time complete data entry, judging whether a train effect phenomenon exists or not, if so, judging that several processes exist at the same time, and recording various characteristics of each monomer and whole belt of the target area. When the train effect is tracked, the attributes of the train effect integral strip echo and the single echo are fused, so that the accurate tracking of the train effect is realized; firstly, starting from the overall angle, determining the position of the overall strip echo, and the specific process is as follows:
A. calculating the centroid position and the long axis direction of the whole strip echo by taking the whole strip echo with the train effect as a research object;
B. finding out all possible motion path combinations of the whole strip-shaped echo between two adjacent moments, and calculating the moving distance of the mass center of the strip-shaped echo and the rotation angle of the long axis under each path combination;
C. because the time delta t (6 minutes) spent on acquiring one volume scan is short, the rotation angle of the whole strip-shaped echo has an upper bound, if the rotation angle of the whole strip-shaped echo is larger than the upper bound in a certain path combination, the path combination is deleted, and only the path combinations with the rotation angles of the whole strip-shaped echo smaller than the upper bound are left;
D. and selecting the shortest path combination as the moving path of the loop band from the remaining path combinations. Since the ratio of the size of the whole band echo to its distance traveled in Δ t (6 minutes) determines that it cannot move far in a short time, and that its previous position, or a position adjacent to it, cannot be occupied by another whole band echo, it is more likely that a combination of paths is left that is shorter, and that is true. Therefore, the path combination with the shortest moving distance of the centroid of the band echo is determined as the real moving path of the whole band echo.
After the position of the whole strip echo is determined, the single echoes can be tracked, and the specific process is as follows:
since the change of the single echo in Δ t (6 minutes) is limited, the two single echoes in the reflectance maps at two adjacent times may be the same single echo as the shapes and sizes of the two single echoes are closer to each other.
And calculating the area difference between each monomer at the current moment and each monomer at the previous moment aiming at the same whole strip echo. If the area difference is less than a threshold, the cells in the previous time instance are stored as a possible match to the current cell. The similarity of the contours of the current monomer to all its possible matching monomers is then calculated by the Hu moments, and the combination with the greatest similarity is considered to be the same monomer, with both labeled with the same ID.
The formula for the Hu moment is as follows:
normal and central moments of order (p + q) of f (x, y):
we process the digitized image after dispersion, and can replace it with the form of integration:
whereinIs the center of gravity.
Normalized central moment ηpq： <math>
<mrow>
<msub>
<msup>
<mi>η</mi>
<mo>′</mo>
</msup>
<mi>pq</mi>
</msub>
<mo>=</mo>
<msub>
<msup>
<mi>μ</mi>
<mo>′</mo>
</msup>
<mi>pq</mi>
</msub>
<mo>/</mo>
<msup>
<mrow>
<mo>(</mo>
<msub>
<msup>
<mi>μ</mi>
<mo>′</mo>
</msup>
<mn>00</mn>
</msub>
<mo>)</mo>
</mrow>
<mi>r</mi>
</msup>
<mo>=</mo>
<msup>
<mi>ρ</mi>
<mrow>
<mi>p</mi>
<mo>+</mo>
<mi>q</mi>
</mrow>
</msup>
<msub>
<mi>μ</mi>
<mi>pq</mi>
</msub>
<mo>/</mo>
<msubsup>
<mi>μ</mi>
<mn>00</mn>
<mi>r</mi>
</msubsup>
<mo>=</mo>
<msup>
<mi>ρ</mi>
<mrow>
<mi>p</mi>
<mo>+</mo>
<mi>q</mi>
</mrow>
</msup>
<msub>
<mi>η</mi>
<mi>pq</mi>
</msub>
</mrow>
</math>
r＝(p+q)/2，p+q＝2，3，4，......
η when the reflectivity map we process undergoes morphological changespqThe following steps are changed:
where ρ is a scaling factor.
Hu constructs 7 geometric moments that satisfy invariant features:
for comparison, we can obtain a degree of similarity with respect to the two monomer profiles:
wherein,
when the profiles of the two monomers are completely the same, I (a, B) ═ 0, the larger the difference between the two profiles, the larger I (a, B);
based on the recognition and judgment result of the train effect for the first time, the area range is defined, the secondary operation time is shortened, and the accuracy is improved. Namely, the suspected area is defined in a certain range, and in the subsequent data entry, the storm body in the suspected area can be directly judged firstly, and then other non-suspected areas are judged. Obtaining the characteristics of the single body and the whole belt shape of the target area again;
the information between the integral belt with the maximum relevance and each monomer in the belt is obtained by comparing various characteristics of the front time sequence monomer and the rear time sequence monomer, comparing integral belt characteristics, comparing the long axis direction of the integral belt unique to the train effect with the moving direction, comparing the speed and the moving direction and the like, so that a one-to-one corresponding relation is formed.
And extrapolating the movement direction, the movement speed and the shape change of the suspected area of the train effect according to the movement inertia of the single body.
Continuously inputting train effect data at three moments, acquiring relevant characteristics, constructing a time sequence of the continuous moments, establishing a monomer family spectrum relationship, checking and comparing the similarity of adjacent secondary monomers, and identifying the motion track of monomer development and extinction through calculation, thereby predicting the motion trend of the monomer, wherein the extrapolation of the train effect is also divided into two parts of extrapolation of integral strip echoes and extrapolation of each monomer echo inside;
firstly, extrapolation of the whole zonal echo is performed, and the body contents are as follows:
A. estimating the moving speed of the whole belt at the next moment according to the inertia of the movement of the monomer and the average speed of the whole belt between the recorded continuous three moments;
B. estimating the moving direction of the whole belt at the next moment according to the inertia of the monomer and the moving direction of the whole belt recorded among three continuous moments;
C. determining the long axis direction of the whole belt at the next moment according to the recorded long axis direction of the whole belt between three continuous moments;
the tracking of the whole strip echo of the train effect is finished by integrating the information.
On the basis of extrapolating the position of the whole strip echo, the position of each single echo can be extrapolated by the same method, but because the change speed of the single echo is relatively fast, the method may have a large error, and therefore the extrapolated position needs to be corrected, and the specific contents are as follows:
A. the principle of extrapolation correction for the monomer is: the speed difference between a single body and the adjacent single body cannot exceed 6m/s, and the difference with the moving direction of the whole strip-shaped echo cannot exceed 10 degrees. If the threshold value is exceeded, the average moving direction and moving speed of the surrounding monomer are used for replacing, so that the distribution situation of the echo of each monomer at the next moment is predicted.
B. In the prediction of the train effect, the single body forming the train effect is generally small, so that the position and the moving speed of the echo of the single body are mainly concerned, and the shape change of the single body does not need to be accurately estimated. Therefore, the patent only roughly extrapolates the shape of the single echo in the "train effect": and expanding and corroding the monomer according to the change rate of the echo area of the monomer.
As shown in fig. 7, the extrapolation of the "train effect" is also divided into two parts, the extrapolation of the whole band echo and the extrapolation of the internal individual echoes, combined, and extrapolated for 6 minutes, 12 minutes, and 18 minutes, respectively. FIGS. 7 a, b, c are real graphs of "train effect" at 6, 12, 18 minutes compared to extrapolated graphs. The a1, b1, c1, a3, b3 and c3 are real reflectivity maps of 6, 12 and 18 minutes 'train effect' and partial enlarged views thereof, and the a2, b2, c2, a4, b4 and c4 are extrapolation maps of 6, 12 and 18 minutes 'train effect' and partial enlarged views thereof.
Claims (7)
1. A method for identifying and forecasting rainstorm caused by train effect is characterized by comprising the following steps:
step S1, preprocessing the real-time radar data: eliminating super refraction, fine monomer and false merging;
step S2, carrying out primary fitting on the monomer subjected to the area primary screening: eliminating interference, clustering the result, performing secondary fitting on the clustering result, and searching a suspected area of the train effect;
step S3, after determining the suspected area of "train effect": automatically identifying the train effect, if so, judging the train effect to be the train effect, performing S4, and if not, skipping to the radar data processing at the next moment, namely, starting from S1;
step S4, tracking the strip echo using the centroid movement distance and the axis rotation angle of the suspected area strip echo: after the position of the whole strip echo is determined, monomer tracking in a suspected area is carried out by using the Hu moment and the form change, and the tracking of a train effect is realized;
and step S5, extrapolating the movement direction, the movement speed and the shape change of the suspected area of the train effect according to the movement inertia of the single body.
2. The method for identifying and forecasting rainstorms caused by "train effect" as claimed in claim 1, wherein the step of preprocessing the real-time radar data in S1 is as follows:
s101: by raising the elevation angle of the radar echo data, super-refraction echoes with strong interference on an automatic identification stage are filtered;
s102: deleting the fine monomers which do not meet the threshold condition by setting an area threshold;
s103: judging whether the reflectivity of the target connected region is greater than 40dBZ, if so, corroding one pixel point, and corroding a strong echo region less; if the detected signal is less than 40dBZ, corroding two pixel points, corroding a weak echo region a little more, then improving a first-level threshold value, judging whether a monomer with a higher first-level threshold value can be detected, if the monomer can be detected, expanding, repeating the process, and if the monomer cannot be detected, stopping.
3. The method for identifying and forecasting rainstorms caused by the "train effect" as claimed in claim 1, wherein said step S2 includes the following steps:
s201: firstly, a centroid method is used for obtaining representative points of each monomer, then straight line fitting is carried out on the points to obtain a fitting straight line of the train effect, then the monomer which deviates from the fitting straight line to a large extent is deleted, and interference of peripheral monomers irrelevant to the train effect is eliminated;
s202: obtaining monomers meeting requirements after preliminary fitting, extracting the mass centers of the monomers, automatically clustering by using a neighbor propagation method, then respectively carrying out straight line fitting on each type of point set again, and deleting the monomers with larger deviation from the fitting straight line degree to obtain a suspected area of the train effect.
4. The method for identifying and forecasting rainstorms caused by the "train effect" as claimed in claim 1, wherein said step S3 includes the following steps:
s301: extracting information of the whole strip-shaped echo in the suspected area, judging whether the length-width ratio of the minimum external rectangle of the whole strip-shaped echo, the average reflectivity and the average speed of the whole strip-shaped echo meet the standard of the train effect, if the length-width ratio of the minimum external rectangle of the whole strip-shaped echo, the average reflectivity and the average speed meet the standard of the train effect, continuing the following judging process, if the length-width ratio does not meet the standard of the train effect, judging that the length-width ratio is not the train effect, then discarding the length-width ratio;
s302: extracting all characteristics of all monomers in the suspected area, putting the characteristics into a rule base, judging whether the monomer echoes meet the 'train effect' monomer echo condition or not by using a standard voting method, counting the proportion of the monomers meeting the 'train effect' monomer echo condition in all the monomers after judging and identifying all the monomers in the suspected area, and judging the monomers to be 'train effect' if the conditions meet threshold conditions.
5. The method for identifying and forecasting rainstorms caused by the "train effect" as claimed in claim 1, wherein said step S4 includes the following steps:
s401: when the train effect is tracked, the attributes of the train effect integral strip echo and the single echo are fused;
s402: calculating the centroid position and the long axis direction of the whole strip echo by taking the whole strip echo with the train effect as a research object, finding out all possible motion path combinations of the whole strip echo between two adjacent moments, calculating the moving distance of the centroid of the strip echo and the rotating angle of the long axis under each path combination, and selecting the shortest path combination as the moving path of the echo band in the path combinations meeting the conditions;
s403: after the position of the whole band echo is determined, calculating the area difference between each monomer at the current moment and each monomer at the previous moment aiming at the same whole band echo, if the area difference is smaller than a threshold value, storing the monomers at the previous moment as possible matching with the current monomer, then calculating the similarity degree of the current monomer and the outlines of all the possible matched monomers through Hu moment, and considering the combination with the maximum similarity degree as the same monomer, and marking the same ID (identity) with the same monomer;
s404: based on the recognition and judgment results of the train effect of the primary pair, the area range is defined, the secondary operation time is shortened, and the accuracy is improved;
s405: the information between the integral belt with the maximum relevance and each monomer in the belt is obtained by comparing various characteristics of the front time sequence monomer and the rear time sequence monomer, comparing integral belt characteristics, comparing the long axis direction of the integral belt unique to the train effect with the moving direction, and comparing the speed with the moving direction, so that a one-to-one correspondence relationship is formed.
6. The method for identifying and forecasting rainstorms caused by the "train effect" as claimed in claim 1, wherein said step S5 includes the following steps:
s501: continuously inputting 'train effect' data at three moments, acquiring relevant characteristics, constructing a time sequence of the continuous moments, establishing a monomer family relation, checking and comparing the similarity of adjacent time monomers, and identifying a motion track of the monomer which develops and disappears through calculation so as to predict the motion trend of the monomer;
s502: firstly, extrapolating the integral band echo, respectively determining the moving speed, the moving direction and the long axis direction of the integral band at the next moment according to the average speed, the moving direction and the long axis direction of the integral band among three recorded continuous moments according to the inertia of the movement of the single body, and integrating the three information to finish the tracking of the integral band echo with the train effect;
s503: on the basis of extrapolating the position of the whole strip echo, the position of each single echo is extrapolated by the same method.
7. The method for identifying and forecasting rainstorms caused by the "train effect" according to claim 6, characterized in that the extrapolation is carried out for 6 minutes, 12 minutes and 18 minutes at the three moments in time.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410446007.6A CN104237890B (en) | 2014-09-03 | 2014-09-03 | The heavy rain identification that one is caused by " train effect " and forecasting procedure |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410446007.6A CN104237890B (en) | 2014-09-03 | 2014-09-03 | The heavy rain identification that one is caused by " train effect " and forecasting procedure |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104237890A true CN104237890A (en) | 2014-12-24 |
| CN104237890B CN104237890B (en) | 2016-06-08 |
Family
ID=52226346
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410446007.6A Expired - Fee Related CN104237890B (en) | 2014-09-03 | 2014-09-03 | The heavy rain identification that one is caused by " train effect " and forecasting procedure |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104237890B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107656256A (en) * | 2017-11-02 | 2018-02-02 | 广州积雨云科技有限公司 | Utilize the method for radar data recognition and tracking heavy rain |
| CN108665486A (en) * | 2018-03-31 | 2018-10-16 | 天津大学 | A kind of labeling method of strong convection monomer sample |
| CN109061650A (en) * | 2018-08-10 | 2018-12-21 | 北京无线电测量研究所 | Networking detection method and system based on array weather radar |
| CN109116358A (en) * | 2018-08-09 | 2019-01-01 | 成都信息工程大学 | Hail identification and occurring area forecast method based on China New Generation Weather Radar |
| CN109164450A (en) * | 2018-09-12 | 2019-01-08 | 天津大学 | A kind of downburst prediction technique based on Doppler Radar Data |
| CN110687618A (en) * | 2019-09-25 | 2020-01-14 | 天津大学 | Automatic nowcasting method for short-time strong rainfall event of multi-monomer convection system |
| CN111507361A (en) * | 2019-01-30 | 2020-08-07 | 富士通株式会社 | Microwave radar-based action recognition device, method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030065445A1 (en) * | 2001-08-30 | 2003-04-03 | White Allen B. | Operational bright-band snow level detection using doppler radar |
| US7231300B1 (en) * | 2004-12-22 | 2007-06-12 | The Weather Channel, Inc. | Producing high-resolution, real-time synthetic meteorological conditions from radar data |
| CN101937078A (en) * | 2009-06-30 | 2011-01-05 | 深圳市气象局 | Nowcasting method and system of thunder cloud cluster based on boundary recognition and tracer technique |
| CN102662172A (en) * | 2012-03-29 | 2012-09-12 | 天津大学 | Stormy cloud cluster extrapolation method based on Doppler radar reflectivity image |
| CN102721987A (en) * | 2012-06-12 | 2012-10-10 | 中国海洋大学 | Method for prewarning Doppler radar remote sensing strong storm |
| CN103529492A (en) * | 2013-09-22 | 2014-01-22 | 天津大学 | Storm body position and form prediction method based on Doppler radar reflectivity image |
-
2014
- 2014-09-03 CN CN201410446007.6A patent/CN104237890B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030065445A1 (en) * | 2001-08-30 | 2003-04-03 | White Allen B. | Operational bright-band snow level detection using doppler radar |
| US7231300B1 (en) * | 2004-12-22 | 2007-06-12 | The Weather Channel, Inc. | Producing high-resolution, real-time synthetic meteorological conditions from radar data |
| CN101937078A (en) * | 2009-06-30 | 2011-01-05 | 深圳市气象局 | Nowcasting method and system of thunder cloud cluster based on boundary recognition and tracer technique |
| CN102662172A (en) * | 2012-03-29 | 2012-09-12 | 天津大学 | Stormy cloud cluster extrapolation method based on Doppler radar reflectivity image |
| CN102721987A (en) * | 2012-06-12 | 2012-10-10 | 中国海洋大学 | Method for prewarning Doppler radar remote sensing strong storm |
| CN103529492A (en) * | 2013-09-22 | 2014-01-22 | 天津大学 | Storm body position and form prediction method based on Doppler radar reflectivity image |
Non-Patent Citations (4)
| Title |
|---|
| J.T.JONHSON: "The Storm Cell Identification and Tracking Algorithm:An Enhanced WSR-88D Algorithm", 《WEATHER AND FORECASTING》 * |
| MICHAEL DIXON,GERRY WIENER: "TITAN:Thunderstorm Identification,Tracking,Analysis,and Nowcasting-A Radar-based Methodology", 《JOURNAL OF ATMOSPHERIC AND OCEANIC TECHNOLOGY》 * |
| 俞小鼎: "雷暴与强对流临近天气预报技术进展", 《气象学报》 * |
| 韩雷等: "《腐蚀方法在风暴自动识别中的应用》", 《北京大学学报》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107656256A (en) * | 2017-11-02 | 2018-02-02 | 广州积雨云科技有限公司 | Utilize the method for radar data recognition and tracking heavy rain |
| CN107656256B (en) * | 2017-11-02 | 2019-06-28 | 广州积雨云科技有限公司 | Utilize the method for radar data recognition and tracking heavy rain |
| CN108665486A (en) * | 2018-03-31 | 2018-10-16 | 天津大学 | A kind of labeling method of strong convection monomer sample |
| CN109116358A (en) * | 2018-08-09 | 2019-01-01 | 成都信息工程大学 | Hail identification and occurring area forecast method based on China New Generation Weather Radar |
| CN109116358B (en) * | 2018-08-09 | 2022-11-22 | 成都信息工程大学 | Hail identification and area forecast method based on new-generation weather radar |
| CN109061650A (en) * | 2018-08-10 | 2018-12-21 | 北京无线电测量研究所 | Networking detection method and system based on array weather radar |
| CN109164450A (en) * | 2018-09-12 | 2019-01-08 | 天津大学 | A kind of downburst prediction technique based on Doppler Radar Data |
| CN109164450B (en) * | 2018-09-12 | 2022-11-22 | 天津大学 | Downburst prediction method based on Doppler radar data |
| CN111507361A (en) * | 2019-01-30 | 2020-08-07 | 富士通株式会社 | Microwave radar-based action recognition device, method and system |
| CN111507361B (en) * | 2019-01-30 | 2023-11-21 | 富士通株式会社 | Action recognition device, method and system based on microwave radar |
| CN110687618A (en) * | 2019-09-25 | 2020-01-14 | 天津大学 | Automatic nowcasting method for short-time strong rainfall event of multi-monomer convection system |
| CN110687618B (en) * | 2019-09-25 | 2021-10-01 | 天津大学 | Automatic nowcasting method for short-time strong rainfall event of multi-monomer convection system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104237890B (en) | 2016-06-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104237890B (en) | The heavy rain identification that one is caused by " train effect " and forecasting procedure | |
| CN106526708B (en) | A kind of intelligent early-warning analysis method of the meteorological strong convective weather based on machine learning | |
| CN104036239B (en) | Fast high-resolution SAR (synthetic aperture radar) image ship detection method based on feature fusion and clustering | |
| CN110223302A (en) | A kind of naval vessel multi-target detection method extracted based on rotary area | |
| KR101258668B1 (en) | Korea local radar processing system | |
| CN108171193B (en) | Polarized SAR (synthetic aperture radar) ship target detection method based on super-pixel local information measurement | |
| Pan et al. | Object-based and supervised detection of potholes and cracks from the pavement images acquired by UAV | |
| CN113345237A (en) | Lane-changing identification and prediction method, system, equipment and storage medium for extracting vehicle track by using roadside laser radar data | |
| CN103353988B (en) | Allos SAR scene Feature Correspondence Algorithm performance estimating method | |
| CN110942111B (en) | Method and device for identifying strong convection cloud cluster | |
| CN103559508B (en) | A kind of based on continuous Adaboost video vehicle detection method | |
| CN109917394B (en) | Short-term intelligent extrapolation method based on weather radar | |
| CN109164450B (en) | Downburst prediction method based on Doppler radar data | |
| KR101255736B1 (en) | Method for classifying meteorological/non-meteorological echoes using single polarization radars | |
| CN113255580A (en) | Method and device for identifying sprinkled objects and vehicle sprinkling and leaking | |
| WO2018168165A1 (en) | Weather forecasting device, weather forecasting method, and program | |
| CN111352113A (en) | Strong convection weather short-term forecasting method and system, storage medium and terminal | |
| CN110261857A (en) | A kind of weather radar spatial interpolation methods | |
| CN115761534A (en) | Method for detecting and tracking small target of infrared unmanned aerial vehicle under air background | |
| CN114019503A (en) | FOD detection system-based airport runway foreign matter detection method, device and storage medium | |
| CN110687618B (en) | Automatic nowcasting method for short-time strong rainfall event of multi-monomer convection system | |
| CN116680604A (en) | Strong convection characteristic classification recognition method, electronic equipment and medium | |
| CN108734122A (en) | A kind of EO-1 hyperion city water body detection method based on adaptive samples selection | |
| Liu et al. | Classified early warning and forecast of severe convective weather based on LightGBM algorithm | |
| CN111736156B (en) | Upwind area identification method and device based on weather radar |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160608 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |