CN112949726B - ISCCP cloud classification method, system, medium and terminal based on FY-4A satellite - Google Patents
ISCCP cloud classification method, system, medium and terminal based on FY-4A satellite Download PDFInfo
- Publication number
- CN112949726B CN112949726B CN202110263471.1A CN202110263471A CN112949726B CN 112949726 B CN112949726 B CN 112949726B CN 202110263471 A CN202110263471 A CN 202110263471A CN 112949726 B CN112949726 B CN 112949726B
- Authority
- CN
- China
- Prior art keywords
- channel
- cloud
- isccp
- latitude
- satellite
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000012545 processing Methods 0.000 claims abstract description 56
- 238000010606 normalization Methods 0.000 claims abstract description 51
- 238000013145 classification model Methods 0.000 claims abstract description 34
- 230000009466 transformation Effects 0.000 claims abstract description 20
- 238000004590 computer program Methods 0.000 claims description 9
- 238000010276 construction Methods 0.000 claims description 9
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Chemical compound O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 8
- 238000011160 research Methods 0.000 abstract description 8
- 238000010586 diagram Methods 0.000 description 19
- 241000208818 Helianthus Species 0.000 description 18
- 235000003222 Helianthus annuus Nutrition 0.000 description 18
- 238000004422 calculation algorithm Methods 0.000 description 18
- 238000004458 analytical method Methods 0.000 description 7
- 238000011161 development Methods 0.000 description 6
- 230000018109 developmental process Effects 0.000 description 6
- 230000003595 spectral effect Effects 0.000 description 5
- 238000012360 testing method Methods 0.000 description 5
- 238000012549 training Methods 0.000 description 5
- 238000007476 Maximum Likelihood Methods 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000003491 array Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000001556 precipitation Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010224 classification analysis Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 238000002310 reflectometry Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Processing (AREA)
Abstract
The invention provides a ISCCP cloud classification method, a system, a medium and a terminal based on an FY-4A satellite, which comprise the following steps: performing projection transformation on the primary data and the primary geographic positioning data of the FY-4A satellite respectively to obtain equal-longitude and latitude primary data and equal-longitude and latitude primary geographic positioning data; carrying out visible light channel normalization processing based on the equal longitude and latitude primary data and the equal longitude and latitude primary geographic positioning data; constructing cloud features based on the normalized visible light channel and the equal longitude and latitude primary data; carrying out standardization processing on the cloud characteristics; based on the cloud characteristics of the standardized processing, a trained ISCCP cloud classification model is adopted to obtain ISCCP cloud classification results. The ISCCP cloud classification method, system, medium and terminal based on the FY-4A satellite realize effective identification of ISCCP cloud types and provide information support for research of weather systems.
Description
Technical Field
The invention relates to a classification method, in particular to a ISCCP cloud classification method, a system, a medium and a terminal based on FY-4A satellites.
Background
Cloud is an important carrier for the energy and water circulation of the earth gas system, and plays an important role in global radiant energy balance and water vapor transmission. The cloud type has clear indication significance for the development change of the weather system, for example, the rain clouds are one of important factors for weather identification and forecasting. Therefore, the accurate cloud classification is of great significance to research and application in the fields of hydrology, weather climate and the like. At present, the international cloud classification principle is mainly based on the characteristics of cloud appearance, height and the like, and development and internal structure of the cloud are properly combined. In the national ground meteorological observation standard, clouds are divided into 3 groups, 10 groups and 29 groups. Wherein group 3 includes high cloud, medium cloud, and low cloud. The high cloud group comprises volume cloud, volume cloud and volume layer cloud 3, the middle cloud group comprises high volume cloud and high layer cloud 2, and the low cloud group comprises rain layer cloud, precipitation cloud and precipitation cloud 5. International satellite cloud climate program (International SatelliteCloud Climatology Project, ISCCP) classifies clouds into 9 categories, namely, coil, deep convection, high cloud, rain, cloud, layer and layer cloud, according to cloud top height and optical thickness of the cloud. Research shows that the classification standard has consistency in distribution in China. However, there is currently no work related to ISCCP cloud classification based on the new generation of stationary weather-wind cloud satellite number four (FY-4A). Therefore, it is necessary to conduct a ISCCP cloud classification method for weather satellites of the wind cloud No. four.
Because the reflectivity of different kinds of clouds in the spectrum band and the characteristic medium of the underlying surface, the brightness temperature of the radiation of the cloud top and other factors have great difference, the satellite cloud image is an important means for constructing a cloud identification algorithm. At present, research work for performing cloud classification analysis according to different cloud characteristics based on multispectral satellite images has been continued for many years. The cloud classifier may use only one or more features to classify. These features overlap in most cases in different cloud classes. Features are largely classified as spectral, texture, contextual or background features. Spectral features describe the average color and hue variation of the cloud. Spectral functions support the division of most cloud classes, but they do not provide information about the color space distribution, because in certain cloud types (like clouds, high clouds and layer clouds) the average color values are similar but the spatial distribution of colors is different. The texture features refer to the spatial statistical distribution of the hue variation of the pixels, and the gray co-occurrence matrix can be used to classify the spatial difference texture features. Contextual or background features refer to the geographic location of pixels in an image, the viewing angle of the instrument, the time and date of the year, and specific local information. Methods of cloud classification can be broadly classified into the following three categories according to principle and computational complexity:
1. Simple algorithm
A simple algorithm represents the processing of cloud features using linear mathematical relationships. The method has low calculation resources required, but has low cloud classification precision, and specifically comprises the following three algorithms:
(1) Minimum distance method
The minimum distance method proposed by Richards and Jia shows that for each point pixel on the image, its distance from the average is calculated and its class is assigned to the class with the smallest distance. The method has higher operation speed but lower precision, so the method is often used for an initial clustering algorithm.
(2) Threshold method
Wang and Sassen propose a thresholding-based parallel classification method. The method belongs to a classification method based on boundary rules, and classification decision parameters of the classification method form an N-dimensional data space. As shown in fig. 1, the upper and lower limits of each dimension of data are determined by the standard deviation of the mean of the feature. When the space where the feature is located is one-dimensional, two-dimensional and three-dimensional, the classification decision parameters of the feature can respectively form line segments, rectangles and parallelepipeds. The one-dimensional feature is simple but the accuracy is low. Yu Fan and Chen Weimin are two-dimensional representations of features.
(3) Proximity algorithm
The K proximity algorithm proposed by Bankert and Wade is mainly used for scenes with less prior knowledge on the distribution of cloud class, and is commonly used for an initial clustering stage of unsupervised learning. By giving a new input instance, using a certain distance measurement method, finding K instances nearest to the instance in the training set, and finally, attributing the cloud class of the input instance to most of the K instances.
2. Mathematical statistics algorithm
With the improvement of the computing capability of a computer, a pattern recognition algorithm based on mathematical statistics and a probability model is introduced into a remote sensing image for cloud recognition. Compared with a simple algorithm, the mathematical statistics algorithm is complex in calculation, but high in accuracy, and specifically comprises the following three algorithms:
(1) Linear discriminant analysis method
Linear discriminant analysis (LINEAR DISCRIMINANT ANALYSIS, LDA) is a supervised statistical technique and Amato proposes combining different cloud features into linear combinations for classification by distinguishing independent features from dependent features of the input data. Shi proposes that a more accurate threshold boundary can be obtained than LDA by using quadratic discriminant analysis (Quadratic DISCRIMINANT ANALYSIS, QDA), and can be used for classifying multi-feature input data. Subsequently Marais proposes a discriminant analysis of the heteroscedastic (Heteroscedastic DISCRIMINANT ANALYSIS, HDA) based on LDA, which combines the visible and near infrared channels into one gray scale image, which is then used for linear classification.
(2) Naive Bayes classifier
A naive bayes classifier is a probabilistic classifier based on bayes hypotheses. And under the prior probability condition of the selected independent features, calculating posterior probability of the cloud class to which the cloud class belongs according to a Bayesian formula, and selecting the maximum probability to obtain a cloud class decision. The method requires a smaller number of samples to estimate the classification parameters, and the independent features mean less computation. However, for satellite data, it is not true that the input features are assumed to be independent.
(3) Maximum likelihood classification method
The maximum likelihood classification method assumes that the input features accord with normal distribution, and the cloud class is attributed to the class with the maximum posterior probability by utilizing the similarity of the distribution situation of the training sample and the prior sample. Compared with the naive Bayes method which requires the input features to have independent characteristics, the maximum likelihood method considers the similarity of sample feature distribution situations, and the maximum likelihood method is more practical.
3. Artificial intelligence algorithm
With the rapid development of computer science, complex machine learning methods, such as fuzzy logic algorithms, artificial neural networks, self-organizing mapping networks, support vector machines, random forests and other algorithms, are verified to be successfully applied to cloud classification tasks. In recent years, with the development of deep learning algorithms in computer vision technology, semantic features having representative and discrimination lines can be learned from data in layers, and great achievement is achieved in the fields of image recognition, signal processing, computer vision and the like. Meanwhile, the deep learning technology is also introduced into the related task of the satellite remote sensing image, the Shi uses a convolutional neural network (Convolutional Neural Networks, CNN) to carry out cloud detection on the satellite image, cai and Wang divide the local cloud image of a wind cloud No. two meteorological satellite (FY-2C) into low cloud, medium cloud, high cloud and direct expansion cloud categories by using the CNN, and compared with the traditional machine learning algorithm, the result is more accurate in precision, but the method is only suitable for identifying the whole Zhang Yuntu.
Disclosure of Invention
In view of the shortcomings of the prior art, the invention aims to provide a ISCCP cloud classification method, a system, a medium and a terminal based on an FY-4A satellite, which realize effective identification of ISCCP cloud types and provide information support for research of weather systems.
To achieve the above and other related objects, the present invention provides a ISCCP cloud classification method based on an FY-4A satellite, comprising the steps of: performing projection transformation on the primary data and the primary geographic positioning data of the FY-4A satellite respectively to obtain equal-longitude and latitude primary data and equal-longitude and latitude primary geographic positioning data; carrying out visible light channel normalization processing based on the equal longitude and latitude primary data and the equal longitude and latitude primary geographic positioning data; constructing cloud features based on the normalized visible light channel and the equal longitude and latitude primary data; carrying out standardization processing on the cloud characteristics; based on the cloud characteristics of the standardized processing, a trained ISCCP cloud classification model is adopted to obtain ISCCP cloud classification results.
In an embodiment of the present invention, the visible light channel normalization processing based on the equal longitude and latitude primary data and the equal longitude and latitude primary geographic positioning data includes the following steps:
calculating normalization operators Wherein,θ1、θ2、/>Respectively representing the zenith angle of the sun, the zenith angle of the satellite and the azimuth angle of the satellite relative to the sun;
The normalized visible light albedo A 0＝As/F is calculated, wherein A s represents the visible light albedo before normalization.
In an embodiment of the present invention, the cloud feature includes:
The visible light near infrared wave bands of the B01 channel, the B02 channel, the B03 channel, the B04 channel, the B05 channel and the B06 channel;
the middle wave infrared wave bands of the B07 channel and the B08 channel;
a water vapor channel of the B09 channel and the B10 channel;
the long-wave infrared wave bands of the B11 channel, the B13 channel and the B14 channel;
The bright temperature difference between the B09 channel and the B12 channel, the bright temperature difference between the B10 channel and the B13 channel, the bright temperature difference between the B11 channel and the B12 channel, the bright temperature difference between the B12 channel and the B13 channel, the bright temperature difference between the B08 channel and the B12 channel and the bright temperature difference between the B08 channel and the B10 channel;
And a B12 channel.
In an embodiment of the present invention, the ISCCP cloud classification model adopts a unet++ network structure.
Correspondingly, the invention provides a ISCCP cloud classification system based on an FY-4A satellite, which comprises a transformation module, a normalization module, a construction module, a standardization processing module and a classification module;
the transformation module is used for respectively carrying out projection transformation on the primary data and the primary geographic positioning data of the FY-4A satellite to obtain equal-longitude and latitude primary data and equal-longitude and latitude primary geographic positioning data;
The normalization module is used for performing visible light channel normalization processing based on the equal-longitude-latitude primary data and the equal-longitude-latitude primary geographic positioning data;
The construction module is used for constructing cloud characteristics based on the visible light channel after normalization processing and the equal longitude and latitude primary data;
the standardized processing module is used for carrying out standardized processing on the cloud characteristics;
the classification module is used for acquiring ISCCP cloud classification results by adopting a trained ISCCP cloud classification model based on cloud characteristics of standardized processing.
In an embodiment of the present invention, the normalization module performs visible light channel normalization processing based on the first-level data of equal longitude and latitude and the first-level geographic positioning data of equal longitude and latitude, and includes the following steps:
calculating normalization operators Wherein,θ1、θ2、/>Respectively representing the zenith angle of the sun, the zenith angle of the satellite and the azimuth angle of the satellite relative to the sun;
The normalized visible light albedo A 0＝As/F is calculated, wherein A s represents the visible light albedo before normalization.
In an embodiment of the present invention, the cloud feature includes:
The visible light near infrared wave bands of the B01 channel, the B02 channel, the B03 channel, the B04 channel, the B05 channel and the B06 channel; the middle wave infrared wave bands of the B07 channel and the B08 channel;
a water vapor channel of the B09 channel and the B10 channel;
the long-wave infrared wave bands of the B11 channel, the B13 channel and the B14 channel;
The bright temperature difference between the B09 channel and the B12 channel, the bright temperature difference between the B10 channel and the B13 channel, the bright temperature difference between the B11 channel and the B12 channel, the bright temperature difference between the B12 channel and the B13 channel, the bright temperature difference between the B08 channel and the B12 channel and the bright temperature difference between the B08 channel and the B10 channel;
And a B12 channel.
In an embodiment of the present invention, the ISCCP cloud classification model adopts a unet++ network structure.
The invention provides a storage medium, on which a computer program is stored, which when being executed by a processor, implements the above-mentioned ISCCP cloud classification method based on FY-4A satellites.
Finally, the present invention provides a terminal comprising: a processor and a memory;
the memory is used for storing a computer program;
The processor is used for executing the computer program stored in the memory, so that the terminal executes the ISCCP cloud classification method based on the FY-4A satellite.
As described above, the ISCCP cloud classification method, system, medium and terminal based on the FY-4A satellite have the following beneficial effects:
(1) ISCCP cloud classification is realized based on a segmentation network, so that the accuracy is high;
(2) Through accurate ISCCP cloud classification, information support is provided for the research of weather systems, and the weather system cloud classification method has high practicability.
Drawings
FIG. 1 is a schematic diagram showing the relationship between the upper and lower limits of each dimension of data and the standard deviation of the mean of the features in the prior art thresholding-based parallel classification method;
FIG. 2 is a flow chart of a method for ISCCP cloud classification based on FY-4A satellites according to one embodiment of the invention;
Fig. 3 shows a diagram of ISCCP cloud classification for sunflower satellite number 8 at time 0130 (UTC) of month 6 of 2020;
fig. 4 shows a schematic diagram of comparison of visible light channel normalization before and after the time of year 2020, month 6, day 3 and day 0130 (UTC);
fig. 5 shows a schematic diagram of ISCCP cloud classification obtained based on ISCCP cloud classification model at time 0130 (UTC) of month 6 of 2020;
fig. 6 (a) is a schematic diagram showing comparison of a ISCCP cloud classification result obtained based on ISCCP cloud classification model at the time of 0130 (UTC) of month 6 of 2020 with a skill score TS of a sunflower satellite cloud classification result;
fig. 6b is a schematic diagram showing a hit ratio POD comparison of a ISCCP cloud classification result obtained based on a ISCCP cloud classification model at a time of 0130 (UTC) of month 6 of 2020 with a sunflower satellite cloud classification result of No. 8;
fig. 6 (c) is a schematic diagram showing the comparison of the false alarm rate FAR of the ISCCP cloud classification result obtained based on the ISCCP cloud classification model at the time of 0130 (UTC) of month 6 of 2020 and the sunflower satellite cloud classification result;
fig. 6 (d) is a schematic diagram showing a report missing rate MAR comparison of a ISCCP cloud classification result obtained based on ISCCP cloud classification model at the time of 0130 (UTC) of month 6 of 2020 with a sunflower satellite cloud classification result of No. 8;
fig. 7 shows a diagram of ISCCP cloud classification for sunflower satellite number 8 at time 0530 (UTC) of month 6 of 2020;
Fig. 8 shows a schematic diagram of comparison of visible light channel normalization before and after the time of 0530 (UTC) at 6/7/2020;
fig. 9 shows a schematic diagram of ISCCP cloud classification obtained based on ISCCP cloud classification model at time of 0530 (UTC) of month 6 of 2020;
fig. 10 (a) is a schematic diagram showing comparison of the skill score TS of ISCCP cloud classification result obtained based on ISCCP cloud classification model at time of 0530 (UTC) of month 6 of 2020 with the skill score TS of sunflower satellite cloud classification result of satellite 8;
fig. 10 (b) is a schematic diagram showing hit ratio POD comparison between ISCCP cloud classification results obtained based on ISCCP cloud classification model at time of 0530 (UTC) of month 6 and 7 of 2020 and sunflower satellite cloud classification results;
Fig. 10 (c) is a schematic diagram showing the comparison of the false alarm rate FAR of the ISCCP cloud classification result obtained based on the ISCCP cloud classification model at the time of 0530 (UTC) of month 6 of 2020 and the sunflower satellite cloud classification result;
Fig. 10 (d) is a schematic diagram showing a report missing rate MAR comparison of a ISCCP cloud classification result obtained based on ISCCP cloud classification model at the time of 0530 (UTC) of month 6 of 2020 with a sunflower satellite cloud classification result of No. 8;
FIG. 11 is a schematic diagram of a ISCCP cloud classification system based on FY-4A satellites according to one embodiment of the present invention;
fig. 12 is a schematic structural diagram of a terminal according to an embodiment of the invention.
Description of element reference numerals
111. Conversion module
112. Normalization module
113. Construction module
114. Standardized processing module
115. Classification module
121. Processor and method for controlling the same
122. Memory device
Detailed Description
Other advantages and effects of the present invention will become apparent to those skilled in the art from the following disclosure, which describes the embodiments of the present invention with reference to specific examples. The invention may be practiced or carried out in other embodiments that depart from the specific details, and the details of the present description may be modified or varied from the spirit and scope of the present invention.
It should be noted that, the illustrations provided in the present embodiment merely illustrate the basic concept of the present invention by way of illustration, and only the components related to the present invention are shown in the drawings and are not drawn according to the number, shape and size of the components in actual implementation, and the form, number and proportion of the components in actual implementation may be arbitrarily changed, and the layout of the components may be more complex.
According to the ISCCP cloud classification method, system, medium and terminal based on the FY-4A satellite, through the data collected based on the FY-4A satellite and the ISCCP cloud classification model, the ISCCP cloud type is effectively identified, the accuracy is high, and information support is provided for the research of a weather system.
As shown in fig. 2, in an embodiment, the method for ISCCP cloud classification based on FY-4A satellites of the present invention comprises the following steps:
And S1, respectively performing projection transformation on the primary data and the primary geographic positioning data of the FY-4A satellite to obtain equal-longitude and latitude primary data and equal-longitude and latitude primary geographic positioning data.
Specifically, projection transformation is carried out on the primary data of the FY-4A satellite to obtain equal longitude and latitude primary data. The first-level data of warp and weft comprise visible light albedo, infrared brightness temperature and other information.
And performing projective transformation on the primary geographic positioning data of the FY-4A satellite to obtain equal longitude and latitude primary geographic positioning data.
And S2, carrying out visible light channel normalization processing based on the equal-longitude and latitude primary data and the equal-longitude and latitude primary geographic positioning data.
Specifically, the normalization processing of the visible light channel is performed based on the visible light albedo in the equal-longitude and latitude primary data and the geographic position information in the equal-longitude and latitude primary geographic positioning data.
In an embodiment of the present invention, a Quasi-lambertian surface adjustment (Quasi-Lambertian Surface Adjustment, QLSA) algorithm is adopted to normalize a visible light channel, and the method specifically includes the following steps:
21 Calculating normalization operators Wherein,θ1、θ2、/>Respectively representing the zenith angle of the sun, the zenith angle of the satellite and the azimuth angle of the satellite relative to the sun.
22 Calculate normalized visible light albedo a 0＝As/F, where a s represents the visible light albedo before normalization.
And S3, constructing cloud features based on the visible light channel after normalization processing and the equal longitude and latitude primary data.
Because the FY-4A satellite is provided with the 14-channel imager, the range of the observation wave band covers the visible light to the long-wave infrared wave band, and the device can be used for observing the characteristics of different targets, and the physical characteristics of the cloud can be effectively reflected by combining different wave bands under certain conditions.
In an embodiment of the present invention, the cloud feature includes:
① The six visible light near infrared bands B01, B02, B03, B04, B05 and B06 can reflect the optical thickness characteristics of the cloud;
② The bright temperature difference between the B09 channel and the B12 channel and between the B10 channel and the B13 channel can reflect the top height information of the cloud;
③ The B12 channel can reflect cloud top temperature information of the cloud;
④ The bright temperature difference between the B11 and the B12 and between the B12 and the B13 channels can reflect the property of the cloud phase state;
⑤ The bright temperature difference between the B08 and the B12 and between the B08 and the B10 channels reflects the characteristic of the cloud water path;
⑥ The B07 and B08 medium wave infrared, B09, B10 water vapor channels and B11, B13 and B14 long wave infrared can reflect the physical characteristics of the cloud to a certain extent.
And S4, carrying out standardization processing on the cloud characteristics.
Specifically, maximum value and minimum value normalization processing is performed on the cloud characteristics, so that standardization is achieved.
And S5, acquiring ISCCP cloud classification results by adopting a trained ISCCP cloud classification model based on the cloud characteristics of the standardized processing.
Specifically, the cloud characteristics subjected to standardization processing are used as the input of a trained ISCCP cloud classification model, and ISCCP cloud classification results can be output through the trained ISCCP cloud classification model.
In an embodiment of the present invention, the normalized cloud feature data is divided into a training set, a test set and a verification set. And constructing a series of 383×384 data blocks from the cloud characteristic data pool subjected to standardized processing, and writing the data blocks into TFRecord files in a data stream mode. In the model training and testing process, TFRecord files are loaded into memory blocks with specified sizes, and data are read in according to specified sequences, so that training and testing of the ISCCP cloud classification model are achieved.
In an embodiment of the present invention, the ISCCP cloud classification model adopts a unet++ network structure.
The ISCCP cloud classification method based on the FY-4A satellite of the invention is further described by the following specific examples.
Example 1
The example selects cloud type distribution conditions in regions with east longitude of 80-139 degrees and north latitude of 5-54 degrees, wherein the time is 0130 (UTC) of 6 months of 2020. The sunflower satellite number 8 cloud classification is shown in fig. 3.
Firstly, projection transformation is respectively carried out by using first-level data and first-level geographic positioning data of FY-4A, and equal-longitude and latitude first-level data and equal-longitude and latitude first-level geographic positioning data containing visible light albedo and infrared brightness temperature are obtained.
And secondly, the visible light channel is normalized by using the quasi-Brownian adjustment method and the equal-longitude and latitude primary geographic positioning data, so that the available time of the visible light image can be effectively increased. As shown in fig. 4, the left side is the visible light albedo before normalization, and the right side is the visible light albedo after normalization. It is known that the data distribution of the albedo of the normalized visible light image in the left part area is obviously improved.
And combining the spectral channel characteristics of different wavebands to obtain characteristic diagrams of 20 channels, and further carrying out maximum value and minimum value normalization processing on the characteristics.
Finally, the constructed test data set is predicted by using the trained ISCCP cloud classification model, and the prediction results are spliced to obtain ISCCP cloud type at the moment shown in fig. 5. Relevant evaluation indexes of ISCCP cloud classification results and sunflower number 8 satellite cloud classification results obtained based on ISCCP cloud classification model at the moment are shown in table 1.
Table 1, 2020, 6/3/0130 (UTC) ISCCP cloud classification model prediction results and sunflower satellite 8 cloud classification result evaluation index results
| Kappa | Ci/TS | Cs/TS | Dc/TS | Ac/TS | As/TS | Ns/TS | Cu/TS | Sc/TS | St/Ts |
| 0.57 | 0.40 | 0.58 | 0.62 | 0.19 | 0.42 | 0.42 | 0.24 | 0.49 | 0.35 |
Wherein Kappa coefficient is used for measuring classification accuracy, TS represents skill score (Threat score), ci represents volume cloud, cs represents volume cloud, dc represents deep convection cloud, ac represents high-level cloud, as represents high-level cloud, ns represents rain layer cloud, cu represents volume cloud, sc represents layer cloud, st represents layer cloud.
As shown in fig. 6 (a) -6 (d), according to the comparison of the predicted result of the cloud classification model at the time of ISCCP of 3 rd month (UTC) of 2020 and the skill score TS, the hit rate POD, the false alarm rate FAR, and the false alarm rate MAR of the satellite cloud classification result of sunflower number 8, the skill score TS, the hit rate POD of the roll cloud, the deep convection cloud, the high-layer cloud, the rain cloud, and the layer cloud are relatively high in terms of the whole, wherein the recognition accuracy of the deep convection cloud is the highest, and the cloud of this type has important indication significance for the development change of the weather system.
Example two
The example chooses a cloud type distribution in the region with a time of 2020, 6,7, 0530 (UTC), east longitude of 80-139 degrees and north latitude of 5-54 degrees. The sunflower satellite number 8 cloud classification is shown in fig. 7.
Firstly, projection transformation is respectively carried out by using first-level data and first-level geographic positioning data of FY-4A, and equal-longitude and latitude first-level data and equal-longitude and latitude first-level geographic positioning data containing visible light albedo and infrared brightness temperature are obtained.
And secondly, normalizing the visible light channel by using the quasi-Brownian adjustment method and using equal-longitude and latitude first-level geographic positioning data. As shown in fig. 8, the left side is the visible light albedo before normalization, and the right side is the visible light albedo after normalization. It is known that normalization can effectively increase the available time of the visible light image.
And combining the spectral channel characteristics of different wavebands to obtain characteristic diagrams of 20 channels, and further carrying out maximum value and minimum value normalization processing on the characteristics.
Finally, the constructed test data set is predicted by using the trained ISCCP cloud classification model, and the prediction results are spliced to obtain ISCCP cloud type at the moment shown in fig. 9. Relevant evaluation indexes of ISCCP cloud classification results and sunflower number 8 satellite cloud classification results obtained based on ISCCP cloud classification model at the moment are shown in table 1.
Table 2, 2020, 6/7/0530 (UTC) ISCCP cloud classification model prediction results and sunflower No. 8 satellite cloud classification result evaluation index results
| Kappa | Ci/TS | Cs/TS | Dc/TS | Ac/TS | As/TS | Ns/TS | Cu/TS | Sc/TS | St/Ts |
| 0.60 | 0.45 | 0.68 | 0.65 | 0.19 | 0.36 | 0.44 | 0.28 | 0.34 | 0.28 |
As shown in fig. 10 (a) -10 (d), according to comparison of the predicted result of cloud classification model and the skill score TS, hit rate POD, false alarm rate FAR, false alarm rate MAR of sunflower satellite cloud 8 at time ISCCP (UTC) of year 6, month 7 of 2020, the skill score TS, hit rate POD of the cloud classification result are relatively high, and these types of clouds have important indication significance for development change of weather system.
As shown in fig. 11, in one embodiment, the ISCCP cloud classification system based on FY-4A satellites of the present invention includes a transformation module 111, a normalization module 112, a construction module 113, a normalization processing module 114, and a classification module 115.
The transformation module 111 is used for respectively performing projection transformation on the primary data and the primary geographic positioning data of the FY-4A satellite to obtain equal-longitude and latitude primary data and equal-longitude and latitude primary geographic positioning data.
The normalization module 112 is connected to the transformation module 111, and is configured to perform visible light channel normalization processing based on the first-level data of equal longitude and latitude and the first-level geographic positioning data of equal longitude and latitude.
The construction module 113 is connected to the transformation module 111 and the normalization module 112, and is configured to construct cloud features based on the normalized visible light channel and the equal longitude and latitude primary data.
The normalization processing module 114 is connected to the construction module 113, and is configured to perform normalization processing on the cloud feature.
The classification module 115 is connected to the normalization processing module 114, and is configured to obtain ISCCP cloud classification results based on the normalized cloud characteristics by using a trained ISCCP cloud classification model.
The structures and principles of the transformation module 111, the normalization module 112, the construction module 113, the normalization processing module 114 and the classification module 115 are in one-to-one correspondence with the steps in the ISCCP cloud classification method based on the FY-4A satellite, so that the description thereof will not be repeated here.
It should be noted that, it should be understood that the division of the modules of the above apparatus is merely a division of a logic function, and may be fully or partially integrated into a physical entity or may be physically separated. And these modules may all be implemented in software in the form of calls by the processing element; or can be realized in hardware; the method can also be realized in a form of calling software by a processing element, and the method can be realized in a form of hardware by a part of modules. For example, the x module may be a processing element that is set up separately, may be implemented in a chip of the apparatus, or may be stored in a memory of the apparatus in the form of program code, and the function of the x module may be called and executed by a processing element of the apparatus. The implementation of the other modules is similar. In addition, all or part of the modules can be integrated together or can be independently implemented. The processing element described herein may be an integrated circuit having signal processing capabilities. In implementation, each step of the above method or each module above may be implemented by an integrated logic circuit of hardware in a processor element or an instruction in a software form.
For example, the modules above may be one or more integrated circuits configured to implement the methods above, such as: one or more Application SPECIFIC INTEGRATED Circuits (ASIC), or one or more microprocessors (DIGITAL SINGNAL Processor DSP), or one or more field programmable gate arrays (FieldProgrammable GATE ARRAY FPGA), etc. For another example, when a module above is implemented in the form of a processing element scheduler code, the processing element may be a general-purpose processor, such as a central processing unit (Central Processing Unit, CPU) or other processor that may invoke the program code. For another example, the modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).
The storage medium of the present invention stores a computer program which, when executed by a processor, implements the above-described method of ISCCP cloud classification based on FY-4A satellites. The storage medium includes: various media capable of storing program codes, such as ROM, RAM, magnetic disk, U-disk, memory card, or optical disk.
As shown in fig. 12, in an embodiment, the terminal of the present invention includes: a processor 121 and a memory 122.
The memory 122 is used for storing a computer program.
The memory 122 includes: various media capable of storing program codes, such as ROM, RAM, magnetic disk, U-disk, memory card, or optical disk.
The processor 121 is connected to the memory 122, and is configured to execute a computer program stored in the memory 122, so that the terminal performs the above-mentioned ISCCP cloud classification method based on the FY-4A satellite.
Preferably, the processor 121 may be a general-purpose processor, including a central processing unit (Central Processing Unit, abbreviated as CPU), a network processor (Network Processor, abbreviated as NP), etc.; but also digital signal processors (DIGITALSIGNAL PROCESSOR, DSP for short), application specific integrated circuits (ASIC for short), field programmable gate arrays (Field Programmable GATE ARRAY, FPGA for short) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
In conclusion, the ISCCP cloud classification method, the system, the medium and the terminal based on the FY-4A satellite realize ISCCP cloud classification based on the segmentation network, and have high accuracy; through accurate ISCCP cloud classification, information support is provided for the research of weather systems, and the weather system cloud classification method has high practicability. Therefore, the invention effectively overcomes various defects in the prior art and has high industrial utilization value.
The above embodiments are merely illustrative of the principles of the present invention and its effectiveness, and are not intended to limit the invention. Modifications and variations may be made to the above-described embodiments by those skilled in the art without departing from the spirit and scope of the invention. Accordingly, it is intended that all equivalent modifications and variations of the invention be covered by the claims, which are within the ordinary skill of the art, be within the spirit and scope of the present disclosure.
Claims (8)
1. A ISCCP cloud classification method based on FY-4A satellites is characterized in that: the method comprises the following steps:
Performing projection transformation on the primary data and the primary geographic positioning data of the FY-4A satellite respectively to obtain equal-longitude and latitude primary data and equal-longitude and latitude primary geographic positioning data;
carrying out visible light channel normalization processing based on the equal longitude and latitude primary data and the equal longitude and latitude primary geographic positioning data;
constructing cloud features based on the normalized visible light channel and the equal longitude and latitude primary data;
Carrying out standardization processing on the cloud characteristics;
based on the cloud characteristics of the standardized processing, acquiring ISCCP cloud classification results by adopting a trained ISCCP cloud classification model;
the visible light channel normalization processing based on the equal longitude and latitude primary data and the equal longitude and latitude primary geographic positioning data comprises the following steps:
calculating normalization operators Wherein/> θ1、θ2、/>Respectively representing the zenith angle of the sun, the zenith angle of the satellite and the azimuth angle of the satellite relative to the sun;
The normalized visible light albedo A 0＝As/F is calculated, wherein A s represents the visible light albedo before normalization.
2. The FY-4A satellite-based ISCCP cloud classification method of claim 1, wherein: the cloud features include:
The visible light near infrared wave bands of the B01 channel, the B02 channel, the B03 channel, the B04 channel, the B05 channel and the B06 channel;
the middle wave infrared wave bands of the B07 channel and the B08 channel;
a water vapor channel of the B09 channel and the B10 channel;
the long-wave infrared wave bands of the B11 channel, the B13 channel and the B14 channel;
The bright temperature difference between the B09 channel and the B12 channel, the bright temperature difference between the B10 channel and the B13 channel, the bright temperature difference between the B11 channel and the B12 channel, the bright temperature difference between the B12 channel and the B13 channel, the bright temperature difference between the B08 channel and the B12 channel and the bright temperature difference between the B08 channel and the B10 channel;
And a B12 channel.
3. The FY-4A satellite-based ISCCP cloud classification method of claim 1, wherein: the ISCCP cloud classification model adopts a UNet++ network structure.
4. ISCCP cloud classification system based on FY-4A satellite, characterized in that: the device comprises a transformation module, a normalization module, a construction module, a standardization processing module and a classification module;
the transformation module is used for respectively carrying out projection transformation on the primary data and the primary geographic positioning data of the FY-4A satellite to obtain equal-longitude and latitude primary data and equal-longitude and latitude primary geographic positioning data;
The normalization module is used for performing visible light channel normalization processing based on the equal-longitude-latitude primary data and the equal-longitude-latitude primary geographic positioning data;
The construction module is used for constructing cloud characteristics based on the visible light channel after normalization processing and the equal longitude and latitude primary data;
the standardized processing module is used for carrying out standardized processing on the cloud characteristics;
the classification module is used for acquiring ISCCP cloud classification results by adopting a trained ISCCP cloud classification model based on the cloud characteristics of standardized processing;
the normalization module performs visible light channel normalization processing based on the equal-longitude-latitude primary data and the equal-longitude-latitude primary geographic positioning data, and comprises the following steps:
calculating normalization operators Wherein/> θ1、θ2、/>Respectively representing the zenith angle of the sun, the zenith angle of the satellite and the azimuth angle of the satellite relative to the sun;
The normalized visible light albedo A 0＝As/F is calculated, wherein A s represents the visible light albedo before normalization.
5. The FY-4A satellite-based ISCCP cloud classification system of claim 4, wherein: the cloud features include:
The visible light near infrared wave bands of the B01 channel, the B02 channel, the B03 channel, the B04 channel, the B05 channel and the B06 channel;
the middle wave infrared wave bands of the B07 channel and the B08 channel;
a water vapor channel of the B09 channel and the B10 channel;
the long-wave infrared wave bands of the B11 channel, the B13 channel and the B14 channel;
The bright temperature difference between the B09 channel and the B12 channel, the bright temperature difference between the B10 channel and the B13 channel, the bright temperature difference between the B11 channel and the B12 channel, the bright temperature difference between the B12 channel and the B13 channel, the bright temperature difference between the B08 channel and the B12 channel and the bright temperature difference between the B08 channel and the B10 channel;
And a B12 channel.
6. The FY-4A satellite-based ISCCP cloud classification system of claim 4, wherein: the ISCCP cloud classification model adopts a UNet++ network structure.
7. A storage medium having stored thereon a computer program, which when executed by a processor, implements the FY-4A satellite-based ISCCP cloud classification method of any one of claims 1 to 3.
8. A terminal, comprising: a processor and a memory;
the memory is used for storing a computer program;
the processor is configured to execute the computer program stored in the memory, so that the terminal performs the ISCCP cloud classification method based on the FY-4A satellite according to any one of claims 1 to 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110263471.1A CN112949726B (en) | 2021-03-11 | 2021-03-11 | ISCCP cloud classification method, system, medium and terminal based on FY-4A satellite |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110263471.1A CN112949726B (en) | 2021-03-11 | 2021-03-11 | ISCCP cloud classification method, system, medium and terminal based on FY-4A satellite |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112949726A CN112949726A (en) | 2021-06-11 |
| CN112949726B true CN112949726B (en) | 2024-05-14 |
Family
ID=76229150
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110263471.1A Active CN112949726B (en) | 2021-03-11 | 2021-03-11 | ISCCP cloud classification method, system, medium and terminal based on FY-4A satellite |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112949726B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113822381A (en) * | 2021-11-22 | 2021-12-21 | 国家卫星气象中心(国家空间天气监测预警中心) | Sea cloud detection method based on sea temperature difference threshold |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6035710A (en) * | 1999-05-28 | 2000-03-14 | Lockheed Martin Missiles & Space Co. | Cloud base height and weather characterization, visualization and prediction based on satellite meteorological observation |
| CN104502999A (en) * | 2014-12-10 | 2015-04-08 | 中国科学院遥感与数字地球研究所 | Day and night cloud detection method and device based on static satellite multichannel data |
| CN108846474A (en) * | 2018-05-18 | 2018-11-20 | 南京信息工程大学 | The satellite cloud picture cloud amount calculation method of convolutional neural networks is intensively connected based on multidimensional |
| CN110826526A (en) * | 2019-11-19 | 2020-02-21 | 上海无线电设备研究所 | Method for cloud detection radar to identify clouds |
| CN111210483A (en) * | 2019-12-23 | 2020-05-29 | 中国人民解放军空军研究院战场环境研究所 | Simulated satellite cloud picture generation method based on generation of countermeasure network and numerical mode product |
| CN112200787A (en) * | 2020-10-15 | 2021-01-08 | 中国科学院空天信息创新研究院 | Cloud detection method, storage medium and system for optical remote sensing image |
| CN112348058A (en) * | 2020-10-20 | 2021-02-09 | 华东交通大学 | Satellite cloud picture classification method based on CNN-LSTM network and computer readable storage medium |
-
2021
- 2021-03-11 CN CN202110263471.1A patent/CN112949726B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6035710A (en) * | 1999-05-28 | 2000-03-14 | Lockheed Martin Missiles & Space Co. | Cloud base height and weather characterization, visualization and prediction based on satellite meteorological observation |
| CN104502999A (en) * | 2014-12-10 | 2015-04-08 | 中国科学院遥感与数字地球研究所 | Day and night cloud detection method and device based on static satellite multichannel data |
| CN108846474A (en) * | 2018-05-18 | 2018-11-20 | 南京信息工程大学 | The satellite cloud picture cloud amount calculation method of convolutional neural networks is intensively connected based on multidimensional |
| CN110826526A (en) * | 2019-11-19 | 2020-02-21 | 上海无线电设备研究所 | Method for cloud detection radar to identify clouds |
| CN111210483A (en) * | 2019-12-23 | 2020-05-29 | 中国人民解放军空军研究院战场环境研究所 | Simulated satellite cloud picture generation method based on generation of countermeasure network and numerical mode product |
| CN112200787A (en) * | 2020-10-15 | 2021-01-08 | 中国科学院空天信息创新研究院 | Cloud detection method, storage medium and system for optical remote sensing image |
| CN112348058A (en) * | 2020-10-20 | 2021-02-09 | 华东交通大学 | Satellite cloud picture classification method based on CNN-LSTM network and computer readable storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 新一代静止卫星葵花-8的云分类研究及其应用;张成伟;《中国博士学位论文全文数据库（电子期刊）》;第2-3章 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112949726A (en) | 2021-06-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | A feature difference convolutional neural network-based change detection method | |
| CN112380952B (en) | Power equipment infrared image real-time detection and identification method based on artificial intelligence | |
| Wang et al. | Multiscale visual attention networks for object detection in VHR remote sensing images | |
| Dong et al. | Object detection in high resolution remote sensing imagery based on convolutional neural networks with suitable object scale features | |
| Su et al. | RCAG-Net: Residual channelwise attention gate network for hot spot defect detection of photovoltaic farms | |
| Lin et al. | Recent advances in intra-hour solar forecasting: A review of ground-based sky image methods | |
| Karami et al. | Automatic plant counting and location based on a few-shot learning technique | |
| Zhuo et al. | Cloud classification of ground-based images using texture–structure features | |
| CN110175615B (en) | Model training method, domain-adaptive visual position identification method and device | |
| CN112347895A (en) | Ship remote sensing target detection method based on boundary optimization neural network | |
| CN108428220A (en) | Satellite sequence remote sensing image sea island reef region automatic geometric correction method | |
| Wang et al. | A center location algorithm for tropical cyclone in satellite infrared images | |
| Bloshchinskiy et al. | Snow and cloud detection using a convolutional neural network and low-resolution data from the Electro-L No. 2 Satellite | |
| CN112949726B (en) | ISCCP cloud classification method, system, medium and terminal based on FY-4A satellite | |
| Alburshaid et al. | Palm trees detection using the integration between gis and deep learning | |
| Wang et al. | MashFormer: A novel multiscale aware hybrid detector for remote sensing object detection | |
| Singh et al. | Detection of changes in Landsat Images using Hybrid PSO-FCM | |
| CN112270285B (en) | SAR image change detection method based on sparse representation and capsule network | |
| CN117689995A (en) | Unknown spacecraft level detection method based on monocular image | |
| CN112365508A (en) | SAR remote sensing image water area segmentation method based on visual attention and residual error network | |
| xi Wang et al. | Detection of MMW radar target based on Doppler characteristics and deep learning | |
| CN113887455B (en) | Face mask detection system and method based on improved FCOS | |
| CN116403071A (en) | Method and device for detecting few-sample concrete defects based on feature reconstruction | |
| Wu et al. | Ground-based vision cloud image classification based on extreme learning machine | |
| Li et al. | Remote sensing image scene classification model based on dual knowledge distillation |
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 |