CN107341449A - A kind of GMS Calculation of precipitation method based on cloud mass changing features - Google Patents

A kind of GMS Calculation of precipitation method based on cloud mass changing features Download PDF

Info

Publication number
CN107341449A
CN107341449A CN201710448042.5A CN201710448042A CN107341449A CN 107341449 A CN107341449 A CN 107341449A CN 201710448042 A CN201710448042 A CN 201710448042A CN 107341449 A CN107341449 A CN 107341449A
Authority
CN
China
Prior art keywords
mrow
msub
mfrac
munderover
mover
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.)
Pending
Application number
CN201710448042.5A
Other languages
Chinese (zh)
Inventor
刘玉
卢涵宇
袁咏仪
张旭
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.)
Liupanshui Sanlida Technology Co Ltd
Guizhou Beidou Technology Co Ltd
Original Assignee
Liupanshui Sanlida Technology Co Ltd
Guizhou Beidou Technology Co Ltd
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 Liupanshui Sanlida Technology Co Ltd, Guizhou Beidou Technology Co Ltd filed Critical Liupanshui Sanlida Technology Co Ltd
Priority to CN201710448042.5A priority Critical patent/CN107341449A/en
Publication of CN107341449A publication Critical patent/CN107341449A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/13—Satellite images
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01W—METEOROLOGY
    • G01W1/00—Meteorology
    • G01W1/10—Devices for predicting weather conditions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
    • G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour
    • 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
    • Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
    • Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
    • 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

Abstract

The invention belongs to Calculation of precipitation technical field, discloses a kind of GMS Calculation of precipitation method based on cloud mass changing features, including：Satellite Simulation of Precipitation parameter attribute collection is obtained to portray the generation of cloud precipitation and evolution, different dimension Cloud-Picture Characteristics parameters are normalized from most value method for normalizing；Structure based on before three layers to the satellite Calculation of precipitation model of type reverse transmittance nerve network, for the region Calculation of precipitation, and using the Simulation of Precipitation precision of multi objective system anlysis model.The rainfall distribution that the present invention is derived using Utilizing Satellite Remote Sensing Data can compensate for the deficiency of conventional meteorological observation, there is provided the precipitation information of more horn of plenty；Beneficial to Nowcasting is carried out, be advantageous to monitor flood, be advantageous to make geological disaster and give warning in advance, have great importance to improving weather forecast accuracy rate and preventing and reducing natural disasters.

Description

A kind of GMS Calculation of precipitation method based on cloud mass changing features
Technical field
The invention belongs to Calculation of precipitation technical field, more particularly to a kind of GMS based on cloud mass changing features Calculation of precipitation method.
Background technology
With the development of meteorological satellite, the distribution of applied satellite data estimation precipitation and intensity, as Satellite data application An importance also grow up simultaneously.Ground observation and aerological sounding website are sparse, are influenceed by earth's surface heating power difference, The local of precipitation differs greatly.
Image segmentation is a classical image processing problem, some significant the purpose is to which piece image is divided into Region, wherein each region has similar a feature and meaning, but the complexity and the diversity of target dealt with objects due to it, The always difficult point and hot issue of computer vision field.Particularly further go deep into recently as technological revolution, image It is segmented in increasing field such as industrial production, Video Applications, Medical Image Processing, biometric image processing, intelligent transportation, electricity Sub- commercial affairs, E-Government, man-machine interface, virtual reality etc. are widely used.Rapid development and net with internet The increase of network type and bandwidth, the decline of imaging device cost, caused image rapidly increase, the product based on image procossing with Our daily life is also increasingly closely related.Image is split as a key technology in image procossing, therefore to figure As the research of cutting techniques is not only with important theory value but also with good practical value
Image segmentation research starts from nineteen seventies, by the development of nearly half a century, it is proposed that a large amount of classical Image segmentation algorithm.Traditional classic map picture segmentation mainly has Threshold segmentation, the segmentation based on edge, point based on region Cut, the segmentation based on global optimization criterion and the segmentation based on statistics etc..Many algorithms all obtain in specific application scenarios Good segmentation effect, but due to the diversification of image imaging mode and equipment, handle object diversity and body form not The reasons such as systematicness cause a kind of no image segmentation algorithm to can be good at handling all types of images.
Image is segmented in also referred to as cluster analysis in statistics.Finite mixture model (Finite based on statistics in recent years Mixture Mode, FMM) research it is very active always.But a distinct disadvantage of the FMM based on Gaussian Profile is that it is right Noise is very sensitive.And Xue Shengshi distributions have heavier afterbody compared with Gaussian Profile, therefore it has well to noise Robustness, it is the healthy and strong replacement of Gaussian Profile.
In summary, the problem of prior art is present be：Ground observation and aerological sounding website are sparse at present, by Ground Heat The influence of power difference, the local of precipitation differ greatly；It is larger deviation to be also present in the data of Calculation of precipitation；Conventional images segmentation side It in method, can not efficiently be split in the presence of noise, reduce the complexity for solving parametric procedure；And it can not improve The robustness of image segmentation, improve the quality of image segmentation.
The content of the invention
The problem of existing for prior art, the invention provides a kind of GMS based on cloud mass changing features Calculation of precipitation method.
The present invention is achieved in that a kind of GMS Calculation of precipitation system based on cloud mass changing features, institute Stating the GMS Calculation of precipitation system based on cloud mass changing features includes：
Normalized module, for obtaining satellite Simulation of Precipitation parameter attribute collection to portray the generation of cloud precipitation and development Process, different dimension Cloud-Picture Characteristics parameters are normalized from most value method for normalizing；
The normalized module utilizes this attribute of shadow region, by being carried out as follows to colored RGB images Normalized：
Wherein：R, G, B are respectively original RGB component, and R ', G ', B ' are respectively the RGB component after normalizing；At B ' points In amount, what shadow region mainly occupied is high pixel value end, the method by using Threshold segmentation to B ' component maps, sets one Higher threshold value just obtains shadow region substantially；
The method of the Threshold segmentation includes：
1) image to be split is inputted, obtains the colouring information of image；Assuming that N number of pixel, these pictures are shared in a sub-picture Element is divided into K class；
2) clusters number K and the likelihood function changing value of iteration ends and the maximum times of iteration are set, calculate pixel Maximum a posteriori probability, the classification of pixel is obtained according to maximum a posteriori probability principle；Formula below is pressed in the solution of the average value of pixel Solved,In formulaRepresent adjacent-systems, NnRepresent the number of neighbour in adjacent-systems；
3) initiation parameter, mean μ and covariance Σ are obtained using K- mean algorithms, then initializing variable, sets and become Measure η=1, precision Λ=(η Σ)-1, v=1,The average of each pixel in pixel is tried to achieve using neighbor relationships Specifically include：
Calculate parameter Λ=(the η Σ) of student t distributions-1；
Gaussian Profile and the multiplication relationship of gamma distribution can be decomposed into according to student t distributions
St (x | μ, Λ, v)=N (x | μ, (η Λ)-1Gam(η|v/2,v/2))
Wherein gamma distribution definition be
Using parameter η, Λ, μ, the v tried to achieve, the value that student t is distributed is tried to achieve；
4) current μ is utilizedkAnd ΣkGaussian Profile is calculated, calculates student t distributions；Calculate scene mixed coefficint πnkWith it is rear Test probability znk；According to the gauss of distribution function and student's t distribution functions tried to achieve, scene mixed stocker is tried to achieve according to following two formula Number πnkWith posterior probability znk
Obtain carrying out image by the image procossing submodule built in normalized module behind shadow region substantially Processing；The processing of described image includes：Image procossing submodule is filtered place to gray level image using high pass/low pass filter Reason is to construct the reference picture of image to be evaluated, using 3*3 mean filters, using each pixel of Filtering Template traversing graph picture, Template center is placed in current pixel every time, the average value of all pixels is newly worth as current pixel using in template, and template is
Calculate before and after image filtering each edge half-tone information respectively, the image F statistical informations to be evaluated before filtering process For sum_orig, the reference picture F2 statistical informations after filtering process are sum_filter, and specific formula for calculation is as follows：
Wherein, w1 and w2 is according to from the weights set with a distance from center pixel, w1=1, w2=1/3；
Using the ratio of the image filtering front and rear edges grey-level statistics drawn as fuzziness index, for convenience of evaluating, Take larger for denominator, less is molecule, keeps the value between (0,1)；
A fuzziness indication range [min, max] according to corresponding to being drawn the DMOS scopes of the best visual effect；Draw Final image be shown in on the intelligent terminal screen of normalized module wireless connection；
Calculation of precipitation module, build based on to the satellite Calculation of precipitation model of type reverse transmittance nerve network, being used before three layers In the region Calculation of precipitation, and using the Simulation of Precipitation precision of multi objective system anlysis model；
The multi objective system anlysis model calculation formula
In formula：--- the average value of meteorological variables in observation period；
yi--- the sampled instantaneous value of i-th of meteorological variables, i.e. sample in observation period, wherein, it is mistake, suspicious anon-normal True sample should be discarded without in calculating, even yi=0；
N --- the total sample number in observation period, determined by sample frequency peace mean time section；
M --- correct sample number (m≤N) in observation period.
Further, the method for the Threshold segmentation also includes：
Calculate mean μkWith covariance Σk；Change ηkValue and try to achieve accuracy value
Further, the method for the Threshold segmentation also includes：The value of log-likelihood function is calculated, calculates its changing value
Or iterations just exits circulate operation more than stated number, initiation parameter step is otherwise performed；
The maximum a posteriori probability of pixel is calculated, the classification of pixel is obtained according to maximum a posteriori probability principle.
Further, in the value for calculating log-likelihood function, according to the log-likelihood function of whole image
The complexity of parameter is solved, introduces implicit variable znkSo that log-likelihood function is written as again
The internal relation that can be made up of according to student t distributions Gaussian Profile and gamma distribution product, log-likelihood letter Number is re-written as following depicted
Further, L (Θ) is found a function on parameter ηkDerivative be
Wherein D represents the dimension of data, image for it is colored when D=3, D=1 during gray scale；SetObtain
Try to achieve ηkValue be
L (Θ) is found a function on parameter ΛkDerivative be
SetObtain equation
Try to achieve ΛkValue be
Further, L (Θ) is found a function on parameter μkDerivative be
SetObtain equation,
Try to achieve μkValue be
Further, the cycle-index for certain limit or iteration being reached when the rate of change of log-likelihood function value reaches certain After input, according to maximum posteriori criterion
Obtain the mark of pixel.
Another object of the present invention is to provide a kind of GMS Calculation of precipitation side based on cloud mass changing features Method, comprise the following steps：
Step 1, satellite Simulation of Precipitation parameter attribute collection is obtained to portray the generation of cloud precipitation and evolution, from most Different dimension Cloud-Picture Characteristics parameters are normalized value method for normalizing；
Step 2, build based on before three layers to the satellite Calculation of precipitation model of type reverse transmittance nerve network, for the ground Domain Calculation of precipitation, and using the Simulation of Precipitation precision of multi objective system anlysis model.
Advantages of the present invention and good effect are：The rainfall distribution derived using Utilizing Satellite Remote Sensing Data can be big The deficiency of conventional meteorological observation is made up greatly, there is provided the precipitation information of more horn of plenty；Beneficial to Nowcasting is carried out, be advantageous to supervise Flood is surveyed, is advantageous to make geological disaster and gives warning in advance, to improving weather forecast accuracy rate and preventing and reducing natural disasters with weight The meaning wanted.To between the estimation result of type reverse transmittance nerve network satellite Calculation of precipitation model and rainfall gauge measured value before layer Correlation can reach 0.57；Model estimation result is systematic to be underestimated less than normal, and imply that will have to the weak precipitation intensity in the region It is preferably indicative.
A distinct disadvantage of FMM based on Gaussian Profile is that it does not account for the spatial relationship of pixel.Therefore to noise Noise immunity it is not strong.The present invention has effectively drawn the direct spatial relationship of pixel, has more preferable robustness；
The present invention is had stronger noise immunity, can obtain preferably segmentation effect；
What scene hybrid parameter proposed by the present invention was shown is expressed as probability vector, and this avoid in most number space It, which is solved, in mixed model needs the extra efficiency remedied calculating, improve algorithm；
There is inherent incidence relation in the present invention, with Gaussian Profile using the model of student t distributions so as to which student t be distributed Gaussian Profile is converted to, on the one hand reduces the number of parameters of solution, on the other hand simplifies solution procedure so that the present invention is easily In realization.
The present invention combines the advantage of spatial variations mixed model and the robustness of student t- distributions, it is proposed that one kind is based on learning The spatial variations mixed model of Sheng Shi distributions.In the present invention, the weight function of definition is used for representing the spatial relationship between pixel, The Gaussian Profile of the spatial relationship and pixel is closely related.Scene mixed coefficint is explicit to be expressed as probability vector.They Automatically this restrictive condition of probability vector is met.Therefore the step for remedying calculating is eliminated in reasoning process, so as to simplify Solution procedure, and then improve the high efficiency of algorithm.The present invention is based on log-likelihood function is maximized, in view of student t is distributed letter Several complicated representative, according to student t distributions and the internal relation of Gaussian Profile, the solution of student's t distributed constants is converted to The solution of Gaussian Distribution Parameters.This mode not only simplifies the number for solving parameter, and also reduces parametric solution process Complexity so that whole Algorithm for Solving process is relatively simple.
The present invention has taken into full account the spatial relationship between pixel, while represents the scene mixed coefficint of pixel space relation Shown is expressed as probability vector.Number of parameters needed for the present invention is much smaller than other models based on markov random file Parameter, therefore it is easily achieved.In addition, carrying out the solution of parameter using expectation-maximization algorithm, the optimization of parameter is obtained Value.
Image blur evaluation method provided by the invention, established different from traditional evaluation method in image to be evaluated certainly On the basis of body structure feature, from the angle of relative evaluation, the reference picture of image to be evaluated is constructed using wave filter, is calculated The ratio of image border statistical information is as evaluation index before and after change.The principle of the present invention is simple, realizes image blur The content independence and real-time of evaluation, fuzziness that can quick and precisely between any image of evaluation comparison.
Brief description of the drawings
Fig. 1 is the GMS Calculation of precipitation method flow provided in an embodiment of the present invention based on cloud mass changing features Figure.
Fig. 2 is the GMS Calculation of precipitation system signal provided in an embodiment of the present invention based on cloud mass changing features Figure.
In figure：1st, normalized module；2nd, Calculation of precipitation module.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to embodiments, to the present invention It is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not used to Limit the present invention.
The application principle of the present invention is explained in detail below in conjunction with the accompanying drawings.
As shown in figure 1, the GMS Calculation of precipitation side provided in an embodiment of the present invention based on cloud mass changing features Method comprises the following steps：
S101：Satellite Simulation of Precipitation parameter attribute collection is obtained to portray the generation of cloud precipitation and evolution, from being most worth Different dimension Cloud-Picture Characteristics parameters are normalized method for normalizing；
S102：Structure based on before three layers to the satellite Calculation of precipitation model of type reverse transmittance nerve network, for the region Calculation of precipitation, and using the Simulation of Precipitation precision of multi objective system anlysis model.
As shown in Fig. 2 the GMS Calculation of precipitation system provided in an embodiment of the present invention based on cloud mass changing features System, the GMS Calculation of precipitation system based on cloud mass changing features include：
Normalized module 1, for obtaining satellite Simulation of Precipitation parameter attribute collection to portray the generation of cloud precipitation and hair Exhibition process, different dimension Cloud-Picture Characteristics parameters are normalized from most value method for normalizing；
The normalized module utilizes this attribute of shadow region, by being carried out as follows to colored RGB images Normalized：
Wherein：R, G, B are respectively original RGB component, and R ', G ', B ' are respectively the RGB component after normalizing；At B ' points In amount, what shadow region mainly occupied is high pixel value end, the method by using Threshold segmentation to B ' component maps, sets one Higher threshold value just obtains shadow region substantially；
The method of the Threshold segmentation includes：
1) image to be split is inputted, obtains the colouring information of image；Assuming that N number of pixel, these pictures are shared in a sub-picture Element is divided into K class；
2) clusters number K and the likelihood function changing value of iteration ends and the maximum times of iteration are set, calculate pixel Maximum a posteriori probability, the classification of pixel is obtained according to maximum a posteriori probability principle；Formula below is pressed in the solution of the average value of pixel Solved,In formulaRepresent adjacent-systems, NnRepresent the number of neighbour in adjacent-systems；
3) initiation parameter, mean μ and covariance Σ are obtained using K- mean algorithms, then initializing variable, sets and become Measure η=1, precision Λ=(η Σ)-1, v=1,The average of each pixel in pixel is tried to achieve using neighbor relationships Specifically include：
Calculate parameter Λ=(the η Σ) of student t distributions-1；
Gaussian Profile and the multiplication relationship of gamma distribution can be decomposed into according to student t distributions
St (x | μ, Λ, v)=N (x | μ, (η Λ)-1Gam(η|v/2,v/2))
Wherein gamma distribution definition be
Using parameter η, Λ, μ, the v tried to achieve, the value that student t is distributed is tried to achieve；
4) current μ is utilizedkAnd ΣkGaussian Profile is calculated, calculates student t distributions；Calculate scene mixed coefficint πnkWith it is rear Test probability znk；According to the gauss of distribution function and student's t distribution functions tried to achieve, scene mixed stocker is tried to achieve according to following two formula Number πnkWith posterior probability znk
Obtain carrying out image by the image procossing submodule built in normalized module behind shadow region substantially Processing；The processing of described image includes：Image procossing submodule is filtered place to gray level image using high pass/low pass filter Reason is to construct the reference picture of image to be evaluated, using 3*3 mean filters, using each pixel of Filtering Template traversing graph picture, Template center is placed in current pixel every time, the average value of all pixels is newly worth as current pixel using in template, and template is
Calculate before and after image filtering each edge half-tone information respectively, the image F statistical informations to be evaluated before filtering process For sum_orig, the reference picture F2 statistical informations after filtering process are sum_filter, and specific formula for calculation is as follows：
Wherein, w1 and w2 is according to from the weights set with a distance from center pixel, w1=1, w2=1/3；
Using the ratio of the image filtering front and rear edges grey-level statistics drawn as fuzziness index, for convenience of evaluating, Take larger for denominator, less is molecule, keeps the value between (0,1)；
A fuzziness indication range [min, max] according to corresponding to being drawn the DMOS scopes of the best visual effect；Draw Final image be shown in on the intelligent terminal screen of normalized module wireless connection；
Calculation of precipitation module 2, build based on to the satellite Calculation of precipitation model of type reverse transmittance nerve network, being used before three layers In the region Calculation of precipitation, and using the Simulation of Precipitation precision of multi objective system anlysis model；
The multi objective system anlysis model calculation formula
In formula：--- the average value of meteorological variables in observation period；
yi--- the sampled instantaneous value of i-th of meteorological variables, i.e. sample in observation period, wherein, it is mistake, suspicious anon-normal True sample should be discarded without in calculating, even yi=0；
N --- the total sample number in observation period, determined by sample frequency peace mean time section；
M --- correct sample number (m≤N) in observation period.
Further, the method for the Threshold segmentation also includes：
Calculate mean μkWith covariance Σk；Change ηkValue and try to achieve accuracy value
The method of the Threshold segmentation also includes：The value of log-likelihood function is calculated, calculates its changing value
Or iterations just exits circulate operation more than stated number, initiation parameter step is otherwise performed；
The maximum a posteriori probability of pixel is calculated, the classification of pixel is obtained according to maximum a posteriori probability principle.
In the value for calculating log-likelihood function, according to the log-likelihood function of whole image
The complexity of parameter is solved, introduces implicit variable znkSo that log-likelihood function is written as again
The internal relation that can be made up of according to student t distributions Gaussian Profile and gamma distribution product, log-likelihood letter Number is re-written as following depicted
L (Θ) is found a function on parameter ηkDerivative be
Wherein D represents the dimension of data, image for it is colored when D=3, D=1 during gray scale；SetObtain
Try to achieve ηkValue be
L (Θ) is found a function on parameter ΛkDerivative be
SetObtain equation
Try to achieve ΛkValue be
L (Θ) is found a function on parameter μkDerivative be
SetObtain equation,
Try to achieve μkValue be
After the cycle-index that the rate of change of log-likelihood function value reaches certain limit or iteration reaches certain input, According to maximum posteriori criterion
Obtain the mark of pixel.
The method GMS infrared band of the present invention can relatively accurately disclose the precipitation mechanism of cloud, higher time point Resolution remote sensing images can monitor the change details of cloud atlas, and obtain the Simulation of Precipitation parameter that can reflect cloud atlas Characteristics of Precipitation； Artificial neural network can preferably portray the non-linear rule of the region satellite Characteristics of Precipitation；To type Back propagation neural before three layers The estimation result of network satellite Calculation of precipitation model can reach 0.57 with the correlation between rainfall gauge measured value.Model estimation knot Fruit is systematic underestimate it is less than normal, imply that will have to the weak precipitation intensity in the region it is preferably indicative.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention All any modification, equivalent and improvement made within refreshing and principle etc., should be included in the scope of the protection.

Claims (8)

1. a kind of GMS Calculation of precipitation system based on cloud mass changing features, it is characterised in that described to be based on cloud mass The GMS Calculation of precipitation system of changing features includes：
Normalized module, the generation of cloud precipitation and developed with portraying for obtaining satellite Simulation of Precipitation parameter attribute collection Journey, different dimension Cloud-Picture Characteristics parameters are normalized from most value method for normalizing；
The normalized module utilizes this attribute of shadow region, by carrying out following normalizing to colored RGB images Change is handled：
<mrow> <msup> <mi>R</mi> <mo>&amp;prime;</mo> </msup> <mo>=</mo> <mfrac> <mi>R</mi> <mrow> <mi>R</mi> <mo>+</mo> <mi>G</mi> <mo>+</mo> <mi>B</mi> </mrow> </mfrac> </mrow>
<mrow> <msup> <mi>G</mi> <mo>&amp;prime;</mo> </msup> <mo>=</mo> <mfrac> <mi>G</mi> <mrow> <mi>R</mi> <mo>+</mo> <mi>G</mi> <mo>+</mo> <mi>B</mi> </mrow> </mfrac> </mrow>
<mrow> <msup> <mi>B</mi> <mo>&amp;prime;</mo> </msup> <mo>=</mo> <mfrac> <mi>B</mi> <mrow> <mi>R</mi> <mo>+</mo> <mi>G</mi> <mo>+</mo> <mi>B</mi> </mrow> </mfrac> <mo>;</mo> </mrow>
Wherein：R, G, B are respectively original RGB component, and R ', G ', B ' are respectively the RGB component after normalizing；In B ' components, What shadow region mainly occupied is high pixel value end, the method by using Threshold segmentation to B ' component maps, is set one higher Threshold value just obtain shadow region substantially；
The method of the Threshold segmentation includes：
1) image to be split is inputted, obtains the colouring information of image；Assuming that N number of pixel, these pixel quilts are shared in a sub-picture It is divided into K class；
2) clusters number K and the likelihood function changing value of iteration ends and the maximum times of iteration are set, calculate the maximum of pixel Posterior probability, the classification of pixel is obtained according to maximum a posteriori probability principle；The solution of the average value of pixel is carried out by formula below Solve,θ in formulanRepresent adjacent-systems, NnRepresent the number of neighbour in adjacent-systems；
3) initiation parameter, mean μ and covariance Σ are obtained using K- mean algorithms, then initializing variable, setting variable η= 1, precision Λ=(η Σ)-1, v=1,The average of each pixel in pixel is tried to achieve using neighbor relationshipsSpecific bag Include：
Calculate parameter Λ=(the η Σ) of student t distributions-1；
Gaussian Profile and the multiplication relationship of gamma distribution can be decomposed into according to student t distributions
St (x | μ, Λ, v)=N (x | μ, (η Λ)-1Gam(η|v/2,v/2))
Wherein gamma distribution definition be
<mrow> <mi>G</mi> <mi>a</mi> <mi>m</mi> <mrow> <mo>(</mo> <mrow> <mi>&amp;eta;</mi> <mo>|</mo> <mi>v</mi> <mo>/</mo> <mn>2</mn> <mo>,</mo> <mi>v</mi> <mo>/</mo> <mn>2</mn> </mrow> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mn>1</mn> <mrow> <mi>&amp;Gamma;</mi> <mrow> <mo>(</mo> <mfrac> <msub> <mi>v</mi> <mi>k</mi> </msub> <mn>2</mn> </mfrac> <mo>)</mo> </mrow> </mrow> </mfrac> <msup> <mrow> <mo>(</mo> <mfrac> <msub> <mi>v</mi> <mi>k</mi> </msub> <mn>2</mn> </mfrac> <mo>)</mo> </mrow> <mrow> <msub> <mi>v</mi> <mi>k</mi> </msub> <mo>/</mo> <mn>2</mn> <mo>-</mo> <mn>1</mn> </mrow> </msup> <mi>exp</mi> <mrow> <mo>(</mo> <mrow> <mo>-</mo> <mfrac> <msub> <mi>v</mi> <mi>k</mi> </msub> <mn>2</mn> </mfrac> <msub> <mi>&amp;eta;</mi> <mi>k</mi> </msub> </mrow> <mo>)</mo> </mrow> </mrow>
Using parameter η, Λ, μ, the v tried to achieve, the value that student t is distributed is tried to achieve；
4) current μ is utilizedkAnd ΣkGaussian Profile is calculated, calculates student t distributions；Calculate scene mixed coefficint πnkAnd posterior probability znk；According to the gauss of distribution function and student's t distribution functions tried to achieve, scene mixed coefficint π is tried to achieve according to following two formulank With posterior probability znk
<mrow> <msub> <mi>&amp;pi;</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mo>=</mo> <mfrac> <mrow> <msub> <mi>&amp;delta;</mi> <mi>k</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>&amp;delta;</mi> <mi>j</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>=</mo> <mfrac> <mrow> <mi>exp</mi> <mrow> <mo>(</mo> <munder> <mi>&amp;Sigma;</mi> <mrow> <mi>m</mi> <mo>&amp;Element;</mo> <msub> <mo>&amp;part;</mo> <mi>n</mi> </msub> </mrow> </munder> <mi>N</mi> <mo>(</mo> <mrow> <msub> <mi>x</mi> <mi>m</mi> </msub> <mo>|</mo> <msub> <mi>&amp;Theta;</mi> <mi>k</mi> </msub> </mrow> <mo>)</mo> <mo>)</mo> </mrow> </mrow> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <mi>exp</mi> <mrow> <mo>(</mo> <munder> <mi>&amp;Sigma;</mi> <mrow> <mi>m</mi> <mo>&amp;Element;</mo> <msub> <mo>&amp;part;</mo> <mi>n</mi> </msub> </mrow> </munder> <mi>N</mi> <mo>(</mo> <mrow> <msub> <mi>x</mi> <mi>m</mi> </msub> <mo>|</mo> <msub> <mi>&amp;Theta;</mi> <mi>j</mi> </msub> </mrow> <mo>)</mo> <mo>)</mo> </mrow> </mrow> </mfrac> </mrow>
<mrow> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mo>=</mo> <mfrac> <mrow> <msub> <mi>&amp;pi;</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mi>S</mi> <mi>t</mi> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>|</mo> <msub> <mi>&amp;theta;</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>j</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>&amp;pi;</mi> <mrow> <mi>n</mi> <mi>j</mi> </mrow> </msub> <mi>S</mi> <mi>t</mi> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>|</mo> <msub> <mi>&amp;theta;</mi> <mi>j</mi> </msub> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>;</mo> </mrow>
Obtain carrying out behind shadow region substantially the processing of image by the image procossing submodule built in normalized module； The processing of described image includes：Image procossing submodule using high pass/low pass filter to gray level image be filtered processing with The reference picture of image to be evaluated is constructed, using 3*3 mean filters, using each pixel of Filtering Template traversing graph picture, every time Template center is placed in current pixel, the average value of all pixels is newly worth as current pixel using in template, and template is
Calculate before and after image filtering each edge half-tone information respectively, the image F statistical informations to be evaluated before filtering process are Sum_orig, the reference picture F2 statistical informations after filtering process are sum_filter, and specific formula for calculation is as follows：
<mrow> <mtable> <mtr> <mtd> <mrow> <mi>s</mi> <mi>u</mi> <mi>m</mi> <mo>_</mo> <mi>o</mi> <mi>r</mi> <mi>i</mi> <mi>g</mi> <mo>=</mo> <mi>w</mi> <mn>1</mn> <mo>&amp;times;</mo> <mrow> <mo>(</mo> <mrow> <mo>|</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>|</mo> <mo>+</mo> <mo>|</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>-</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <mo>|</mo> <mo>+</mo> <mo>|</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>+</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <mo>|</mo> <mo>+</mo> <mo>|</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>+</mo> <mn>1</mn> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>|</mo> </mrow> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>+</mo> <mi>w</mi> <mn>2</mn> <mo>&amp;times;</mo> <mrow> <mo>(</mo> <mrow> <mo>|</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mi>j</mi> <mo>-</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <mo>|</mo> <mo>+</mo> <mo>|</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mi>j</mi> <mo>+</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <mo>|</mo> <mo>+</mo> <mo>|</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>+</mo> <mn>1</mn> <mo>,</mo> <mi>j</mi> <mo>-</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <mo>|</mo> <mo>+</mo> <mo>|</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>+</mo> <mn>1</mn> <mo>,</mo> <mi>j</mi> <mo>+</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <mo>|</mo> </mrow> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> </mtable> <mo>,</mo> </mrow>
<mrow> <mtable> <mtr> <mtd> <mrow> <mi>s</mi> <mi>u</mi> <mi>m</mi> <mo>_</mo> <mi>f</mi> <mi>i</mi> <mi>l</mi> <mi>t</mi> <mi>e</mi> <mi>r</mi> <mo>=</mo> <mi>w</mi> <mn>1</mn> <mo>&amp;times;</mo> <mrow> <mo>(</mo> <mrow> <mo>|</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>|</mo> <mo>+</mo> <mo>|</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>-</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <mo>|</mo> <mo>+</mo> <mo>|</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> <mo>+</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <mo>|</mo> <mo>+</mo> <mo>|</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>+</mo> <mn>1</mn> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>|</mo> </mrow> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>+</mo> <mi>w</mi> <mn>2</mn> <mo>&amp;times;</mo> <mrow> <mo>(</mo> <mrow> <mo>|</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mi>j</mi> <mo>-</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <mo>|</mo> <mo>+</mo> <mo>|</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>-</mo> <mn>1</mn> <mo>,</mo> <mi>j</mi> <mo>+</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <mo>|</mo> <mo>+</mo> <mo>|</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>+</mo> <mn>1</mn> <mo>,</mo> <mi>j</mi> <mo>-</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <mo>|</mo> <mo>+</mo> <mo>|</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>,</mo> <mi>j</mi> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mi>F</mi> <mn>2</mn> <mrow> <mo>(</mo> <mrow> <mi>i</mi> <mo>+</mo> <mn>1</mn> <mo>,</mo> <mi>j</mi> <mo>+</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <mo>|</mo> </mrow> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> </mtable> <mo>,</mo> </mrow>
Wherein, w1 and w2 is according to from the weights set with a distance from center pixel, w1=1, w2=1/3；
Using the ratio of the image filtering front and rear edges grey-level statistics drawn as fuzziness index, for convenience of evaluating, take compared with It is big for denominator, less is molecule, keeps the value between (0,1)；
A fuzziness indication range [min, max] according to corresponding to being drawn the DMOS scopes of the best visual effect；Draw most Whole image be shown in on the intelligent terminal screen of normalized module wireless connection；
Calculation of precipitation module, build based on before three layers to the satellite Calculation of precipitation model of type reverse transmittance nerve network, for this Region Calculation of precipitation, and using the Simulation of Precipitation precision of multi objective system anlysis model；
The multi objective system anlysis model calculation formula
<mrow> <mover> <mi>Y</mi> <mo>&amp;OverBar;</mo> </mover> <mo>=</mo> <mfrac> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>i</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>y</mi> <mi>i</mi> </msub> </mrow> <mi>m</mi> </mfrac> <mo>;</mo> </mrow>
In formula：--- the average value of meteorological variables in observation period；
yi--- the sampled instantaneous value of i-th of meteorological variables, i.e. sample in observation period, wherein, mistake, suspicious incorrect sample It should be discarded without in calculating, even yi=0；
N --- the total sample number in observation period, determined by sample frequency peace mean time section；
M --- correct sample number (m≤N) in observation period.
2. the GMS Calculation of precipitation system based on cloud mass changing features as claimed in claim 1, it is characterised in that The method of the Threshold segmentation also includes：
Calculate mean μkWith covariance Σk；Change ηkValue and try to achieve accuracy value
3. the GMS Calculation of precipitation system based on cloud mass changing features as claimed in claim 1, it is characterised in that The method of the Threshold segmentation also includes：The value of log-likelihood function is calculated, calculates its changing value
<mrow> <mfrac> <mrow> <mo>|</mo> <msup> <mi>L</mi> <mrow> <mo>(</mo> <mi>t</mi> <mo>)</mo> </mrow> </msup> <mrow> <mo>(</mo> <mi>&amp;Theta;</mi> <mo>)</mo> </mrow> <mo>-</mo> <msup> <mi>L</mi> <mrow> <mo>(</mo> <mi>t</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> </msup> <mrow> <mo>(</mo> <mi>&amp;Theta;</mi> <mo>)</mo> </mrow> <mo>|</mo> </mrow> <mrow> <msup> <mi>L</mi> <mrow> <mo>(</mo> <mi>t</mi> <mo>+</mo> <mn>1</mn> <mo>)</mo> </mrow> </msup> <mrow> <mo>(</mo> <mi>&amp;Theta;</mi> <mo>)</mo> </mrow> </mrow> </mfrac> <mo>&lt;</mo> <msup> <mn>10</mn> <mrow> <mo>-</mo> <mn>5</mn> </mrow> </msup> </mrow>
Or iterations just exits circulate operation more than stated number, initiation parameter step is otherwise performed；
The maximum a posteriori probability of pixel is calculated, the classification of pixel is obtained according to maximum a posteriori probability principle.
4. the GMS Calculation of precipitation system based on cloud mass changing features as claimed in claim 1, it is characterised in that In the value for calculating log-likelihood function, according to the log-likelihood function of whole image
<mrow> <mi>L</mi> <mrow> <mo>(</mo> <mi>&amp;Theta;</mi> <mo>)</mo> </mrow> <mo>=</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <mi>log</mi> <mi> </mi> <mi>f</mi> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>)</mo> </mrow> <mo>=</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <mi>l</mi> <mi>o</mi> <mi>g</mi> <mrow> <mo>(</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>&amp;pi;</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mi>S</mi> <mi>t</mi> <mo>(</mo> <mrow> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>|</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> <mo>,</mo> <msub> <mi>&amp;Lambda;</mi> <mi>k</mi> </msub> <mo>,</mo> <msub> <mi>v</mi> <mi>k</mi> </msub> </mrow> <mo>)</mo> <mo>)</mo> </mrow> </mrow>
The complexity of parameter is solved, introduces implicit variable znkSo that log-likelihood function is written as again
<mrow> <mi>L</mi> <mrow> <mo>(</mo> <mi>&amp;Theta;</mi> <mo>)</mo> </mrow> <mo>=</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mo>{</mo> <msub> <mi>log&amp;pi;</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mo>+</mo> <mi>log</mi> <mi> </mi> <mi>S</mi> <mi>t</mi> <mrow> <mo>(</mo> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>|</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> <mo>,</mo> <msub> <mi>&amp;Lambda;</mi> <mi>k</mi> </msub> <mo>,</mo> <msub> <mi>v</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mo>}</mo> <mo>;</mo> </mrow>
The internal relation that can be made up of according to student t distributions Gaussian Profile and gamma distribution product, log-likelihood function quilt It is re-written as following depicted
<mfenced open = "" close = ""> <mtable> <mtr> <mtd> <mrow> <mi>L</mi> <mrow> <mo>(</mo> <mi>&amp;Theta;</mi> <mo>)</mo> </mrow> <mo>=</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mo>{</mo> <msub> <mi>log&amp;pi;</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mo>-</mo> <mfrac> <mi>D</mi> <mn>2</mn> </mfrac> <mi>l</mi> <mi>o</mi> <mi>g</mi> <mrow> <mo>(</mo> <mn>2</mn> <mi>&amp;pi;</mi> <mo>)</mo> </mrow> <mo>+</mo> <mfrac> <mn>1</mn> <mn>2</mn> </mfrac> <mi>l</mi> <mi>o</mi> <mi>g</mi> <mo>|</mo> <msub> <mi>&amp;eta;</mi> <mi>k</mi> </msub> <msub> <mi>&amp;Lambda;</mi> <mi>k</mi> </msub> <mo>|</mo> <mo>-</mo> <mfrac> <mn>1</mn> <mn>2</mn> </mfrac> <msup> <mrow> <mo>(</mo> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mi>T</mi> </msup> <msub> <mi>&amp;eta;</mi> <mi>k</mi> </msub> <msub> <mi>&amp;Lambda;</mi> <mi>k</mi> </msub> <mrow> <mo>(</mo> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mo>}</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mo>+</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mrow> <mo>{</mo> <mrow> <mo>-</mo> <mi>log</mi> <mi>&amp;Gamma;</mi> <mrow> <mo>(</mo> <mfrac> <msub> <mi>v</mi> <mi>k</mi> </msub> <mn>2</mn> </mfrac> <mo>)</mo> </mrow> <mo>+</mo> <mfrac> <msub> <mi>v</mi> <mi>k</mi> </msub> <mn>2</mn> </mfrac> <mi>log</mi> <mrow> <mo>(</mo> <mfrac> <msub> <mi>v</mi> <mi>k</mi> </msub> <mn>2</mn> </mfrac> <mo>)</mo> </mrow> <mo>+</mo> <mrow> <mo>(</mo> <mrow> <mfrac> <msub> <mi>v</mi> <mi>k</mi> </msub> <mn>2</mn> </mfrac> <mo>-</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <msub> <mi>log&amp;eta;</mi> <mi>k</mi> </msub> <mo>-</mo> <mfrac> <msub> <mi>v</mi> <mi>k</mi> </msub> <mn>2</mn> </mfrac> <msub> <mi>&amp;eta;</mi> <mi>k</mi> </msub> </mrow> <mo>}</mo> </mrow> <mo>.</mo> </mrow> </mtd> </mtr> </mtable> </mfenced>
5. the GMS Calculation of precipitation system based on cloud mass changing features as claimed in claim 4, it is characterised in that L (Θ) is found a function on parameter ηkDerivative be
<mrow> <mfrac> <mrow> <mo>&amp;part;</mo> <mi>L</mi> <mrow> <mo>(</mo> <mi>&amp;Theta;</mi> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <msub> <mi>&amp;eta;</mi> <mi>k</mi> </msub> </mrow> </mfrac> <mo>=</mo> <munderover> <mi>&amp;Sigma;</mi> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mo>&amp;lsqb;</mo> <mfrac> <mi>D</mi> <mrow> <mn>2</mn> <msub> <mi>&amp;eta;</mi> <mi>k</mi> </msub> </mrow> </mfrac> <mo>-</mo> <mfrac> <mn>1</mn> <mn>2</mn> </mfrac> <msup> <mrow> <mo>(</mo> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mi>T</mi> </msup> <msub> <mi>&amp;Lambda;</mi> <mi>k</mi> </msub> <mrow> <mo>(</mo> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mo>-</mo> <mrow> <mo>(</mo> <mfrac> <msub> <mi>v</mi> <mi>k</mi> </msub> <mn>2</mn> </mfrac> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> <mfrac> <mn>1</mn> <msub> <mi>&amp;eta;</mi> <mi>k</mi> </msub> </mfrac> <mo>-</mo> <mfrac> <msub> <mi>v</mi> <mi>k</mi> </msub> <mn>2</mn> </mfrac> <mo>&amp;rsqb;</mo> </mrow>
Wherein D represents the dimension of data, image for it is colored when D=3, D=1 during gray scale；SetObtain
<mrow> <munderover> <mi>&amp;Sigma;</mi> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mrow> <mo>&amp;lsqb;</mo> <mrow> <mfrac> <mi>D</mi> <mrow> <mn>2</mn> <msub> <mi>&amp;eta;</mi> <mi>k</mi> </msub> </mrow> </mfrac> <mo>-</mo> <mfrac> <mn>1</mn> <mn>2</mn> </mfrac> <msup> <mrow> <mo>(</mo> <mrow> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> </mrow> <mo>)</mo> </mrow> <mi>T</mi> </msup> <msub> <mi>&amp;Lambda;</mi> <mi>k</mi> </msub> <mrow> <mo>(</mo> <mrow> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mrow> <mo>(</mo> <mrow> <mfrac> <msub> <mi>v</mi> <mi>k</mi> </msub> <mn>2</mn> </mfrac> <mo>-</mo> <mn>1</mn> </mrow> <mo>)</mo> </mrow> <mfrac> <mn>1</mn> <msub> <mi>&amp;eta;</mi> <mi>k</mi> </msub> </mfrac> <mo>-</mo> <mfrac> <msub> <mi>v</mi> <mi>k</mi> </msub> <mn>2</mn> </mfrac> </mrow> <mo>&amp;rsqb;</mo> </mrow> <mo>=</mo> <mn>0</mn> </mrow>
Try to achieve ηkValue be
<mrow> <msub> <mi>&amp;eta;</mi> <mi>k</mi> </msub> <mo>=</mo> <mfrac> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>D</mi> <mo>-</mo> <msub> <mi>v</mi> <mi>k</mi> </msub> <mo>+</mo> <mn>2</mn> <mo>)</mo> </mrow> </mrow> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mo>&amp;lsqb;</mo> <msup> <mrow> <mo>(</mo> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mi>T</mi> </msup> <msub> <mi>&amp;Lambda;</mi> <mi>k</mi> </msub> <mrow> <mo>(</mo> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>v</mi> <mi>k</mi> </msub> <mo>&amp;rsqb;</mo> </mrow> </mfrac> <mo>;</mo> </mrow>
L (Θ) is found a function on parameter ΛkDerivative be
<mrow> <mfrac> <mrow> <mo>&amp;part;</mo> <mi>L</mi> <mrow> <mo>(</mo> <mi>&amp;Theta;</mi> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <msub> <mi>&amp;Lambda;</mi> <mi>k</mi> </msub> </mrow> </mfrac> <mo>=</mo> <mfrac> <mn>1</mn> <mn>2</mn> </mfrac> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mrow> <mo>(</mo> <msubsup> <mi>&amp;Lambda;</mi> <mi>k</mi> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msubsup> <mo>-</mo> <msub> <mi>&amp;eta;</mi> <mi>k</mi> </msub> <msup> <mrow> <mo>(</mo> <mrow> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> </mrow> <mo>)</mo> </mrow> <mi>T</mi> </msup> <mo>(</mo> <mrow> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> </mrow> <mo>)</mo> <mo>)</mo> </mrow> </mrow>
SetObtain equation
<mrow> <mfrac> <mn>1</mn> <mn>2</mn> </mfrac> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mrow> <mo>(</mo> <msubsup> <mi>&amp;Lambda;</mi> <mi>k</mi> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msubsup> <mo>-</mo> <msub> <mi>&amp;eta;</mi> <mi>k</mi> </msub> <msup> <mrow> <mo>(</mo> <mrow> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> </mrow> <mo>)</mo> </mrow> <mi>T</mi> </msup> <mo>(</mo> <mrow> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> </mrow> <mo>)</mo> <mo>)</mo> </mrow> <mo>=</mo> <mn>0</mn> </mrow>
Try to achieve ΛkValue be
<mrow> <msub> <mi>&amp;Lambda;</mi> <mi>k</mi> </msub> <mo>=</mo> <msup> <mrow> <mo>&amp;lsqb;</mo> <mfrac> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <msub> <mi>&amp;eta;</mi> <mi>k</mi> </msub> <msup> <mrow> <mo>(</mo> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mi>T</mi> </msup> <mrow> <mo>(</mo> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> </mrow> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> </mrow> </mfrac> <mo>&amp;rsqb;</mo> </mrow> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msup> <mo>.</mo> </mrow>
6. the GMS Calculation of precipitation system based on cloud mass changing features as claimed in claim 5, it is characterised in that L (Θ) is found a function on parameter μkDerivative be
<mrow> <mfrac> <mrow> <mo>&amp;part;</mo> <mi>L</mi> <mrow> <mo>(</mo> <mi>&amp;Theta;</mi> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> </mrow> </mfrac> <mo>=</mo> <mo>-</mo> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mo>&amp;lsqb;</mo> <msub> <mi>&amp;Lambda;</mi> <mi>k</mi> </msub> <mrow> <mo>(</mo> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mo>&amp;rsqb;</mo> </mrow>
Set <mrow> <mfrac> <mrow> <mo>&amp;part;</mo> <mi>L</mi> <mrow> <mo>(</mo> <mi>&amp;Theta;</mi> <mo>)</mo> </mrow> </mrow> <mrow> <mo>&amp;part;</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> </mrow> </mfrac> <mo>=</mo> <mn>0</mn> <mo>,</mo> </mrow> Obtain equation,
<mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mo>&amp;lsqb;</mo> <msub> <mi>&amp;Lambda;</mi> <mi>k</mi> </msub> <mrow> <mo>(</mo> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mo>&amp;rsqb;</mo> <mo>=</mo> <mn>0</mn> </mrow>
Try to achieve μkValue be
<mrow> <msub> <mi>&amp;mu;</mi> <mi>k</mi> </msub> <mo>=</mo> <mfrac> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mover> <msub> <mi>x</mi> <mi>n</mi> </msub> <mo>&amp;OverBar;</mo> </mover> </mrow> <mrow> <munderover> <mo>&amp;Sigma;</mo> <mrow> <mi>n</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> </mrow> </mfrac> <mo>.</mo> </mrow>
7. the GMS Calculation of precipitation system based on cloud mass changing features as claimed in claim 3, it is characterised in that After the cycle-index that the rate of change of log-likelihood function value reaches certain limit or iteration reaches certain input, according to maximum Posterior probability criterion
<mrow> <mi>arg</mi> <munder> <mrow> <mi>m</mi> <mi>a</mi> <mi>x</mi> </mrow> <mi>k</mi> </munder> <mo>{</mo> <msub> <mi>z</mi> <mrow> <mi>n</mi> <mi>k</mi> </mrow> </msub> <mo>}</mo> </mrow>
Obtain the mark of pixel.
8. a kind of GMS Calculation of precipitation system based on cloud mass changing features as claimed in claim 1 based on cloud mass The GMS Calculation of precipitation method of changing features, it is characterised in that the static meteorology based on cloud mass changing features Satellite Calculation of precipitation method comprises the following steps：
Step 1, satellite Simulation of Precipitation parameter attribute collection is obtained to portray the generation of cloud precipitation and evolution, is returned from most value Different dimension Cloud-Picture Characteristics parameters are normalized one change method；
Step 2, build based on to the satellite Calculation of precipitation model of type reverse transmittance nerve network, being dropped before three layers for the region Water is estimated, and uses the Simulation of Precipitation precision of multi objective system anlysis model.
CN201710448042.5A 2017-06-14 2017-06-14 A kind of GMS Calculation of precipitation method based on cloud mass changing features Pending CN107341449A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710448042.5A CN107341449A (en) 2017-06-14 2017-06-14 A kind of GMS Calculation of precipitation method based on cloud mass changing features

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710448042.5A CN107341449A (en) 2017-06-14 2017-06-14 A kind of GMS Calculation of precipitation method based on cloud mass changing features

Publications (1)

Publication Number Publication Date
CN107341449A true CN107341449A (en) 2017-11-10

Family

ID=60220592

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710448042.5A Pending CN107341449A (en) 2017-06-14 2017-06-14 A kind of GMS Calculation of precipitation method based on cloud mass changing features

Country Status (1)

Country Link
CN (1) CN107341449A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109410313A (en) * 2018-02-28 2019-03-01 南京恩瑞特实业有限公司 A kind of meteorology three-dimensional information 3D simulation inversion method
CN111860562A (en) * 2020-03-30 2020-10-30 中国人民解放军国防科技大学 Self-adaptive dry-wet distinguishing method based on multiple statistics of microwave link
CN111983732A (en) * 2020-07-27 2020-11-24 南京信息工程大学 Rainfall intensity estimation method based on deep learning
US11131790B2 (en) 2019-03-25 2021-09-28 Yandex Europe Ag Method of and system for generating weather forecast

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104200480A (en) * 2014-09-17 2014-12-10 西安电子科技大学宁波信息技术研究院 Image fuzzy degree evaluation method and system applied to intelligent terminal
CN104820754A (en) * 2015-05-13 2015-08-05 南京信息工程大学 Space statistical downscaling rainfall estimation method based on geographical difference analysis method
CN105354409A (en) * 2015-10-14 2016-02-24 成都信息工程大学 Research method for diagnosis identification and estimation of plateau mountain precipitation information
CN106709918A (en) * 2017-01-20 2017-05-24 成都信息工程大学 Method for segmenting images of multi-element student t distribution mixed model based on spatial smoothing

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104200480A (en) * 2014-09-17 2014-12-10 西安电子科技大学宁波信息技术研究院 Image fuzzy degree evaluation method and system applied to intelligent terminal
CN104820754A (en) * 2015-05-13 2015-08-05 南京信息工程大学 Space statistical downscaling rainfall estimation method based on geographical difference analysis method
CN105354409A (en) * 2015-10-14 2016-02-24 成都信息工程大学 Research method for diagnosis identification and estimation of plateau mountain precipitation information
CN106709918A (en) * 2017-01-20 2017-05-24 成都信息工程大学 Method for segmenting images of multi-element student t distribution mixed model based on spatial smoothing

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
叶颖: "《基于ARM的水文气象数据采集系统的设计》", 《中国优秀硕士学位论文全文数据库 基础科技辑》 *
夏双 等: "《基于FY-2C卫星数据的藏北高原降水估算研究》", 《长江流域资源与环境》 *
杨俊 等: "《基于归一化RGB色彩模型的阴影处理方法》", 《光电工程》 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109410313A (en) * 2018-02-28 2019-03-01 南京恩瑞特实业有限公司 A kind of meteorology three-dimensional information 3D simulation inversion method
CN109410313B (en) * 2018-02-28 2023-03-24 南京恩瑞特实业有限公司 Meteorological three-dimensional information 3D simulation inversion method
US11131790B2 (en) 2019-03-25 2021-09-28 Yandex Europe Ag Method of and system for generating weather forecast
RU2757591C1 (en) * 2019-03-25 2021-10-19 Общество С Ограниченной Ответственностью «Яндекс» Method and system for generating weather forecast
CN111860562A (en) * 2020-03-30 2020-10-30 中国人民解放军国防科技大学 Self-adaptive dry-wet distinguishing method based on multiple statistics of microwave link
CN111860562B (en) * 2020-03-30 2022-04-12 中国人民解放军国防科技大学 Self-adaptive dry-wet distinguishing method based on multiple statistics of microwave link
CN111983732A (en) * 2020-07-27 2020-11-24 南京信息工程大学 Rainfall intensity estimation method based on deep learning

Similar Documents

Publication Publication Date Title
CN108573276B (en) Change detection method based on high-resolution remote sensing image
CN106355151B (en) A kind of three-dimensional S AR images steganalysis method based on depth confidence network
CN106228185B (en) A kind of general image classifying and identifying system neural network based and method
Yu et al. Context-based hierarchical unequal merging for SAR image segmentation
WO2018023734A1 (en) Significance testing method for 3d image
CN103413151B (en) Hyperspectral image classification method based on figure canonical low-rank representation Dimensionality Reduction
CN107563355A (en) Hyperspectral abnormity detection method based on generation confrontation network
CN107330357A (en) Vision SLAM closed loop detection methods based on deep neural network
CN102982338B (en) Classification of Polarimetric SAR Image method based on spectral clustering
CN112395987B (en) SAR image target detection method based on unsupervised domain adaptive CNN
CN109726748B (en) GL-CNN remote sensing image scene classification method based on frequency band feature fusion
CN107341449A (en) A kind of GMS Calculation of precipitation method based on cloud mass changing features
CN106600595A (en) Human body characteristic dimension automatic measuring method based on artificial intelligence algorithm
CN113160062B (en) Infrared image target detection method, device, equipment and storage medium
CN101216942A (en) An increment type characteristic background modeling algorithm of self-adapting weight selection
CN108460391A (en) Based on the unsupervised feature extracting method of high spectrum image for generating confrontation network
CN103208001A (en) Remote sensing image processing method combined with shape self-adaption neighborhood and texture feature extraction
CN104299232B (en) SAR image segmentation method based on self-adaptive window directionlet domain and improved FCM
CN103279957A (en) Method for extracting remote sensing image interesting area based on multi-scale feature fusion
CN107368852A (en) A kind of Classification of Polarimetric SAR Image method based on non-down sampling contourlet DCGAN
CN103366184B (en) Polarization SAR data classification method based on hybrid classifer and system
CN110163213A (en) Remote sensing image segmentation method based on disparity map and multiple dimensioned depth network model
CN102902956A (en) Ground-based visible cloud image recognition processing method
CN105469111A (en) Small sample set object classification method on basis of improved MFA and transfer learning
CN104484890A (en) Video target tracking method based on compound sparse model

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20171110

RJ01 Rejection of invention patent application after publication