CN114997534A - Similar rainfall forecasting method and equipment based on visual features - Google Patents
Similar rainfall forecasting method and equipment based on visual features Download PDFInfo
- Publication number
- CN114997534A CN114997534A CN202210909459.8A CN202210909459A CN114997534A CN 114997534 A CN114997534 A CN 114997534A CN 202210909459 A CN202210909459 A CN 202210909459A CN 114997534 A CN114997534 A CN 114997534A
- Authority
- CN
- China
- Prior art keywords
- rainfall
- historical
- similarity
- similar
- current
- 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
- 230000000007 visual effect Effects 0.000 title claims abstract description 86
- 238000013277 forecasting method Methods 0.000 title claims abstract description 11
- 238000009826 distribution Methods 0.000 claims abstract description 141
- 238000000034 method Methods 0.000 claims abstract description 73
- 238000010586 diagram Methods 0.000 claims description 48
- 230000008569 process Effects 0.000 claims description 35
- 239000011159 matrix material Substances 0.000 claims description 26
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 13
- 238000005192 partition Methods 0.000 claims description 12
- 238000011156 evaluation Methods 0.000 claims description 8
- 150000001875 compounds Chemical class 0.000 claims description 6
- 230000002093 peripheral effect Effects 0.000 claims description 6
- 239000003086 colorant Substances 0.000 claims description 5
- 238000000691 measurement method Methods 0.000 claims description 4
- 238000012163 sequencing technique Methods 0.000 claims description 4
- 238000002372 labelling Methods 0.000 claims description 3
- 239000002245 particle Substances 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims description 3
- 238000004088 simulation Methods 0.000 claims 2
- 238000012545 processing Methods 0.000 abstract description 9
- 238000004364 calculation method Methods 0.000 description 7
- 238000011160 research Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 230000001186 cumulative effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000001556 precipitation Methods 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 238000010998 test method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/587—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using geographical or spatial information, e.g. location
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Tourism & Hospitality (AREA)
- Human Resources & Organizations (AREA)
- Databases & Information Systems (AREA)
- Library & Information Science (AREA)
- Strategic Management (AREA)
- Data Mining & Analysis (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- General Health & Medical Sciences (AREA)
- Development Economics (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Educational Administration (AREA)
- Multimedia (AREA)
- Primary Health Care (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a similar rainfall forecasting method and equipment based on visual features, wherein the method comprises the steps of constructing a historical rainfall picture library and extracting the visual features of the historical rainfall; the historical rainfall visual characteristics comprise total rainfall, rainfall spatial distribution and rainfall center; constructing a current rainfall picture set, extracting current rainfall visual characteristics, measuring the distance between the historical rainfall visual characteristics and the current rainfall visual characteristics, and calculating the rainfall similarity between the current rainfall visual characteristics and the historical rainfall visual characteristics in each period; and giving weight to each rainfall similarity to obtain comprehensive rainfall similarity, and outputting the comprehensive rainfall similarity in sequence. The rainfall data spatial characteristics can be displayed more intuitively, and the data processing speed and the similarity contrast accuracy are improved.
Description
Technical Field
The invention belongs to the field of flood forecasting, and particularly relates to a similar rainfall forecasting method based on visual characteristics.
Background
In recent years, problems caused by global climate change are becoming more severe, and rainstorm flood disasters are frequent. Rain flood prediction is increasingly regarded as a work affecting economic construction and life and property safety. To this end, the skilled person proposes a number of solutions.
In the existing solution, the analysis method of the rainfall flood similarity is mostly biased to the correlation statistics of a single feature, and the processing of multiple features and multiple samples is relatively limited. In addition, the rainfall flood characteristic information is input with relative shortage, and the effective characteristic analysis aspect is weak; however, the performance of the machine learning algorithm is good and bad, and depends on the characteristics of the model input to a great extent. Finally, the rainfall similarity is directly analyzed according to rainfall data of the rainfall station, the data are difficult to visually display, and meanwhile, the rainfall flood process cannot be accurately described from the system level.
Disclosure of Invention
The purpose of the invention is as follows: a similar rainfall forecasting method based on visual characteristics is provided to solve the problems in the prior art.
The technical scheme is as follows: the similar rainfall forecasting method based on the visual characteristics comprises the following steps:
s1, constructing historical rainfall picture library
S11, collecting historical rainfall data in a plurality of periods in a preset area, and drawing a plurality of groups of historical rainfall distribution maps based on the historical rainfall data;
s12, extracting historical rainfall visual features based on each group of historical rainfall distribution map; the historical rainfall visual characteristics comprise total rainfall, rainfall spatial distribution and rainfall center;
s2, constructing a current rainfall picture set, extracting the current rainfall visual characteristics,
s3, measuring the distance between the historical rainfall visual characteristics and the current rainfall visual characteristics, and calculating the rainfall similarity between the current rainfall visual characteristics and the historical rainfall visual characteristics in each period; and giving weight to each rainfall similarity to obtain comprehensive rainfall similarity, and outputting the comprehensive rainfall similarity in sequence.
According to an aspect of the application, the step S11 is further:
acquiring geographic data of a preset area and rainfall data of each rainfall station in each period; drawing a map of a predetermined area based on the geographic data and rasterizing; and filling color parameters for the rasterized map based on the rainfall data to obtain a historical rainfall distribution diagram.
According to an aspect of the application, the step S12 is further:
reading the historical rainfall distribution map, converting the historical rainfall distribution map into a gray scale map, setting a numerical interval of the gray scale map according to the rainfall level, and converting the gray scale map into an NxN gray scale picture matrix;
and binarizing the gray level picture matrix according to the rainfall level to obtain a binarized rainfall matrix corresponding to different rainfall levels.
According to one aspect of the application, the total rainfall isRTComprises the following steps:
；
in the formula (I), the compound is shown in the specification,kthe number of colors in each historical rainfall distribution graph is obtained;p（C i ）is as followsiThe frequency with which the seed color appears is,v （C i ）is as followsiAnd the rainfall value corresponding to each color.
According to one aspect of the application, the rainfall spatial distributionRComprises the following steps:
；
m and n are the number of rows and columns after each historical rainfall distribution graph is partitioned;r m×n the rainfall value of the block corresponding to the mth row and the nth column is obtained.
According to one aspect of the application, the rain centerC=（C x ，C y ）；
In the formula (I), the compound is shown in the specification,
，r ab and m and n are the number of rows and columns after each historical rainfall distribution graph is partitioned.
According to an aspect of the present application, the process of calculating the rainfall similarity between the current rainfall visual characteristics and the historical rainfall visual characteristics in each period in step S3 further includes:
step S31, using Minkowski distanced1Calculating the total rainfall similarity of the preset basin, d1=RT；
Step S32, calculating the rainfall center similarity of the preset watershed by adopting the Euclidean distance d2,
step S33, adopting the rainfall approximation rate of the partitioned areas to express the rainfall spatial distribution approximation rate:
step S330, determining whether two blocks in the two rainfall spatial distributions R are similar;
respectively reading rainfall values of the two blocks, wherein the larger value is max, the smaller value is min, the rainfall difference ratio is defined as (max-min)/max, the threshold value of the rainfall difference ratio is alpha, and when the rainfall difference ratio is smaller than or equal to the threshold value, the two blocks are similar;
step S331, calculating logarithms of similar blocks in two rainfall spatial distributions R;
repeating the step S330, calculating the similarity of all the blocks in the two rainfall spatial distributions R, and counting the number of similar blocks;
step S332, calculating a rainfall spatial distribution approximation rate; and if each rainfall spatial distribution has c blocks, wherein d pairs of blocks in the two rainfall spatial distributions are similar blocks, the approximation rate of the two rainfall spatial distributions is (1-d/c).
According to an aspect of the present application, in the step 3, a weight is given to each rainfall similarity, and the process of obtaining the comprehensive rainfall similarity further includes:
when the single-day rainfall similarity is calculated, adopting a subjective and objective integrated weighting method:
selecting part of current rainfall picture sets as labeling samples to label data:
the similarity degree of each labeled sample and all samples in the historical rainfall picture library is scored;
and linearly weighting each rainfall visual characteristic index of the marked sample according to the random weight to obtain an evaluation result, calculating according to the evaluation result and the marking result to obtain an NDCG @ K value, and searching for the optimal weight by using a particle swarm algorithm to enable the average NDCG value of the whole marked sample to be maximum.
According to one aspect of the application, when calculating the similarity of rainfall for multiple days, a rainfall sequence searching method based on a sliding window and weighted dynamic time warping is adopted:
inputting a current rainfall visual characteristic sequence Q with the length of e and a historical rainfall visual characteristic sequence S with the length of f in a certain period, wherein e is smaller than f;
sliding a window with the size of e on the historical rainfall visual characteristic sequence S, wherein the sliding step length is 1, and taking the sequence in each window as a subsequence to obtain a sequence set;
obtaining the distance between the current rainfall visual characteristic sequence Q and each subsequence P in the sequence set by adopting a weighted dynamic time warping algorithm, wherein the distance between any two points in the sequence is calculated by using a single-day rainfall characteristic multivariate characteristic distance measurement method;
and sequencing all the subsequences according to the distance to obtain a specified number of subsequences with the minimum distance as a search result.
According to one aspect of the application, further comprising:
s333, reading the similar blocks in the step S332, and respectively communicating the adjacent similar blocks of the two rainfall pictures to form at least one similar domain;
respectively judging whether the current historical rainfall distribution diagram and the rainfall centers of the historical rainfall distribution diagram are located in the similar domain;
if the current rainfall distribution diagram and the historical rainfall distribution diagram are larger than the threshold value, the historical rainfall distribution diagram is used as an output similar image alternative;
and S334, if the current historical rainfall distribution diagram and the rainfall centers of the historical rainfall distribution diagram are both located in the similar domain, obtaining the peripheral profiles of the similar domain, calculating the curve similarity of the two peripheral profile curves, and if the curve similarity is greater than a preset value, outputting the historical rainfall distribution diagram as a local similar alternative.
According to one aspect of the present application, the process of calculating the rainfall similarity between the current rainfall visual characteristic and the rainfall visual characteristic in each historical period further comprises:
acquiring water system characteristic data of a preset area, and dividing the preset area into a preset number of sub-partitions according to the water system characteristics;
respectively calculating the surface rainfall of each current sub-partition, combining the surface rainfall into a surface rainfall matrix, and further calculating the total rainfall of a preset area;
respectively calculating the surface rainfall of each sub-partition in each period of history based on the historical rainfall data and the water system characteristics, combining the surface rainfall into a surface rainfall matrix, and further calculating the total rainfall of a preset area;
and calculating Euclidean distances of the current rainfall matrix and the rainfall matrix of each historical period, taking the Euclidean distances as similarity numerical values and arranging the similarity numerical values in a descending order.
There is further provided an apparatus comprising:
at least one processor; and
a memory communicatively coupled to at least one of the processors; wherein the content of the first and second substances,
the memory stores instructions executable by the processor to implement a visual feature-based rain forecasting method of any embodiment.
Has the advantages that: a new rainfall forecast thought is provided, the spatial characteristics of rainfall data can be displayed more intuitively, and the speed of data processing and the accuracy of similarity comparison are improved.
Drawings
FIG. 1 is a flow chart of the first embodiment.
Fig. 2 is a flowchart of similarity calculation.
FIG. 3 is a flowchart of the second embodiment.
Detailed Description
As shown in fig. 1, the technical principle and related processes of the technical solution of the present application are described in detail.
In order to solve the problems in the prior art, the applicant has conducted intensive research, and in the existing rainfall and flood prediction methods, direct rainfall data are basically adopted, a historical database is constructed through rainfall or flood data containing geographic position information and time information, then data mining is conducted through relevant technical parameters, and similarity matching is generally conducted through methods such as a neural network. For example, in one embodiment, the similarity of rainfall is calculated by statistically processing the sum of the rainfall for a single rainfall, the number of days of rainfall for a single rainfall, the average rainfall, the maximum rainfall for a single day, the sum of the rainfall for different rainfall levels, for example, the sum of the rainfall for a single rainfall day with a rainfall of less than 5mm, or the sum of the rainfall for a single rainfall day with a rainfall of greater than 50mm, and clustering. Then, optimization is carried out in the rainfall process with higher similarity, and corresponding results are obtained. The existing various similarity research and comparison methods are difficult to visually display the data spatial distribution characteristics. The existing rainfall flood prediction method added with the geographical position information only generates a grid matrix from a basin outline, then generates a matrix sequence from rainfall data, predicts the rainfall process by a method of calculating a similarity matrix, and basically directly calculates through a data hierarchy, so that the intuitiveness is not enough.
For this purpose, the applicant proposes the following method.
In one embodiment of the application, a similar rainfall forecasting method based on visual features is provided, which mainly comprises the following processes. The rainfall test method comprises the steps of firstly making a rainfall distribution map based on historical rainfall data, building a rainfall graph database of each rainfall period in history through the rainfall distribution map, then making part of rainfall data into a test set by adopting the same or similar process, calculating the similarity through a graph similarity algorithm, judging the credibility of a scheme, and adjusting relevant parameters and details of the algorithm. Through the process, the theoretical feasibility and operability of the method are established. When the rainfall monitoring system is used, the current rainfall data are obtained, the current rainfall picture set is constructed, and a plurality of historical rainfall processes with certain similarity can be found, so that a prevention scheme can be formulated through data of historically related rainfall processes, and reference is provided for the current rainfall process through processing experience of the historical rainfall processes.
Specifically, the step S1 of constructing the historical rainfall picture library specifically includes:
historical rainfall data in a plurality of periods of a preset area is collected, and a plurality of groups of historical rainfall distribution graphs are drawn based on the historical rainfall data.
Specifically, firstly, acquiring geographic data of a preset area and rainfall data of each rainfall station in each period; then drawing a map of a predetermined area based on the geographic data and rasterizing; and finally, filling color parameters for the rasterized map based on rainfall data to obtain a historical rainfall distribution diagram.
In the historical rainfall distribution diagram, the spatial distribution characteristics of rainfall can be more clearly displayed, such as the position of a rainfall center, the distribution of rainfall in different areas and the like. Because the requirement of rainfall similarity search on data precision is not high, the advantage of using the rainfall distribution graph is more obvious, and the defects that the calculated amount is increased due to the fact that the prior art is too much limited by data details and the description of rainfall distribution on the space-time whole is inaccurate are avoided. Through the display of rainfall data on regional distribution, the spatial change characteristics of rainfall can be more accurately mastered, and the research on the evolution process of rainfall in time and space is facilitated. It should be noted that the rainfall value in a certain geographical position can be obtained by using data of a plurality of rainfall stations and then using a spatial interpolation method.
In order to better study the rainfall distribution diagram, the extraction of rainfall features in the historical rainfall distribution diagram, particularly the distribution features of rainfall on the space, is the key point for improving the search efficiency and accuracy. For this purpose, research is conducted to extract the total rainfall amount of a research area (for example, on a certain river basin), the spatial distribution of rainfall and rainfall characteristics of a rainfall center as a historical rainfall distribution map. Specifically, the procedure is as follows: extracting historical rainfall visual features based on each group of historical rainfall distribution map; the historical rainfall visual characteristics include total rainfall, rainfall spatial distribution and rainfall center.
In other words, the principle of similarity search of the rainfall distribution map is to extract rainfall features, quantify distances among the rainfall features, and then sort the rainfall features according to the quantified distances to obtain the top K results with the smallest distances as the result of the similarity search.
Rainfall feature extraction and distance measurement are key in the rainfall similarity search process. Rainfall has multiple characteristics, so the rainfall similarity search relates to the problem of multiple comprehensive evaluation, the rainfall needs to be represented by the multiple characteristics, the distances among the characteristics of different rainfalls are measured and subjected to distance fusion, and then the rainfall similarity search can be carried out.
The whole rainfall similarity searching process is as follows: extracting rainfall characteristics, measuring rainfall characteristic distance, fusing a plurality of rainfall characteristics, and sequencing according to the fused distance to obtain a search result.
Specifically, any one of the historical rainfall distribution graphs is converted into a gray-scale graph, and the gray-scale graph is converted into a gray-scale picture matrix based on the rainfall distribution level corresponding to the gray scale of each range. Namely: reading the historical rainfall distribution map, converting the historical rainfall distribution map into a gray scale map, setting a numerical interval of the gray scale map according to the rainfall level, and converting the gray scale map into an NxN gray scale picture matrix; and binarizing the gray level picture matrix according to the rainfall level to obtain a binarized rainfall matrix corresponding to different rainfall levels. N is a natural number. For example, 32 × 32 may be used.
In a further embodiment of the method of the invention,
total rainfallRTComprises the following steps:
(ii) a RT is an abbreviation for total rainfall (rain total).
In the formula (I), the compound is shown in the specification,kthe number of colors in each historical rainfall distribution graph is obtained;p（C i ）is as followsiThe frequency with which the seed color appears is,v （C i ）is a firstiThe rainfall value corresponding to each color.
Spatial distribution of rainfallRComprises the following steps:
(ii) a m and n are the number of rows and columns of each historical rainfall distribution graph after being partitioned;r m×n the rainfall value of the block corresponding to the mth row and the nth column is obtained.
Rainfall centerC=（C x ，C y ）；
In the formula (I), the compound is shown in the specification,
，r ab the rainfall values of the a-th row and the b-th column are m and n are the row number and the column number of each historical rainfall distribution graph after being partitioned.
And S2, collecting rainfall data, generating a current rainfall distribution map, constructing a current rainfall picture set, and extracting current rainfall visual characteristics.
During testing or using, the rainfall picture to be inquired is extracted by the method, and whether similar rainfall pictures exist or not is searched in the historical rainfall picture library, so that the rainfall process in a certain period is corresponded. It should be noted that, in order to distinguish the usage of the picture library, the picture library is divided into a historical rainfall picture library and a current rainfall picture set. However, the same method is adopted for processing the rainfall pictures related to the rainfall pictures, such as extracting rainfall features. Therefore, the current rainfall visual characteristics also include characteristics of total rainfall, rainfall spatial distribution, rainfall center distribution and the like.
After the rainfall pictures are obtained and the rainfall visual characteristics are established, how to compare the rainfall processes of the areas in different periods and calculate the similarity of the rainfall processes is particularly important. For this reason, a method is proposed as described in step S3.
S3, measuring the distance between the historical rainfall visual characteristics and the current rainfall visual characteristics, and calculating the rainfall similarity between the current rainfall visual characteristics and the historical rainfall visual characteristics in each period; and giving weight to each rainfall similarity to obtain comprehensive rainfall similarity, and outputting the comprehensive rainfall similarity in sequence.
That is, during a period of time, such as a current river basin, during rainfall, a current rainfall profile, such as a rainfall profile of a river basin on a day, may be drawn according to current rainfall data. Meanwhile, according to the calculation process, the rainfall visual characteristics of the current rainfall distribution map can be obtained. In a simpler case, a historical rainfall profile that is closer to the current rainfall visual characteristics may be looked up in a historical rainfall database. Briefly, if the visual characteristics of rainfall in a certain historical rainfall distribution map are close to those of rainfall in the current rainfall distribution map, the two are similar. For example, under a certain environment, three rainfall visual characteristic values, such as total rainfall, rainfall spatial distribution and rainfall center, are very close, and then the two values are very close. In practice, however, the method is very complex, and due to the existence of three parameters, the method needs to be given an authority so as to perform comprehensive judgment, otherwise, in a certain situation, the rainfall centers in a historical rainfall distribution graph are very close, the rainfall spatial distribution is also very close, but the total rainfall is not close. The rainfall spatial distribution in another historical rainfall distribution map is very close, and the rainfall center is not close to the total rainfall, so that which historical rainfall distribution map is the closest cannot be given.
In a further embodiment, the process of calculating the rainfall similarity between the current rainfall visual characteristic and the historical rainfall visual characteristic in each period is further as follows:
step S31, using Minkowski distanced1Calculating the total rainfall similarity of the preset basin, d1=RT；
Namely the difference value between the total rainfall capacity of the current rainfall picture to be detected and the total rainfall capacity of a certain historical rainfall picture.
Step S32, calculating the rainfall center similarity of the preset watershed by adopting the Euclidean distance d2,
step S33, adopting the partitioned area rainfall approximation to represent the rainfall spatial distribution approximation:
step S330, determining whether two blocks in the two rainfall spatial distributions R are similar;
respectively reading rainfall values of the two blocks, wherein the larger value is max, the smaller value is min, the rainfall difference ratio is defined as (max-min)/max, the threshold value of the rainfall difference ratio is alpha, and when the rainfall difference ratio is smaller than or equal to the threshold value, the two blocks are similar;
step S331, calculating logarithms of similar blocks in two rainfall spatial distributions R;
repeating the step S330, calculating the similarity of all the blocks in the two rainfall spatial distributions R, and counting the number of similar blocks;
step S332, calculating a rainfall spatial distribution approximation rate; and (3) c blocks are arranged in each rainfall spatial distribution, wherein d pairs of blocks in the two rainfall spatial distributions are similar blocks, and the approximation rate of the two rainfall spatial distributions is (1-d/c).
In the above embodiment, given the mathematical representation of the visual features of rainfall, the similarity of the individual visual features of rainfall is calculated using different methods, which can be calculated using minkowski distance since the total rainfall is one-dimensional data, whereas the rainfall center relates to two-dimensional coordinates and is thus calculated by euclidean distance. Of course in other embodiments a similar approach may be used, for example by scaling the total rainfall, i.e. d1= RT1/RT 0.
For spatial rainfall distribution, the rainfall space is partitioned, for example, a river basin rainfall picture is partitioned into a plurality of small blocks, and whether the current rainfall picture is similar to the historical rainfall distribution map or not is judged by calculating the similarity of the small blocks and the proportion of the similar small blocks in all the small blocks. For example, in the current rainfall distribution diagram, there are more than threshold blocks similar to the blocks of the historical rainfall distribution diagram in a certain period, that is, there are more similar blocks in the two pictures, so that the existence similarity between the two pictures can be inferred, and a basis is provided for judging the similarity between the current rainfall distribution diagram and the historical rainfall distribution diagram.
Experiments prove that the calculation of the rainfall spatial distribution approximation rate is more practical and obtains better effect. However, in some situations, the overall distribution similarity is large, but the similarity of other parameter characteristics such as the distribution of rainfall centers is not large, and the overall similarity is small, but through manual comparison, it is found that the similarity of a local area is large, and a historical rainfall distribution map with the large local area similarity can be actually used for reference and reference.
To this end, a preferred embodiment is given to solve the above-mentioned problems.
S333, reading the similar blocks in the step S332, and respectively communicating the adjacent similar blocks of the two rainfall pictures (the current rainfall distribution diagram and the historical rainfall distribution diagram) to form at least one similar domain;
respectively judging whether the current historical rainfall distribution diagram and the rainfall centers of the historical rainfall distribution diagram are located in the similar domain;
if the rainfall center of the current historical rainfall distribution diagram is located in the similar domain of the diagram, and the rainfall center of the historical rainfall distribution diagram is located in the similar domain of the diagram, calculating the maximum radius of the similar domain, (the maximum value of the center coordinates of each block and the rainfall center coordinates in the similar domain), calculating the generalized similar area (pi multiplied by the square of the maximum radius) through the maximum radius of the similar domain, and finally calculating the proportion of the generalized similar area to the rainfall distribution diagram in the current rainfall distribution diagram and the historical rainfall distribution diagram respectively, and if the generalized similar area and the rainfall distribution diagram in the current rainfall distribution diagram and the historical rainfall distribution diagram are both larger than a threshold value, taking the historical rainfall distribution diagram as an output similar image alternative;
and S334, if the current historical rainfall distribution diagram and the rainfall centers of the historical rainfall distribution diagram are both located in the similar domain, obtaining the peripheral profiles of the similar domain, calculating the curve similarity of the two peripheral profile curves, and if the curve similarity is greater than a preset value, outputting the historical rainfall distribution diagram as a local similar alternative.
Through above-mentioned embodiment, can solve in the great condition of local similarity, and whole similarity is not big, the condition of regional precipitation evolution. In some cases, the distribution of the rainfall has some regularity, but may be done in different places, outwards along the rain center. In other words, a certain historical rainfall process evolves at a place A, while the current rainfall process evolves at a place B away from the place A by a certain distance, and from the rainfall visual characteristics such as rainfall capacity, rainfall center and rainfall distribution, the two processes are similar, but in different areas.
This is common because the area of investigation is generally large. Therefore, the local similarity is measured, the area and the outer contour characteristics of the similar domain are calculated, and whether the evolution processes of the two rains are similar or not can be judged. Thereby further improving the accuracy of similarity search. That is to say, the scenes which do not satisfy the similar conditions in the whole area but partially satisfy the similar conditions are searched out, and a basis is provided for subsequent decisions.
In addition, since the historical rainfall distribution map includes data of rainfall in multiple periods and multiple levels, the data size is very large, and even if there is only one current rainfall distribution map, comparison needs to be performed in the historical rainfall distribution map one by one, and then a conclusion needs to be given, so that the calculation amount is very large, and therefore, a similarity search algorithm needs to be optimized. This is especially true for a multi-day rainfall profile.
In a further embodiment, besides the calculation of the local similarity by using the rainfall communication domain, the rainfall distribution and the flow direction can be analyzed and predicted from the aspect of the water system communication by performing the partition according to the water system characteristics, that is, by performing the partition of the sub-regions according to the water system distribution of the region. Specifically, the method mainly comprises the following steps:
the process of calculating the rainfall similarity of the current rainfall visual characteristics and the rainfall visual characteristics of each historical period further comprises the following steps:
acquiring water system characteristic data of a preset area, and dividing the preset area into a preset number of sub-partitions according to the water system characteristics;
respectively calculating the surface rainfall of each current sub-partition, combining the surface rainfall into a surface rainfall matrix, and further calculating the total rainfall of a preset area;
respectively calculating the surface rainfall of each sub-partition in each period of history based on the historical rainfall data and the water system characteristics, combining the surface rainfall into a surface rainfall matrix, and further calculating the total rainfall of a preset area;
and calculating Euclidean distances of the current rainfall matrix and the rainfall matrix of each historical period, taking the Euclidean distances as similarity numerical values and arranging the similarity numerical values in a descending order.
In this embodiment, in order to improve the calculation of the local rainfall similarity, a certain area, for example, the jialing river area, may be divided into 9 sub-areas, and then the surface rainfall of each sub-area is calculated by using the above method, so as to form a rainfall matrix, and calculate the total surface rainfall. Then, the historical rainfall data is partitioned according to the water system characteristics according to the process, and then the rainfall of the sub-partition surface is calculated. And finally, calculating the Euclidean distance between the current rainfall and the historical rainfall, wherein the Euclidean distance is the size of the representation similarity. So that several most similar rainfall events can be obtained.
In this embodiment, the specific processing procedure may be as follows: for current rainfall and historical rainfall, firstly, a rainfall distribution live image is generated through rainfall data, then the rainfall distribution live image is subjected to gray level processing, and the gray level image is blocked according to the corresponding relation between the rainfall and the colors in the rainfall distribution image.
Then, according to the corresponding relation between the rainfall and the colors in the rainfall distribution diagram, the gray level picture is matrixed, namely each rainfall corresponds to one gray level value in the diagram.
After the matrixed picture is obtained, the picture is subjected to binarization processing according to the corresponding relation between the rainfall and the gray scale value, and a plurality of 0-1 rainfall pictures corresponding to different rainfall are formed.
And after the pictures are obtained, calculating the similarity of the rainfall process according to a similarity measurement method.
In this embodiment, if the rainfall process is more than two days, the binarized rainfall pictures are merged into one picture and presented in a multi-dimensional array mode. The merging is performed by taking a union set for each data.
In the above embodiment, the similarity calculation method may further include:
firstly, taking an intersection of the binarized current rainfall distribution map and the binarized historical rainfall distribution map, namely recording as 1 if the numerical values of the same positions are all 1, and otherwise recording as 0;
then, taking a union set of the binarized current rainfall distribution diagram and the binarized historical rainfall distribution diagram, wherein one numerical value at the same position is 1, namely 1;
counting the number of grid points with the value of 1 in the intersection and the union;
and dividing the number of the grid points with the numerical value of 1 in the intersection by the number of the grid points with the numerical value of 1 in the union set to obtain the similarity of the two processes.
In the step 3, a weight is given to each rainfall similarity, and the process of obtaining the comprehensive rainfall similarity further comprises the following steps:
when the single-day rainfall similarity is calculated, adopting a subjective and objective integrated weighting method:
selecting part of current rainfall picture sets as labeling samples to label data:
the similarity degree of each labeled sample and all samples in the historical rainfall picture library is scored;
and linearly weighting each rainfall visual characteristic index of the marked sample according to the random weight to obtain an evaluation result, calculating according to the evaluation result and the marking result to obtain an NDCG @ K value, and searching for the optimal weight by using a particle swarm algorithm to enable the average NDCG value of the whole marked sample to be maximum.
The NDCG Normalized broken Cumulative Gain (Normalized discrete Cumulative Gain) is an evaluation index that takes the return order into consideration. The value range [0,1] is larger, and the effect is better.
The method specifically comprises the following steps:
。
in a further embodiment, when calculating the similarity of rainfall for multiple days, a rainfall sequence search method based on a sliding window and weighted dynamic time warping is employed:
inputting a current rainfall visual characteristic sequence Q with the length of e and a historical rainfall visual characteristic sequence S with the length of f in a certain period, wherein e is smaller than f;
sliding a window with the size of e on the historical rainfall visual characteristic sequence S, wherein the sliding step length is 1, and taking the sequence in each window as a subsequence to obtain a sequence set;
obtaining the distance between the current rainfall visual characteristic sequence Q and each subsequence P in the sequence set by adopting a weighted dynamic time warping algorithm, wherein the distance between any two points in the sequence is calculated by using a single-day rainfall characteristic multivariate characteristic distance measurement method;
and sequencing all the subsequences according to the distance to obtain a specified number of subsequences with the minimum distance as a search result.
In a further embodiment, the process of the weighted dynamic time warping method is as follows:
assuming that the current rainfall visual characteristic is sequence C, and the historical rainfall visual characteristic of a certain period is sequence Q, their lengths are respectively β and γ, when β = γ, the euclidean distance can be used to measure the distance between the two sequences. When the two are not equal, a weighting algorithm is adopted to solve the problem.
Wherein the content of the first and second substances,w max represents all the weightswThe maximum value of (a) is generally 1;n mid is half of the maximum value of lambda-kappa (which is actually the length of the longer of the two sequences);gthe scale of the function with respect to lambda-kappa is determined by empirical constants that control the curvature, and is commonly used in the range of 0.01-0.06. Each element (λ, κ) in the lattice represents q λ And c κ Distance d (q) between these two points λ ，c κ ），d（x λ ,y κ ) Denotes x λ 、y κ The distance between two points. d is a radical of w Representing the weighted distance.
The present application further provides an apparatus comprising:
at least one processor; and
a memory communicatively coupled to at least one of the processors; wherein the memory stores instructions executable by the processor for execution by the processor to implement the method for similar visual characteristic-based rainfall forecasting according to any of the embodiments above.
Although the preferred embodiments of the present invention have been described in detail, the present invention is not limited to the details of the embodiments, and various equivalent modifications can be made within the technical spirit of the present invention, and the scope of the present invention is also within the scope of the present invention.
Claims (12)
1. The similar rainfall forecasting method based on the visual characteristics is characterized by comprising the following steps of:
s1, constructing historical rainfall picture library
S11, collecting historical rainfall data in a plurality of periods in a preset area, and drawing a plurality of groups of historical rainfall distribution maps based on the historical rainfall data;
s12, extracting historical rainfall visual features based on each group of historical rainfall distribution map; the historical rainfall visual characteristics comprise total rainfall, rainfall spatial distribution and rainfall center;
s2, constructing a current rainfall picture set, extracting the current rainfall visual characteristics,
s3, measuring the distance between the historical rainfall visual characteristics and the current rainfall visual characteristics, and calculating the rainfall similarity between the current rainfall visual characteristics and the historical rainfall visual characteristics in each period; and giving weight to each rainfall similarity to obtain comprehensive rainfall similarity, and outputting the comprehensive rainfall similarity in sequence.
2. The method for forecasting similar rainfall based on visual characteristics according to claim 1, wherein the step S11 is further to:
acquiring geographic data of a preset area and rainfall data of each rainfall station in each period; drawing a map of a predetermined area based on the geographic data and rasterizing; and filling color parameters for the rasterized map based on the rainfall data to obtain a historical rainfall distribution diagram.
3. The method for forecasting similar rainfall based on visual characteristics according to claim 2, wherein said step S12 is further comprising:
reading the historical rainfall distribution graph, converting the historical rainfall distribution graph into a gray-scale image, setting a numerical value interval of the gray-scale image according to the rainfall level, and converting the gray-scale image into an NxN gray-scale image matrix;
and binarizing the gray level picture matrix according to the rainfall level to obtain a binarized rainfall matrix corresponding to different rainfall levels.
4. The visual-feature-based rainfall simulation forecasting method of claim 2, wherein the total rainfall amountRTComprises the following steps:
；
in the formula (I), the compound is shown in the specification,kthe number of colors in each historical rainfall distribution graph is obtained;p（C i ）is as followsiThe frequency with which the seed color appears is,v（C i ）is a firstiThe rainfall value corresponding to each color.
5. The visual feature-based rainfall simulation forecasting method of claim 4, wherein the rainfall spatial distributionRComprises the following steps:
；
m and n are the number of rows and columns of each historical rainfall distribution graph after being partitioned;r m×n the rainfall value of the block corresponding to the mth row and the nth column is obtained.
6. The method of claim 5, wherein the rainfall center is a similar rainfall forecast based on visual characteristicsC=（C x ，C y ）；
In the formula (I), the compound is shown in the specification,
，r ab the rainfall values of the a-th row and the b-th column are m and n are the row number and the column number of each historical rainfall distribution graph after being partitioned.
7. The method for forecasting similar rainfall based on visual characteristics according to claim 6, wherein the step S3 of calculating the rainfall similarity between the visual characteristics of the current rainfall and the visual characteristics of the historical rainfall at each period further comprises:
step S31, using Minkowski distanced1Calculating the total rainfall similarity of the preset basin, d1=RT；
Step S32, calculating the rainfall center similarity of the preset watershed by adopting the Euclidean distance d2,
step S33, adopting the partitioned area rainfall approximation to represent the rainfall spatial distribution approximation:
step S330, determining whether two blocks in the two rainfall spatial distributions R are similar;
respectively reading rainfall values of the two blocks, wherein the larger value is max, the smaller value is min, the rainfall difference ratio is defined as (max-min)/max, the threshold value of the rainfall difference ratio is alpha, and when the rainfall difference ratio is smaller than or equal to the threshold value, the two blocks are similar;
step S331, calculating logarithms of similar blocks in the two rainfall spatial distributions R;
repeating the step S330, calculating the similarity of all the blocks in the two rainfall spatial distributions R, and counting the number of the similar blocks;
step S332, calculating a rainfall spatial distribution approximation rate; and if each rainfall spatial distribution has c blocks, wherein d pairs of blocks in the two rainfall spatial distributions are similar blocks, the approximation rate of the two rainfall spatial distributions is (1-d/c).
8. The method for forecasting similar rainfall based on visual characteristics according to claim 7, wherein the weighting is given to each rainfall similarity in the step 3, and the process of obtaining the comprehensive rainfall similarity further comprises:
when the single-day rainfall similarity is calculated, adopting a subjective and objective integrated weighting method:
selecting part of current rainfall picture sets as labeling samples to label data:
the similarity degree of each labeled sample and all samples in the historical rainfall picture library is scored;
and linearly weighting each rainfall visual characteristic index of the marked sample according to the random weight to obtain an evaluation result, calculating to obtain an NDCG @ K value according to the evaluation result and the marking result, and finding the optimal weight by using a particle swarm algorithm to maximize the average NDCG value of the whole marked sample.
9. The method for forecasting the similar rainfall based on the visual characteristics according to claim 7, wherein when the similarity of the rainfall for a plurality of days is calculated, the following search method is adopted:
inputting a current rainfall visual characteristic sequence Q with the length of e and a historical rainfall visual characteristic sequence S with the length of f in a certain period, wherein e is smaller than f;
sliding a window with the size of e on the historical rainfall visual characteristic sequence S, wherein the sliding step length is 1, and taking the sequence in each window as a subsequence to obtain a sequence set;
obtaining the distance between the current rainfall visual characteristic sequence Q and each subsequence P in the sequence set by adopting a weighted dynamic time warping algorithm, wherein the distance between any two points in the sequence is calculated by using a single-day rainfall characteristic multivariate characteristic distance measurement method;
and sequencing all the subsequences according to the distance to obtain a specified number of subsequences with the minimum distance as a search result.
10. The method for forecasting the similar rainfall based on the visual characteristics of claim 7, further comprising:
s333, reading the similar blocks in the step S332, and respectively communicating the adjacent similar blocks of the two rainfall pictures to form at least one similar domain;
respectively judging whether the current historical rainfall distribution diagram and the rainfall centers of the historical rainfall distribution diagram are located in the similar domain;
if the current rainfall distribution diagram and the historical rainfall distribution diagram are larger than the threshold value, the historical rainfall distribution diagram is used as an output similar image alternative;
and S334, if the current historical rainfall distribution diagram and the rainfall centers of the historical rainfall distribution diagram are both located in the similar domain, obtaining the peripheral profiles of the similar domain, calculating the curve similarity of the two peripheral profile curves, and if the curve similarity is greater than a preset value, outputting the historical rainfall distribution diagram as a local similar alternative.
11. The method for forecasting similar rainfall based on visual characteristics according to claim 6, wherein the process of calculating the rainfall similarity between the current rainfall visual characteristics and the rainfall visual characteristics in each historical period further comprises:
acquiring water system characteristic data of a preset area, and dividing the preset area into a preset number of sub-partitions according to the water system characteristics;
respectively calculating the surface rainfall of each current sub-partition, combining the surface rainfall into a surface rainfall matrix, and further calculating the total rainfall of a preset area;
respectively calculating the surface rainfall of each sub-partition in each period of history based on the historical rainfall data and the water system characteristics, combining the surface rainfall into a surface rainfall matrix, and further calculating the total rainfall of a preset area;
and (4) calculating Euclidean distances of the current rainfall matrix and the rainfall matrix of each historical period, taking the Euclidean distances as similarity numerical values and arranging the similarity numerical values in a descending order.
12. An apparatus, comprising:
at least one processor; and
a memory communicatively coupled to at least one of the processors; wherein the content of the first and second substances,
the memory stores instructions executable by the processor for implementing the method for visual feature-based rain alike forecast of any one of claims 1-10.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210909459.8A CN114997534B (en) | 2022-07-29 | 2022-07-29 | Similar rainfall forecasting method and equipment based on visual features |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210909459.8A CN114997534B (en) | 2022-07-29 | 2022-07-29 | Similar rainfall forecasting method and equipment based on visual features |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114997534A true CN114997534A (en) | 2022-09-02 |
| CN114997534B CN114997534B (en) | 2022-10-21 |
Family
ID=83022558
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210909459.8A Active CN114997534B (en) | 2022-07-29 | 2022-07-29 | Similar rainfall forecasting method and equipment based on visual features |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114997534B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115271255A (en) * | 2022-09-19 | 2022-11-01 | 长江水利委员会水文局 | Rainfall flood similarity analysis method and system based on knowledge graph and machine learning |
| CN116010795A (en) * | 2023-03-17 | 2023-04-25 | 河海大学 | Similar scene rainfall pattern library construction method based on image features and deep learning |
| CN116091801A (en) * | 2023-03-07 | 2023-05-09 | 河海大学 | Rainfall image similarity searching method based on deep learning |
| CN116701553A (en) * | 2023-05-12 | 2023-09-05 | 中国长江电力股份有限公司 | Similar rainfall runoff process searching method based on rainfall time distribution histogram |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110852526A (en) * | 2019-11-19 | 2020-02-28 | 南京中禹智慧水利研究院有限公司 | Real-time flood forecasting method based on rainfall flood process similarity judgment |
| CN111008259A (en) * | 2019-12-06 | 2020-04-14 | 中国水利水电科学研究院 | River basin rainfall similarity searching method |
| CN111832810A (en) * | 2020-06-23 | 2020-10-27 | 西安电子科技大学 | Flood forecasting model, information processing method, storage medium, and computer device |
| CN111948739A (en) * | 2020-07-21 | 2020-11-17 | 重庆交通大学 | Flood forecasting system |
| CN112508237A (en) * | 2020-11-20 | 2021-03-16 | 北京师范大学 | Rain type region division method based on data analysis and real-time rain type prediction method |
| CN112785053A (en) * | 2021-01-15 | 2021-05-11 | 北京市水科学技术研究院 | Method and system for forecasting urban drainage basin flood |
| CN112801342A (en) * | 2020-12-31 | 2021-05-14 | 国电大渡河流域水电开发有限公司 | Adaptive runoff forecasting method based on rainfall runoff similarity |
| JP2021076565A (en) * | 2019-11-13 | 2021-05-20 | 良悦 土門 | Method for obtaining high-accuracy rainfall prediction value up to several-hours ahead in dam basin based on correlation of interpolated (interpolation) value of rainfall prediction value by low (middle) resolution model and rainfall prediction value (including interpolated value) waveform of high resolution model |
| US20210263187A1 (en) * | 2018-06-01 | 2021-08-26 | Nippon Telegraph And Telephone Corporation | Information presenting method, information presenting device, and information presenting program |
| CN113742327A (en) * | 2021-09-02 | 2021-12-03 | 中国水利水电科学研究院 | Automatic rainfall station abnormal value screening method based on rain-measuring radar data |
| CN113779113A (en) * | 2021-11-12 | 2021-12-10 | 水利部交通运输部国家能源局南京水利科学研究院 | Flood dynamic estimation method and system based on rainfall flood space-time process similarity excavation |
| CN114330102A (en) * | 2021-12-07 | 2022-04-12 | 武汉大学 | Rapid flood forecasting method and device based on rainfall similarity and intelligent model parameter adaptation |
| CN114706900A (en) * | 2022-02-17 | 2022-07-05 | 沈阳工业大学 | Precipitation similarity forecasting method based on image feature combination |
-
2022
- 2022-07-29 CN CN202210909459.8A patent/CN114997534B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210263187A1 (en) * | 2018-06-01 | 2021-08-26 | Nippon Telegraph And Telephone Corporation | Information presenting method, information presenting device, and information presenting program |
| JP2021076565A (en) * | 2019-11-13 | 2021-05-20 | 良悦 土門 | Method for obtaining high-accuracy rainfall prediction value up to several-hours ahead in dam basin based on correlation of interpolated (interpolation) value of rainfall prediction value by low (middle) resolution model and rainfall prediction value (including interpolated value) waveform of high resolution model |
| CN110852526A (en) * | 2019-11-19 | 2020-02-28 | 南京中禹智慧水利研究院有限公司 | Real-time flood forecasting method based on rainfall flood process similarity judgment |
| CN111008259A (en) * | 2019-12-06 | 2020-04-14 | 中国水利水电科学研究院 | River basin rainfall similarity searching method |
| CN111832810A (en) * | 2020-06-23 | 2020-10-27 | 西安电子科技大学 | Flood forecasting model, information processing method, storage medium, and computer device |
| CN111948739A (en) * | 2020-07-21 | 2020-11-17 | 重庆交通大学 | Flood forecasting system |
| CN112508237A (en) * | 2020-11-20 | 2021-03-16 | 北京师范大学 | Rain type region division method based on data analysis and real-time rain type prediction method |
| CN112801342A (en) * | 2020-12-31 | 2021-05-14 | 国电大渡河流域水电开发有限公司 | Adaptive runoff forecasting method based on rainfall runoff similarity |
| CN112785053A (en) * | 2021-01-15 | 2021-05-11 | 北京市水科学技术研究院 | Method and system for forecasting urban drainage basin flood |
| CN113742327A (en) * | 2021-09-02 | 2021-12-03 | 中国水利水电科学研究院 | Automatic rainfall station abnormal value screening method based on rain-measuring radar data |
| CN113779113A (en) * | 2021-11-12 | 2021-12-10 | 水利部交通运输部国家能源局南京水利科学研究院 | Flood dynamic estimation method and system based on rainfall flood space-time process similarity excavation |
| CN114330102A (en) * | 2021-12-07 | 2022-04-12 | 武汉大学 | Rapid flood forecasting method and device based on rainfall similarity and intelligent model parameter adaptation |
| CN114706900A (en) * | 2022-02-17 | 2022-07-05 | 沈阳工业大学 | Precipitation similarity forecasting method based on image feature combination |
Non-Patent Citations (3)
| Title |
|---|
| L.BROCCA: "Distributed rainfall-runoff modeling for flood frequency estimation and flood forecasting", 《HYDROLOGICAL PROCESSES》 * |
| 常远勇等: "基于TRMM 3B42数据的1998-2010年中国暴雨时空特征分析", 《水资源与水工程学报》 * |
| 王丽娟: "基于大数据分析方法的汉江流域安康段洪水预报研究", 《中国博士学位论文全文数据库 基础科学辑》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115271255A (en) * | 2022-09-19 | 2022-11-01 | 长江水利委员会水文局 | Rainfall flood similarity analysis method and system based on knowledge graph and machine learning |
| CN116091801A (en) * | 2023-03-07 | 2023-05-09 | 河海大学 | Rainfall image similarity searching method based on deep learning |
| CN116091801B (en) * | 2023-03-07 | 2023-06-16 | 河海大学 | Rainfall image similarity searching method based on deep learning |
| CN116010795A (en) * | 2023-03-17 | 2023-04-25 | 河海大学 | Similar scene rainfall pattern library construction method based on image features and deep learning |
| CN116701553A (en) * | 2023-05-12 | 2023-09-05 | 中国长江电力股份有限公司 | Similar rainfall runoff process searching method based on rainfall time distribution histogram |
| CN116701553B (en) * | 2023-05-12 | 2024-01-02 | 中国长江电力股份有限公司 | Similar rainfall runoff process searching method based on rainfall time distribution histogram |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114997534B (en) | 2022-10-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114997534B (en) | Similar rainfall forecasting method and equipment based on visual features | |
| Veregin | Error modeling for the map overlay operation | |
| CN112506990B (en) | Hydrological data anomaly detection method based on spatiotemporal information | |
| CN111651545A (en) | Urban marginal area extraction method based on multi-source data fusion | |
| CN104899437A (en) | Early-warning method for heavy-rainfall type landslide hazard | |
| WO2020107712A1 (en) | Method for implementing antenna azimuth correction based on user data | |
| WO2020043027A1 (en) | Method for optimizing boundaries of region when predicting atmospheric pollution | |
| CN112308292A (en) | Method for drawing fire risk grade distribution map | |
| WO2013009160A2 (en) | A geometric method for predicting landslide disaste | |
| CN116363601B (en) | Data acquisition and analysis method and system for pollution monitoring equipment | |
| CN111080356A (en) | Method for calculating residence price influence factors by using machine learning regression model | |
| CN114254802B (en) | Prediction method for vegetation coverage space-time change under climate change drive | |
| CN112907113B (en) | Vegetation change cause identification method considering spatial correlation | |
| CN105488316A (en) | Air quality prediction system and method | |
| CN113632101B (en) | Method for predicting atmospheric pollution through vectorization analysis | |
| CN114444791A (en) | Flood disaster remote sensing monitoring and evaluation method based on machine learning | |
| Cappelaere et al. | Hydrologic process simulation of a semiarid, endoreic catchment in Sahelian West Niger. 2. Model calibration and uncertainty characterization | |
| CN108764527B (en) | Screening method for soil organic carbon library time-space dynamic prediction optimal environment variables | |
| CN113836808A (en) | PM2.5 deep learning prediction method based on heavy pollution feature constraint | |
| CN117171533B (en) | Real-time acquisition and processing method and system for geographical mapping operation data | |
| CN109190783B (en) | Urban water network leakage space aggregation detection and key influence factor identification method | |
| Lu et al. | Uncertainty quantification of machine learning models to improve streamflow prediction under changing climate and environmental conditions | |
| Zahraie et al. | SST clustering for winter precipitation prediction in southeast of Iran: Comparison between modified K-means and genetic algorithm-based clustering methods | |
| Noor et al. | Prediction map of rainfall classification using random forest and inverse distance weighted (IDW) | |
| CN116384627A (en) | Geological disaster evaluation method based on machine learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |