A method for extracting the extent of arctic ice lakes

Abstract

The present application relates to the technical field of monitoring and forecasting of sea ice, and specifically discloses a method for extracting the range of Arctic ice-lake, comprising: obtaining original sea ice density data, and supplementing missing data to obtain sea ice density data; performing threshold segmentation and connected domain analysis on the sea ice density data to obtain sea ice density partitions; determining initial growth cores of the ice-lake growth area and the edge ice-water mixed growth area; taking pixels in the neighborhood of the initial growth cores of the ice-lake growth area and the edge ice-water mixed growth area that satisfy a preset rule as corresponding growth points, and taking the growth points as new growth cores for iteration until the range is stable to stop iteration, thereby obtaining an intermediate ice-lake area and an intermediate ice-water mixed area; and performing connectivity analysis on the central discrete low-density ice area to output the range of the ice-lake in the whole Arctic, the range of the edge ice, and the range of the open water area. The present application realizes automatic extraction of the real shape and range of the ice-lake without limitation.

Description

Technical Field

[0001] This invention relates to the field of sea ice monitoring and forecasting technology, specifically to a method for extracting the extent of interglacial lakes in the Arctic. Background Technology

[0002] As an important mesoscale physical oceanographic phenomenon in polar ice regions, interglacial lakes play a crucial role in the exchange of matter and energy between the sea, ice, and atmosphere. They possess unique thermodynamic characteristics, such as high heat flux, ice production, and salinity, which have profound impacts on the polar climate system and its ecosystem. Due to their high sensitivity to even minor changes in the polar cryosphere system, the dynamic changes in the geographical location and extent of interglacial lakes have become important indicators reflecting climate change and alterations in the polar marine environment.

[0003] Against the backdrop of global warming, particularly the significant warming effect in the Arctic (i.e., the Arctic amplification effect), Arctic sea ice cover is experiencing an unprecedented and rapid decline. However, the spatiotemporal distribution characteristics and dynamic response mechanisms of interglacial lakes during this long-term change process are still poorly understood by the scientific community. Therefore, strengthening the observation and analysis of long-term changes in interglacial lakes and exploring their response patterns under global warming has become an urgent task in current polar scientific research. This is not only crucial for the protection and management of the polar environment but also has immeasurable scientific value for the accurate simulation and prediction of the global climate system.

[0004] There is no clear boundary between interglacial lakes and the surrounding thick ice, and quantitative studies require defining the extent of these lakes. Within interglacial lakes, the rate of new ice formation is high, sea ice freezes rapidly, there are numerous brine cells, and the sea ice has high usable salinity. These characteristics influence their attribute values ​​in remote sensing imagery, thus distinguishing them from the surrounding thick ice. The sea ice remote sensing data used to infer the extent of interglacial lakes mainly include sea ice concentration, thin ice thickness, sea ice passive microwave brightness temperature, and SAR backscattering coefficient.

[0005] Sea ice concentration refers to the percentage of sea ice coverage in a region. The sea ice concentration of interglacial lakes is lower than that of the surrounding sea ice, and this characteristic can be used to extract their extent. Previous methods mainly include the threshold method, the system point method, and the water integration method. The threshold method is the most commonly used method for determining the extent of interglacial lakes based on sea ice concentration. If pixels below a specified threshold are surrounded by sea ice concentrations above a specified threshold, then the low-concentration area is considered an interglacial lake. Researchers typically select a threshold between 60% and 85%. This method is independent of other sea ice parameters and is currently one of the most direct and efficient methods for defining the extent of interglacial lakes. Furthermore, the scientific community has achieved the commercial release of Arctic sea ice concentration data products, with long data time series (from 2002 to the present) and convenient access. In addition, the spatial resolution of these data products can reach several kilometers, and the temporal resolution can reach one day.

[0006] Currently, thresholding is used to extract the extent of interglacial lakes. This often involves pre-defining a "boundary box" for the lake, and within that specific box, thresholding is used to extract its extent. This leads to several drawbacks and errors, such as: it's often used to extract interglacial lakes in a specific area or region, resulting in poor comparability between different extracted lakes; there's no method for extracting all interglacial lakes in the Arctic; it cannot extract the extent of lakes outside the bounding box, potentially leading to underestimation of some lakes; different researchers using different bounding boxes for the same lake can result in significant discrepancies in extraction results; when the area within the bounding box connects to open water outside, the existence or disappearance of the interglacial lake is uncertain, but it's still identified as an interglacial lake because it's within the bounding box. Furthermore, current research on Arctic interglacial lakes is mostly based on specific land coasts or islands, areas typically far from edge ice zones with relatively high sea ice concentrations and long-lasting interglacial lake durations. Few studies focus on interglacial lakes near edge ice zones. Edge ice zones, similar to interglacial lakes, have low sea ice concentrations, making pixel-level differentiation difficult.

[0007] Therefore, there is an urgent need for a method that can extract the entire range of Arctic interglacial lakes without relying on other parameters or being constrained by artificial boundaries. Summary of the Invention

[0008] To address the aforementioned problems, the purpose of this invention is to provide a method for extracting the extent of Arctic interglacial lakes, thereby overcoming artificial boundary constraints and enabling the unified extraction of the entire extent of Arctic interglacial lakes, thus providing technical support for the observation and analysis of long-term changes in Arctic interglacial lakes.

[0009] This invention provides a method for extracting the extent of interglacial lakes in the Arctic, comprising:

[0010] Step S1: Obtain the original sea ice concentration data of Arctic sea ice, and supplement the missing data in the original sea ice concentration data to obtain the sea ice concentration data.

[0011] Step S2: Threshold segmentation and connected component analysis are performed on the sea ice concentration data to obtain sea ice concentration partitions; the sea ice concentration partitions include: a central discrete low-density ice zone, an edge ice zone, an interglacial lake growth zone, and an edge ice-water mixed growth zone.

[0012] Step S3: Determine the initial growth nucleus of the interglacial lake growth zone based on the boundary of the interglacial lake growth zone, and determine the initial growth nucleus of the edge ice-water mixed growth zone based on the boundary of the edge ice-water mixed growth zone;

[0013] Step S4: Pixels that meet the preset rules in the neighborhood of the initial growth kernel of the interglacial lake growth area and the edge ice-water mixed growth area are respectively taken as the corresponding growth points, and the growth points are used as new growth kernels for iteration until the range of the interglacial lake growth area and the edge ice-water mixed growth area is stable, and the iteration stops to obtain the intermediate interglacial lake area and the intermediate ice-water mixed area.

[0014] Step S5: Perform connectivity analysis on the central discrete low-density ice zone, and combine the intermediate interglacial lake zone and the intermediate ice-water mixing zone to output the range of interglacial lakes, edge ice range and open water range of the entire Arctic.

[0015] In one possible implementation,

[0016] Step S1 includes:

[0017] Obtain an ocean mask;

[0018] The missing ocean data in the original sea ice concentration data is assigned a value of 0 using the ocean mask.

[0019] The remaining missing data in the original sea ice concentration data are assigned as the first data; the time corresponding to the original sea ice concentration data is the first time; the first data is the interpolation of the sea ice concentration data of two second times adjacent to the first time.

[0020] In one possible implementation, step S2 includes:

[0021] The sea ice concentration data is segmented based on a first preset sea ice concentration threshold to obtain dense ice areas where the sea ice concentration data is greater than the first preset sea ice concentration threshold and low-density ice areas where the sea ice concentration data is not greater than the first preset sea ice concentration threshold.

[0022] Connectivity analysis was performed on the low-density ice region to obtain the two largest connected regions and the central discrete low-density ice region.

[0023] The connected region is segmented based on a second preset sea ice concentration threshold to obtain the edge ice zone that does not contain open water and the edge low-density ice zone that contains open water; the second preset sea ice concentration threshold is less than the first preset sea ice concentration threshold.

[0024] Connectivity analysis was performed on the edge low-density ice zone to obtain the edge ice-water mixed growth zone and the interglacial lake growth zone that is not connected to open water and is close to the edge ice zone.

[0025] In one possible implementation, step S3 includes:

[0026] The interglacial lake growth area and the edge ice-water mixed growth area are binarized respectively, and the boundaries of all sub-regions of the interglacial lake growth area and the boundaries of all sub-regions of the edge ice-water mixed growth area are obtained by a region boundary tracking algorithm.

[0027] The boundaries of all sub-regions of the interglacial lake growth zone are used as the growth nuclei of the interglacial-water mixed growth zone, and the boundaries of all sub-regions of the marginal interglacial-water mixed growth zone are used as the growth nuclei of the interglacial lake growth zone.

[0028] In one possible implementation, the growth kernel is a set of region boundary pixels.

[0029] In one possible implementation, step S4 includes:

[0030] The pixels in the eight neighborhoods of the initial growth nucleus of the interglacial lake growth region that are not land, not interglacial lake, and not densely ice-covered areas are taken as the first undetermined growth points.

[0031] The first undetermined growth point, whose sea ice concentration data is not less than that of the growth nucleus and whose eight neighboring regions do not contain open water, is determined as the new growth nucleus of the interglacial lake growth area. The next iteration is then carried out until the range of the interglacial lake growth area stops growing, at which point the iteration stops, and the preliminary interglacial lake area is obtained.

[0032] In one possible implementation, step S4 further includes:

[0033] The pixels in the eight neighborhoods of the initial growth kernel of the edge ice-water mixed growth zone that are not land, not interglacial lakes, and not dense ice areas are used as the second undetermined growth points.

[0034] The second undetermined growth point in the eight neighboring regions, excluding the initial interglacial lake region, is determined as the new growth nucleus of the marginal ice-water mixed growth region, and the next iteration is performed until the range of the marginal ice-water mixed growth region no longer increases, at which point the iteration stops, and the initial ice-water mixed region is obtained.

[0035] Based on the unclassified pixels in the edge ice zone that are not classified as preliminary interglacial lake areas and the pixels in the preliminary ice-water mixing area, a connected component analysis is performed. Pixels in the eight-neighborhood of the unclassified pixels that contain the preliminary ice-water mixing area are added to the preliminary ice-water mixing area to obtain the intermediate ice-water mixing area; the remaining unclassified pixels are added to the preliminary interglacial lake area to obtain the intermediate interglacial lake area.

[0036] In one possible implementation, step S5 includes:

[0037] Connectivity analysis is performed sequentially on the pixels of the central discrete low-density ice zone to determine that the pixels in the eight neighborhoods of the central discrete low-density ice zone including the middle ice-water mixing zone are the first pixels, and the rest are the second pixels.

[0038] Adding the first pixel to the intermediate ice-water mixing zone yields the ice-water mixing zone, and adding the second pixel to the intermediate interglacial lake zone yields the interglacial lake range.

[0039] In one possible implementation, step S5 further includes:

[0040] Threshold screening is performed on the ice-water mixing area. The ice-water mixing area with sea ice concentration data greater than 0% is classified as the edge ice range, and the remaining ice-water mixing area is classified as the open water range.

[0041] The method for extracting the extent of Arctic interglacial lakes provided by this invention utilizes a hierarchical partitioning and refinement strategy and formulates pixel-level region growth rules to achieve automatic extraction of the true shape and extent of unrestricted interglacial lakes. Attached Figure Description

[0042] Figure 1 A flowchart illustrating the method for extracting the extent of interglacial lakes in the Arctic, provided as an embodiment of the present invention;

[0043] Figure 2 A density partitioning diagram provided for an embodiment of the present invention;

[0044] Figure 3 A schematic diagram illustrating the extraction of interglacial lake extent provided for an embodiment of the present invention. Detailed Implementation

[0045] The embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and examples. The following detailed description of the embodiments and the accompanying drawings are used to illustrate the principles of the present invention by way of example, but should not be used to limit the scope of the present invention. That is, the present invention is not limited to the described preferred embodiments, and the scope of the present invention is defined by the claims.

[0046] In the description of this invention, it should be noted that, unless otherwise stated, "a plurality of" means two or more; the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance; those skilled in the art can understand the specific meaning of the above terms in this invention as appropriate.

[0047] Figure 1 A flowchart illustrating the method for extracting the extent of Arctic interglacial lakes provided as an embodiment of the present invention is shown below. Figure 1 As shown, this invention provides a method for extracting the extent of interglacial lakes in the Arctic, comprising:

[0048] Step S1: Obtain the raw sea ice concentration data of Arctic sea ice, and supplement the missing data in the raw sea ice concentration data to obtain the sea ice concentration data.

[0049] In one possible implementation, the sea ice concentration product has a time span from 2012 to the present, a spatial resolution of 6.25 km, and a temporal resolution of 1 day. The raw sea ice concentration data has valid values ​​from 0 to 100, expressed as a percentage. Due to inherent limitations of the observation satellite system, large areas of no-value zones appear in some regions. Both the land area and the no-value zones have the following attributes: Data quality control is necessary because it can easily lead to confusion and affect subsequent results.

[0050] Based on daily raw sea ice concentration data of Arctic sea ice, the following data was obtained from 2012 to the present. The cumulative frequency map shows that the most frequent pixels are land, thus yielding land mask data. ( (The region with the highest frequency). According to The cumulative frequency map creates a partial ocean mask for non-land areas with a cumulative frequency of at least ten occurrences. ( (Ocean areas far from land, with a frequency greater than 10 times).

[0051] Obtain an ocean mask; use the ocean mask to assign 0 to the missing ocean data in the original sea ice concentration data; assign the remaining missing data in the original sea ice concentration data to the first data; the time corresponding to the original sea ice concentration data is the first time; the first data is the interpolation of the sea ice concentration data of the two second times adjacent to the first time.

[0052] In one example, the obtained daily sea ice concentration data is first used... The missing portion of the ocean is assigned a value of 0. Then, the attribute is... For non-land areas, connected component analysis is performed to detect regions outside the ocean mask area where data is missing, and the area of ​​each connected component is calculated. If a large area of ​​missing data exists in a connected region, valid raw sea ice concentration data for the two adjacent days are obtained for that region. The missing data for the current day is then supplemented by linear weighted interpolation of the raw sea ice concentration data for the two adjacent days, as shown in the following formula:

[0053] ;

[0054] in, For the missing sea ice concentration data to be interpolated, For the number of consecutive days of missing or corrupted data, and For two complete and valid sea ice concentration data points with adjacent missing dates, The date of the missing data is relatively recent. The dates that are relatively far from the missing data are handled by first processing the missing dates with the nearest adjacent complete and valid data. This is done using a "closed" method of interpolation from both ends towards the middle, until all missing dates are filled.

[0055] Step S2: Threshold segmentation and connected component analysis are performed on the sea ice concentration data to obtain sea ice concentration partitions;

[0056] The sea ice concentration zones include: a central discrete low-density ice zone, a marginal ice zone, an interglacial lake growth zone, and a marginal ice-water mixed growth zone. Figure 2 A density partitioning diagram provided for an embodiment of the present invention.

[0057] In one possible implementation, the sea ice concentration data is segmented based on a first preset sea ice concentration threshold to obtain dense ice areas where the sea ice concentration data is greater than the first preset sea ice concentration threshold and low-density ice areas where the sea ice concentration data is not greater than the first preset sea ice concentration threshold.

[0058] Connectivity analysis of the low-density ice region yielded the two largest connected regions and the central discrete low-density ice region.

[0059] Based on the second preset sea ice concentration threshold, the connected components are segmented by threshold to obtain the edge ice zone that does not contain open water and the edge low-density ice zone that contains open water; the second preset sea ice concentration threshold is less than the first preset sea ice concentration threshold.

[0060] Connectivity analysis was performed on the marginal low-density ice zone to obtain the marginal ice-water mixed growth zone and the interglacial lake growth zone that is not connected to the open water area but is close to the marginal ice zone.

[0061] Connected regions are areas that are connected geographically and pixelally.

[0062] In one example, the first preset sea ice concentration threshold The second preset sea ice concentration threshold is 75%. 15%, designated as dense ice zone It does not include interglacial lakes.

[0063] In one example, a divide-and-conquer strategy was adopted to process the entire Arctic region into smaller, more detailed sections, ultimately dividing the image into four independent blocks to be processed. The goal was to obtain the interglacial lake growth zone and the edge ice-water mixed growth zone.

[0064] First, use the first preset sea ice concentration threshold. Sea ice concentration data Threshold segmentation is performed to obtain dense ice areas. (marked as) , Low-density ice areas ( ).

[0065] Among them, low-density ice areas Includes all interglacial lakes (denoted as , All edge ice zones located at the edge of sea ice (denoted as...) , ) and large areas of open water (denoted as , ).

[0066] Afterwards, for low-density ice areas Perform connected component analysis. After binarization, all pixels connected in four directions are labeled as the same connected component. All connected component attributes are statistically analyzed, including sequence number, region pixel index, region centroid index, and total number of pixels (i.e., range). Except for the Arctic Ocean, the Arctic sea ice concentration data includes portions of the North Atlantic and North Pacific Oceans. The two connected components with the largest ranges are identified. These include vast areas of open water in the North Atlantic and the North Pacific, as well as the adjacent marginal ice belts and a small number of interglacial lakes; the remaining connected area is the central discrete low-density ice zone. This area is mainly composed of interglacial lakes within the Arctic ice zone.

[0067] Then, a lower sea ice concentration threshold was used. Connected components The region is divided into peripheral low-density ice zones. ( ) and marginal ice zone ( Low-density ice zone at the edge. It includes open ocean, a small amount of marginal ice zone, and a few interglacial lakes. Marginal Ice Zone It mainly consists of marginal ice and a few interglacial lakes.

[0068] Finally, for the low-density ice areas at the edges After binarization, all four-directionally connected pixels are labeled as the same connected region. All connected region attributes are then statistically analyzed, including sequence number, region pixel index, region centroid index, and total number of pixels. At this point, the larger independent connected regions are mainly located in the North Pacific, North Atlantic, Baltic Sea, Hudson Bay, and Sea of ​​Okhotsk. Based on the location and extent of the connected regions, the large-scale connected regions located in these areas are designated as marginal ice-water mixed growth zones. This area is primarily characterized by a large expanse of open water and a small amount of marginal ice. The remaining area is designated as the interglacial lake growth zone. This area does not include large areas of open water; it mainly consists of a few interglacial lakes near the edge of the ice sheet. Marginal Glacial-Water Mixture Zone and interglacial lake growth area Each contains multiple sub-blocks.

[0069] Among them, based on 6.25km spatial resolution data from 2012 to the present. Statistical analysis of region segmentation results, including the establishment of a marginal ice-water mixed growth zone. The minimum number of pixels in a neutron block is 4000.

[0070] Step S3: Determine the initial growth nuclei of the interglacial lake growth zone based on the boundary of the interglacial lake growth zone, and determine the initial growth nuclei of the edge interglacial water-mixed growth zone based on the boundary of the edge interglacial water-mixed growth zone;

[0071] In one possible implementation, a first label value is assigned to the interglacial lake growth zone, and a second label value is assigned to the marginal ice-water mixed growth zone. The interglacial lake growth zone and the marginal ice-water mixed growth zone are binarized respectively, and a region boundary tracing algorithm is used to obtain the boundaries of all sub-regions of the interglacial lake growth zone and the boundaries of all sub-regions of the marginal ice-water mixed growth zone. The boundaries of all sub-regions of the interglacial lake growth zone are used as the initial growth kernel of the interglacial lake growth zone, and the boundaries of all sub-regions of the marginal ice-water mixed growth zone are used as the initial growth kernel of the ice-water mixed growth zone.

[0072] The initial growth kernel contains the initial point at the start of the region growth algorithm, which is the set of pixels representing the region boundary.

[0073] In one example, the interglacial lake growth region and edge ice-water mixed growth zone Assign a tag value to each and After binarizing the interglacial lake growth region, the interglacial lake growth region was retrospectively analyzed. The boundaries of each sub-region are used as growth nuclei for interglacial lake regions. After binarizing the marginal ice-water mixed growth region, the marginal ice-water mixed growth region is traced. The boundaries of each subregion within the region serve as the growth nuclei for the marginal ice-water mixed growth zone.

[0074] Establish interglacial lake growth zone have Each sub-block has a growth kernel. , , These are the row and column numbers of the pixels, respectively; edge ice-water mixed growth zone. have Each sub-block has a growth kernel. , , These are the row and column numbers of the pixels, respectively.

[0075] Step S4: Pixels that meet the preset rules in the neighborhood of the initial growth kernel of the interglacial lake growth area and the edge ice-water mixed growth area are respectively taken as the corresponding growth points, and the growth points are used as new growth kernels for iteration until the range of the interglacial lake growth area and the edge ice-water mixed growth area is stable, and the iteration stops to obtain the middle interglacial lake area and the middle ice-water mixed area.

[0076] During this stage, the increase in pixels in the interglacial lake growth zone and the marginal glacial-water mixed growth zone all originated from the marginal ice zone. .

[0077] In one possible implementation, the non-terrestrial regions in the eight-neighborhood of the initial growth nucleus in the interglacial lake growth zone are considered. Non-glacial lakes Non-dense ice areas The pixels are selected as the first undetermined growth points; the density data of the undetermined growth points is not less than that of the growth kernel and there are no edge ice-water mixed growth regions in the eight neighborhoods. The first undetermined growth point was identified as an interglacial lake growth point and assigned an interglacial lake marker value. This completes the current growth cycle. The interglacial lake growth point is used as the new growth nucleus, and the next iteration continues until the extent of the interglacial lake growth region stops increasing. This results in a preliminary interglacial lake region, and the pixels in this region are assigned a third label value. .

[0078] Let the growth point of the interglacial lake be ( , , ), which satisfies ,and ,and ,and And the neighboring pixels of this growth point ( , , ).

[0079] The non-terrestrial regions in the eight neighborhoods of the initial growth nucleus in the marginal ice-water mixed growth zone Non-glacial lake areas Non-dense ice areas The pixels were used as the second undetermined growth point; the eight neighboring regions excluding the initial interglacial lake area were considered. The second undetermined growth point was identified as a new growth nucleus in the marginal ice-water mixed growth zone, and was assigned the marginal ice-water mixed zone label value. This process completes the current growth cycle and proceeds to the next iteration until the extent of the edge ice-water mixed growth zone stops increasing, at which point the iteration stops, resulting in the initial ice-water mixed zone.

[0080] Let the growth point of the edge ice-water mixed growth zone be... ( , , ), which satisfies ,and ,and And the neighboring pixels of this growth point ( , , ).

[0081] According to the edge ice zone Connectivity analysis was performed on unclassified pixels not classified as part of the initial interglacial lake region and pixels in the initial ice-water mixing region. Pixels whose eight-neighbor regions contain elements of the initial ice-water mixing region were added to the initial ice-water mixing region to obtain the intermediate ice-water mixing region. And assign a fourth tag value to the pixels in that area. The remaining unclassified pixels are added to the initial interglacial lake area to obtain the intermediate interglacial lake area. .

[0082] Step S5 involves performing connectivity analysis on the central discrete low-density ice zone, and combining this with the intermediate interglacial lake zone and the intermediate ice-water mixing zone to output the extent of interglacial lakes, marginal ice zones, and open water zones across the entire Arctic.

[0083] In one possible implementation, connectivity analysis is performed sequentially on the pixels of the central discrete low-density ice zone to determine that the pixels in the eight-neighborhood of the central discrete low-density ice zone that include the intermediate ice-water mixing zone are the first pixels, and the rest are the second pixels. The first pixels are added to the intermediate ice-water mixing zone to obtain the ice-water mixing zone, and the second pixels are added to the intermediate interglacial lake zone to obtain the interglacial lake range. The ice-water mixing zone is then threshold-filtered, and the ice-water mixing areas with sea ice concentration data greater than 0% are classified as edge ice ranges, and the remaining ice-water mixing areas are classified as open water ranges.

[0084] Among them, the interglacial lake range is a low-density ice area surrounded by sea ice or surrounded by both sea ice and land; the sea ice density data in the interglacial lake growth area is not greater than the first preset sea ice density threshold; the marginal ice range is the sea ice edge area connected to open water on one side; the sea ice density data in the marginal ice area is not greater than the second preset sea ice density threshold; the open water range refers to the open ocean water with a density value of 0%.

[0085] In one example, connectivity is determined sequentially for pixels in the central discrete low-density ice zone. The eight neighboring regions contain... The pixels are classified as Otherwise, it is classified as At this point, the extraction of the ice lake area is complete; subsequently, [further details will be provided]. Threshold-based screening is performed on the regions; areas with sea ice concentration greater than 0% are classified as marginal ice zones. Otherwise, it is classified as an open water area. . Figure 3 A schematic diagram illustrating the extraction of interglacial lake extent provided for an embodiment of the present invention.

[0086] The method for extracting the extent of Arctic interglacial lakes provided by this invention has the following beneficial effects:

[0087] 1) This invention addresses the problem of confusion between interglacial lakes and edge ice areas, as well as the problem of pre-defining the growth area of ​​interglacial lakes. It utilizes a step-by-step partitioning and refinement processing strategy to formulate pixel-level region growth rules, thereby achieving automatic extraction of the true shape and extent of interglacial lakes without limitations.

[0088] 2) This invention enables the automatic extraction of data from all interglacial lakes in the entire Arctic;

[0089] 3) This invention realizes the automatic separation and extraction technology between the edge ice zone and the interglacial lake growth zone.

[0090] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for extracting the extent of interglacial lakes in the Arctic, characterized in that, include: Step S1: Obtain the original sea ice concentration data of Arctic sea ice, and supplement the missing data in the original sea ice concentration data to obtain the sea ice concentration data. Step S2: Perform threshold segmentation and connected component analysis on the sea ice concentration data to obtain sea ice concentration partitions; The sea ice concentration zones include: a central discrete low-density ice zone, a marginal ice zone, an interglacial lake growth zone, and a marginal ice-water mixed growth zone. Step S3: Determine the initial growth nucleus of the interglacial lake growth zone based on the boundary of the interglacial lake growth zone, and determine the initial growth nucleus of the edge ice-water mixed growth zone based on the boundary of the edge ice-water mixed growth zone; Step S4: Pixels that meet the preset rules in the neighborhood of the initial growth kernel of the interglacial lake growth area and the edge ice-water mixed growth area are respectively taken as the corresponding growth points, and the growth points are used as new growth kernels for iteration until the range of the interglacial lake growth area and the edge ice-water mixed growth area is stable, and the iteration stops to obtain the intermediate interglacial lake area and the intermediate ice-water mixed area. Step S5: Perform connectivity analysis on the central discrete low-density ice zone, and combine the intermediate interglacial lake zone and the intermediate ice-water mixing zone to output the range of interglacial lakes, edge ice range and open water range of the entire Arctic.

2. The method for extracting the extent of Arctic interglacial lakes according to claim 1, characterized in that, Step S1 includes: Obtain an ocean mask; The missing ocean data in the original sea ice concentration data is assigned a value of 0 using the ocean mask. The remaining missing data in the original sea ice concentration data are assigned as the first data; the time corresponding to the original sea ice concentration data is the first time; the first data is the interpolation of the sea ice concentration data of two second times adjacent to the first time.

3. The method for extracting the extent of Arctic interglacial lakes according to claim 1, characterized in that, Step S2 includes: The sea ice concentration data is segmented based on a first preset sea ice concentration threshold to obtain dense ice areas where the sea ice concentration data is greater than the first preset sea ice concentration threshold and low-density ice areas where the sea ice concentration data is not greater than the first preset sea ice concentration threshold. Connectivity analysis was performed on the low-density ice region to obtain the two largest connected regions and the central discrete low-density ice region. The connected region is segmented based on a second preset sea ice concentration threshold to obtain the edge ice zone that does not contain open water and the edge low-density ice zone that contains open water; the second preset sea ice concentration threshold is less than the first preset sea ice concentration threshold. Connectivity analysis was performed on the edge low-density ice zone to obtain the edge ice-water mixed growth zone and the interglacial lake growth zone that is not connected to open water and is close to the edge ice zone.

4. The method for extracting the extent of Arctic interglacial lakes according to claim 1, characterized in that, Step S3 includes: The interglacial lake growth area and the edge ice-water mixed growth area are binarized respectively, and the boundaries of all sub-regions of the interglacial lake growth area and the boundaries of all sub-regions of the edge ice-water mixed growth area are obtained by a region boundary tracking algorithm. The boundaries of all sub-regions of the interglacial lake growth zone are used as the growth nuclei of the interglacial-water mixed growth zone, and the boundaries of all sub-regions of the marginal interglacial-water mixed growth zone are used as the growth nuclei of the interglacial lake growth zone.

5. The method for extracting the extent of Arctic interglacial lakes according to claim 1, characterized in that, The growth kernel is a set of pixels representing the region boundary.

6. The method for extracting the extent of Arctic interglacial lakes according to claim 1, characterized in that, Step S4 includes: The pixels in the eight neighborhoods of the initial growth nucleus of the interglacial lake growth region that are not land, not interglacial lake, and not densely ice-covered areas are taken as the first undetermined growth points. The first undetermined growth point, whose sea ice concentration data is not less than that of the growth nucleus and whose eight neighboring regions do not contain open water, is determined as the new growth nucleus of the interglacial lake growth area. The next iteration is then carried out until the range of the interglacial lake growth area stops growing, at which point the iteration stops, and the preliminary interglacial lake area is obtained.

7. The method for extracting the extent of Arctic interglacial lakes according to claim 6, characterized in that, Step S4 further includes: The pixels in the eight neighborhoods of the initial growth kernel of the edge ice-water mixed growth zone that are not land, not interglacial lakes, and not dense ice areas are used as the second undetermined growth points. The second undetermined growth point in the eight neighboring regions, excluding the initial interglacial lake region, is determined as the new growth nucleus of the marginal ice-water mixed growth region, and the next iteration is performed until the range of the marginal ice-water mixed growth region no longer increases, at which point the iteration stops, and the initial ice-water mixed region is obtained. Based on the unclassified pixels in the edge ice zone that are not classified as preliminary interglacial lake areas and the pixels in the preliminary ice-water mixing area, a connected component analysis is performed. Pixels in the eight-neighborhood of the unclassified pixels that contain the preliminary ice-water mixing area are added to the preliminary ice-water mixing area to obtain the intermediate ice-water mixing area; the remaining unclassified pixels are added to the preliminary interglacial lake area to obtain the intermediate interglacial lake area.

8. The method for extracting the extent of Arctic interglacial lakes according to claim 1, characterized in that, Step S5 includes: Connectivity analysis is performed sequentially on the pixels of the central discrete low-density ice zone to determine that the pixels in the eight neighborhoods of the central discrete low-density ice zone including the middle ice-water mixing zone are the first pixels, and the rest are the second pixels. Adding the first pixel to the intermediate ice-water mixing zone yields the ice-water mixing zone, and adding the second pixel to the intermediate interglacial lake zone yields the interglacial lake range.

9. The method for extracting the extent of Arctic interglacial lakes according to claim 8, characterized in that, Step S5 further includes: Threshold screening is performed on the ice-water mixing zone. The ice-water mixing zone with sea ice concentration data greater than 0% is classified as the edge ice range, and the remaining ice-water mixing zone is classified as the open water range.

Images