A method for rapid extraction and classification of astronaut sea ice hummocks

Abstract

This invention discloses a method for rapidly extracting and classifying interglacial lakes in the Antarctic Cosmonaut Sea. The steps include: quality control of sea ice concentration data, identification and extraction of interglacial lakes, classification of interglacial lakes, and data storage. This method improves the speed and accuracy of extracting spatial distribution, spatial morphology, and area parameters of interglacial lakes. It enables the extraction of interglacial lakes by category, facilitating research on the formation mechanisms of different types of interglacial lakes. It also improves the degree of automated data processing, reduces manual intervention and workload, standardizes data processing criteria, improves data quality, and obtains long-term data sequences. Furthermore, it helps researchers to promptly grasp the near-real-time status of interglacial lakes during the sea ice growth season (May-August) in the Antarctic Cosmonaut Sea, enhancing understanding of the formation and dissipation processes of interglacial lakes, further strengthening scientific understanding of the Cosmonaut Sea, providing data support for future scientific expeditions, and serving my country's national strategic interests in Antarctica.

Description

Technical Field

[0001] This invention relates to the field of polar sea ice exploration science and technology, specifically a method for rapidly extracting and classifying inter-sea ice lakes used by astronauts. Background Technology

[0002] Polar sea ice is a crucial component of the global climate system. By reflecting most of the incident solar radiation, it blocks energy exchange between the atmosphere and the ocean, playing a vital role in the climate system. It is also one of the most sensitive elements to global climate change; against the backdrop of global warming, it influences atmospheric and oceanic circulation through energy balance processes, significantly impacting the climate system. Antarctica, perpetually covered by ice sheets and sea ice, is one of Earth's cold sources and has a significant influence on the global climate system.

[0003] A polyglacial lake is an ice-free body of water that forms and remains frozen during the sea ice freezing period. Based on their formation mechanism, they are classified into sensible and latent heat polyglacial lakes. Sensible polyglacial lakes, also known as open-water polyglacial lakes, are mainly formed by the mixing of ocean water and the heating of sea ice. The continuous heating of the ocean causes sea ice to melt and simultaneously prevents the formation of new ice. Latent heat polyglacial lakes often form near the coast and are also known as coastal polyglacial lakes. They are mainly influenced by cascading winds along the coast. These continuous cascading winds constantly blow newly grown sea ice away from the coast, and latent heat polyglacial lakes often have a high sea ice growth rate. Past observations and studies have found two long-standing sensible polyglacial lakes in Antarctica: one in the Weddell Sea and the other in the Cosmonaut Sea. Numerous coastal polyglacial lakes also exist along the Antarctic coast. The presence of polyglacial lakes alters the energy exchange between the ocean and the atmosphere, affecting sea ice growth and further impacting the marine ecosystem.

[0004] Cosmonaut Sea contains several coastal and open-water interglacial lakes. The open-water interglacial lakes, discovered in 1987, appear almost annually from May to August. The presence of these lakes influences the ocean's thermal and dynamic processes, as well as sea ice growth. Cosmonaut Sea is a key area of ​​focus for my country's Antarctic expeditions this year; therefore, the identification, extraction, classification (open-water and coastal interglacial lakes), and construction of long-term time series of these lakes are of great significance.

[0005] However, the Antarctic environment is extremely harsh, with severe natural conditions, and the spatial distribution of interglacial lakes is vast, making on-site observation extremely difficult. Currently, the identification of interglacial lakes mainly relies on satellite remote sensing observations, primarily microwave remote sensing, optical satellite observations, and synthetic aperture radar (SAR) observations. However, there are still certain problems with the extraction of interglacial lakes. High-resolution optical satellite data has a high accuracy rate in extracting interglacial lakes, but the observational data is affected by clouds, impacting the spatial and temporal continuity of the data, making it impossible to construct long-term series and analyze long-term changes in interglacial lakes in the Cosmonaut Sea. High-resolution SAR satellite imagery is limited by satellite revisit time and swath width, making continuous daily observation of the Cosmonaut Sea impossible.

[0006] Microwave remote sensing data offers long observation periods and wide coverage, making it the optimal choice for obtaining long-term, continuous data on the distribution and changes of interglacial lakes, although certain challenges remain. Interglacial lake observations based on microwave data primarily involve inverting sea ice concentration from the microwave data, extracting interglacial lakes using sea ice concentration thresholds, and calculating their areas. However, most microwave sea ice concentration data extraction methods are research-oriented. Currently, in the Cosmonaut Sea, the focus is mainly on calculating the area of ​​interglacial lakes in open waters, and the data is generally multi-day averages without classifying the lakes, often requiring manual intervention.

[0007] Therefore, there is a need to find a method and device that can identify quickly, without human intervention, and can operate stably for a long time.

[0008] Currently, the commonly used method is to invert sea ice concentration based on microwave data and extract the distribution and area of ​​interglacial lakes through sea ice concentration thresholds. However, this method has drawbacks. First, it requires manual intervention, which can easily introduce subjective errors, inconsistent judgment standards, large workload, slow processing speed, long processing time, and low efficiency. Second, it can only identify specific individual interglacial lakes and cannot extract the distribution and number of interglacial lakes in the entire Astronaut Sea. Third, it cannot classify and identify interglacial lakes in open water and those along the coast. Fourth, it cannot provide long-term continuous time series data. Fifth, there are no publicly available interglacial lake data products.

[0009] Therefore, a method for rapidly extracting and classifying interglacial lakes for astronauts is designed to solve the above problems. Summary of the Invention

[0010] The purpose of this invention is to provide a method for rapidly extracting and classifying interglacial lakes in the astronaut sea, improving the speed and accuracy of extracting spatial distribution, spatial morphology, and area parameters of interglacial lakes, enabling the extraction of interglacial lakes by category, reducing manual intervention and workload, unifying the judgment criteria for data processing, and improving the quality of data use, thereby solving the problems mentioned in the background art.

[0011] To achieve the above objectives, the present invention provides the following technical solution: a method for rapidly extracting and classifying interglacial lakes in astronaut seas, characterized by the following steps:

[0012] Step 1: Sea ice concentration data quality control; Perform quality control on sea ice concentration (SIC), identify invalid data such as missing measurements, convert the valid range of sea ice concentration values, mark land values, and then use a land mask to assign values ​​to land values ​​to perform sea ice concentration data quality control.

[0013] Step 2: Identification and extraction of interglacial lakes; Based on the results of Step 1, a sea ice concentration threshold is set within the quality-controlled sea ice concentration data, and land values ​​are assigned values; Connected component identification is performed on the data, and the identification results are divided into spatial distribution of connected components and number of connected components; The spatial distribution of connected components is the result of interglacial lake identification; Frequency statistics are performed on the values ​​in the spatial distribution of connected components, and each value is sorted from high to low frequency; The area of ​​each interglacial lake is calculated;

[0014] Step 3: Classification of interglacial lakes; Based on the results of Step 1, a sea ice concentration threshold is set in the data, and the land values ​​are reassigned; Connectivity component identification is performed on the data again, and the identification results are divided into spatial distribution of connected components and number of connected components, while extracting the values ​​of land markers; Based on the results of Step 2, the data is compared and classified, and interglacial lakes are classified into coastal interglacial lakes and offshore interglacial lakes.

[0015] Step four, data storage.

[0016] Preferably, the spatial resolution of the sea ice concentration product is set to 25km, the time period is set to 1979 to the present, the time resolution is once every two days before August 20, 1987, and once a day thereafter. The identification, extraction and classification of interglacial lakes are mainly concentrated in May to August each year, which is the growth period of Antarctic sea ice.

[0017] In sea ice concentration (SIC), 0-100 is a valid value. To convert it to 0-1.0, land is set to nan, and some invalid values ​​are also nan. Some data may contain invalid values ​​due to missing measurements, so it is necessary to perform quality control on the original sea ice concentration data. Using a land mask, the nan values ​​of non-land in the seawater are set to 0, resulting in the quality-controlled sea ice concentration data.

[0018] Preferably, in step two, the threshold for SIC1 data is set to 0.7, 0≤SIC<0.7 is set to 1.0, SIC≥0.7 is set to 0.0, and land (nan) is set to 0.0;

[0019] In the frequency statistics of the values ​​in the spatial distribution of connected domains, the range of the statistics is 0 to (number of connected domains - 1); based on the frequency statistics ranking results, seawater and sea ice connected domains are removed, and interglacial lake connected domains are retained. The area of ​​each interglacial lake is calculated by combining grid area data and interglacial lake indicator values ​​in the statistical data.

[0020] Preferably, in step three, the threshold is 0.7, 0≤SIC<0.7 is set to 1.0; SIC≥0.7 is set to 0.0, and land (nan) is set to 1.5. Connectivity identification is performed again. Based on the statistical ranking results of the value frequency in the spatial distribution of connected components, seawater and sea ice connected components are removed, and interglacial lake connected components are retained. The results of step three and step two are compared to classify interglacial lakes.

[0021] Preferably, in step four, the stored data includes the region's latitude and longitude, sea ice concentration, interglacial lake identification results, interglacial lake markings, interglacial lake classification results, and the number of interglacial lakes.

[0022] Compared with existing technologies, the beneficial effects of this invention are: the advantages of this method for rapidly extracting and classifying interglacial lakes for astronauts compared to other inventions are:

[0023] 1. This invention proposes a method and apparatus for rapid and automatic extraction and classification of interglacial lakes in astronaut seas, which improves the speed and accuracy of extracting spatial distribution, spatial morphology, and area parameters of interglacial lakes;

[0024] 2. It can extract interglacial lakes by category, which facilitates researchers' study on the formation mechanism of different types of interglacial lakes;

[0025] 3. It improved the level of automated data processing, reduced manual intervention and workload, standardized the judgment criteria for data processing, improved the quality of data use, and obtained long-term data sequences.

[0026] 4. It can help researchers to grasp the near-real-time status of interglacial lakes in the Antarctic astronaut sea during the sea ice growth season (May-August), which is conducive to enhancing the understanding of the formation and dissipation processes of interglacial lakes, further strengthening the scientific understanding of the astronaut sea, providing data support for future scientific expeditions, and serving my country's national strategic interests in Antarctica. Attached Figure Description

[0027] Figure 1 This is a cloud map of the original sea ice concentration data;

[0028] Figure 2 A cloud map of sea ice concentration data after data quality control;

[0029] Figure 3 A schematic diagram of the sea ice concentration ice water identification result 1 (SIC1);

[0030] Figure 4 A schematic diagram of the sea ice concentration ice water identification result 2 (SIC2);

[0031] Figure 5 This is a schematic diagram of the spatial distribution results of connected component identification based on SIC1 (Lsic1);

[0032] Figure 6 This is a schematic diagram of the spatial distribution results of connected component identification based on SIC1 (Lsic1);

[0033] Figure 7 This is a schematic diagram showing the spatial distribution results of connected component identification based on SIC1 (Lsic1), with landmass set to nan.

[0034] Figure 8 This is a schematic diagram showing the spatial distribution results of connected component identification based on SIC1 (Lsic1), with the land and interglacial lake regions set to nan.

[0035] Figure 9 This is a schematic diagram of the spatial distribution results of connected component identification based on SIC2 (Lsic2);

[0036] Figure 10 A schematic diagram showing the classification results of interglacial lakes;

[0037] Figure 11 A flowchart for identifying, extracting, and classifying interglacial lakes based on sea ice concentration (SIC) data;

[0038] Figure 12 A schematic diagram illustrating the storage of identification and classification results for interglacial lakes. Detailed Implementation

[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] This invention provides a technical solution:

[0041] A method for rapidly extracting and classifying interglacial lakes in astronaut seas comprises the following steps:

[0042] Step 1: Sea ice concentration data quality control; Perform quality control on sea ice concentration (SIC), identify invalid data such as missing measurements, convert the valid range of sea ice concentration values, mark land values, and then use a land mask to assign values ​​to land values ​​to perform sea ice concentration data quality control.

[0043] In step one, the spatial resolution of the sea ice concentration product is set to 25km, the time length is set to 1979-present, the time resolution is once every two days before August 20, 1987, and once a day thereafter. The identification, extraction and classification of interglacial lakes are mainly concentrated in May to August each year, which is the growth period of Antarctic sea ice.

[0044] In sea ice concentration (SIC), 0-100 is considered valid values. Converting this to 0-1.0, land is represented as nan, and some invalid values ​​are also nan. nan indicates an invalid value. Some data may contain invalid values ​​due to missing measurements, etc. Therefore, it is necessary to modify the original sea ice concentration data (e.g., ...). Figure 1 (The black line in the figure represents the 70% sea ice concentration contour line, and the white dots represent missing measurements.) Quality control was performed using a land mask, where the grid types in the mask data correspond to both land and ocean. The NaN values ​​for non-land elements in the seawater were set to 0. Figure 2 The data represents sea ice concentration after quality control. The black line represents the 70% sea ice concentration contour line.

[0045] Step two, interglacial lake identification and extraction; based on the results of step one, a sea ice concentration threshold is set within the quality-controlled sea ice concentration data, and land values ​​are assigned values. This set of data is denoted as SIC1 (e.g., Figure 3 The data is connected component identification. The method involves binary processing the image, assigning labels, and determining if the labels of the surrounding eight points are the same. Through iterative loops, all regions with the same connectivity features are identified as connected regions. The identification results are divided into connected component spatial distribution (Lsic) and connected component count (Ln); the connected component spatial distribution represents the results of interglacial lake identification. Frequency statistics are performed on the values ​​in the connected component spatial distribution, and each value is sorted from high to low frequency. The area of ​​each interglacial lake is then calculated.

[0046] In step two, for the SIC1 data, the threshold is set to 0.7, 0≤SIC<0.7 is set to 1.0, SIC≥0.7 is set to 0.0, and land (nan) is set to 0.0;

[0047] In the frequency statistics of values ​​in the spatial distribution of connected components, the range of the statistics is 0 to (number of connected components - 1); in the statistical data, the area of ​​each interglacial lake is calculated by combining grid area data and interglacial lake indicator values.

[0048] Specifically, the spatial distribution (Lsic) and the number of connected components (Ln) of the identification results are denoted as Lsic1 and Ln1, respectively. Figure 5 Lsic1 represents the interglacial lake identification result, denoted as Polynyas. In Polynyas, the land value is set to 999 based on the land mask. Figure 6 ).

[0049] Using a land mask, set the land value in Lsic1 to nan( Figure 7 ), denoted as Lsic1T.

[0050] Perform frequency statistics on the values ​​in Lsic1, denoted as xL (xL contains the frequency corresponding to each value), with the statistical range being 0 to (Ln1-1), and sort each value from high to low frequency as xL2.

[0051] Based on the studied sea area, it can be seen that in xL2, the regions with the most pixels are the sea ice and seawater regions, namely... Figure 7 Values ​​of 0 and 1 represent sea ice and seawater regions (land is set to nan and is not included in the statistics). Other values ​​represent interglacial lake regions. The interglacial lake indicator value is denoted as PolynyasFlag. When the number of connected components Ln is less than 2, it indicates that no interglacial lake exists, and PolynyasFlag is denoted as -1. Combining the grid area data and the interglacial lake indicator value PolynyasFlag, the area of ​​each interglacial lake is calculated and denoted as PolynyasA.

[0052] Step 3, classifying interglacial lakes; based on the results of Step 1, a sea ice concentration threshold is set within the data, and the land values ​​are reassigned. This set of data is denoted as SIC2 (e.g., Figure 4 The data is then re-identified as a connected component, with the results divided into spatial distribution and number of connected components. The values ​​of land markers are also extracted. Based on the results of step two, the data is compared and classified into coastal and offshore interglacial lakes.

[0053] In the SIC2 data, the threshold is 0.7, 0≤SIC<0.7 is set to 1.0; SIC≥0.7 is set to 0.0, and land (nan) is set to 1.5;

[0054] Specifically:

[0055] Based on the results in step two, the labeling value of the interglacial lake region in Lsic1T (i.e., the connected domain that is not seawater and sea ice) is further set to nan, denoted as Lsic1T2. Figure 8 );

[0056] Connectivity component identification was performed on SIC2, and the spatial distribution and number of connected components were denoted as Lsic2 and Ln2, respectively. Figure 9 (The black dots in the image represent land markers). Extract the values ​​of the land markers and denote them as Landflag;

[0057] Mark the values ​​in Lsic1T2 that are Nan and where Lsic2 equals Landflag as -1, and denote the new array as Lsic1T3. Mark the values ​​in Lsic1T2 that are Nan and where Lsic1T3 is not equal to -1 as -2, and denote the new array as Lsic1T4. Use a landmass mask to set the land portion of Lsic1T4 to 999, and denote it as PolynyasC. This is the classification result for interglacial lakes. In PolynyasC, a value of -1 indicates a coastal interglacial lake, and -2 indicates an offshore interglacial lake. Figure 10 In the figure, the area with a value of -1 (white) represents coastal interglacial lakes, and the area with a value of -2 (white) represents open water interglacial lakes.

[0058] Step four, data storage.

[0059] In step four, the stored data includes the region's latitude and longitude (Lon), latitude (Lat), sea ice concentration (Sic), interglacial lake identification results (Polynyas), interglacial lake flags (PolynyasFlag), interglacial lake classification results (PolynyasC), and the number of interglacial lakes (PolynyasN).

[0060] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for rapidly extracting and classifying interglacial lakes in astronaut seas, characterized in that, The steps are as follows: Step 1: Quality control of sea ice concentration data; perform quality control on sea ice concentration (SIC), identify missing and invalid data, convert the valid range of sea ice concentration values, mark land values, and then use a land mask to assign values ​​to land values. Step 2: Identification and extraction of interglacial lakes; Based on the results of Step 1, a sea ice concentration threshold is set within the quality-controlled sea ice concentration data, and land values ​​are assigned values; Connected component identification is performed on the data, and the identification results are divided into spatial distribution of connected components and number of connected components; The spatial distribution of connected components is the result of interglacial lake identification; Frequency statistics are performed on the values ​​in the spatial distribution of connected components, and each value is sorted from high to low frequency; The area of ​​each interglacial lake is calculated; Step 3, classification of interglacial lakes; Based on the results of Step 1, a sea ice concentration threshold is set in the data, and the land values ​​are reassigned; Connectivity component identification is performed on the data again, and the identification results are divided into spatial distribution of connected components and number of connected components, while extracting the values ​​of land markers; Based on the results of Step 2, the data is compared and classified, and interglacial lakes are classified into coastal interglacial lakes and offshore interglacial lakes. Step four, data storage; In step one, the valid values ​​of sea ice concentration (SIC) from 0 to 100 are converted to 0 to 1.

0. Land is set to nan, and some invalid values ​​are also nan. Using a land mask, the nan values ​​of non-land in the seawater are set to 0, which is the sea ice concentration data after quality control. In step two, for the SIC1 data, the threshold is set to 0.7, 0≤SIC<0.7 is set to 1.0, SIC≥0.7 is set to 0.0, and land (nan) is set to 0.0; In the frequency statistics of the values ​​in the spatial distribution of connected domains, the range of the statistics is 0 to (number of connected domains - 1); based on the frequency statistics ranking results, the seawater and sea ice connected domains are removed, and the interglacial lake connected domains are retained. The area of ​​each interglacial lake is calculated by combining grid area data and interglacial lake indicator values ​​in the statistical data. In step three, the threshold is 0.7, 0≤SIC<0.7 is set to 1.0, SIC≥0.7 is set to 0.0, and land (nan) is set to 1.

5. Connectivity identification is performed again. Based on the statistical ranking results of the value frequency in the spatial distribution of connected components, seawater and sea ice connected components are removed, and interglacial lake connected components are retained. The results of step three and step two are compared to classify interglacial lakes.

2. The method for rapidly extracting and classifying interglacial lakes for astronauts according to claim 1, characterized in that: The spatial resolution of the sea ice concentration product is set at 25km, and the time period is set from 1979 to the present. The time resolution is once every two days until August 20, 1987, and once a day thereafter. The identification, extraction and classification of interglacial lakes are concentrated in May to August each year, which is the growth period of Antarctic sea ice. Some data may contain invalid values ​​due to missing measurements, necessitating quality control of the raw sea ice concentration data.

3. The method for rapidly extracting and classifying interglacial lakes for astronauts according to claim 2, characterized in that: In step four, the stored data includes the latitude and longitude of the region, sea ice concentration, interglacial lake identification results, interglacial lake markings, interglacial lake classification results, and the number of interglacial lakes.

Images