A method and system for monitoring changes in remote sensing information of sea areas and islands

Abstract

The application relates to a sea area and island remote sensing information change monitoring method and system, and relates to the technical field of remote sensing monitoring, and comprises the following steps: acquiring remote sensing image time series of a target sea area and island, the remote sensing image time series comprising a plurality of remote sensing images, and each remote sensing image having a collection time stamp; collecting tide height information of the target sea area and island corresponding to each collection time stamp and water depth distribution information of a monitoring sea area where the target sea area and island are located; and correcting and processing each remote sensing image according to the tide height information corresponding to each collection time stamp and the water depth distribution information, to generate island boundary profile change time series. The application solves the problem that traditional sea area and island remote sensing monitoring uses a single moment image for analysis, cannot distinguish between real coastline changes and temporary boundary shifts caused by tides, and is interfered by shallow water area seabed reflection, thereby causing boundary extraction precision to decrease, causing large monitoring errors, and failing to meet the application precision requirements of sea area management, coastal zone protection and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of remote sensing monitoring, in particular to a sea area and island remote sensing information change monitoring method and system. BACKGROUND

[0002] With the deepening of sea area management, coastal protection and other work, the sea area and island remote sensing information change monitoring precision has become a key technical problem.

[0003] At present, the traditional monitoring method cannot distinguish the temporary shift of the island visual boundary caused by the tide from the real coastline change, and is disturbed by the shallow water bottom reflection, which will cause the boundary extraction precision to decrease greatly. At the same time, it mainly relies on single moment image for analysis, and lacks consistency benchmark of time sequence monitoring. Not only does it reduce the utilization value of island remote sensing monitoring data, but also increases the complexity and cost of subsequent sea area management and resource investigation, resulting in large monitoring result error, which cannot meet the precision requirement of actual application. SUMMARY

[0004] In order to solve the above technical problems, the present application provides a sea area and island remote sensing information change monitoring method and system, which improves the present situation that the monitoring error is large and it is difficult to meet the precision requirement of actual application due to the fact that the boundary shift caused by the tide cannot be distinguished, the shallow water reflection disturbance and the lack of time sequence benchmark in the traditional monitoring.

[0005] The embodiments of the present application disclose the following technical solutions:

[0006] In a first aspect, the embodiments of the present application provide a sea area and island remote sensing information change monitoring method, which comprises:

[0007] acquiring a remote sensing image time sequence of a target sea area and island, wherein the remote sensing image time sequence comprises a plurality of remote sensing images, and each remote sensing image has a collection time stamp;

[0008] collecting tide height information of the target sea area and island corresponding to each collection time stamp, and water depth distribution information of a monitoring sea area where the target sea area and island are located;

[0009] correcting and processing each remote sensing image according to the tide height information corresponding to each collection time stamp and the water depth distribution information, to generate an island boundary profile change time sequence.

[0010] In a second aspect, the embodiments of the present application provide a sea area and island remote sensing information change monitoring system, which comprises:

[0011] an image time sequence acquisition module, configured to acquire a remote sensing image time sequence of a target sea area and island, wherein the remote sensing image time sequence comprises a plurality of remote sensing images, and each remote sensing image has a collection time stamp;

[0012] a time-series tidal level water depth collection module, configured to collect tidal height information corresponding to each of the collection time stamps of the target sea area islands, and water depth distribution information of a monitoring sea area where the target sea area islands are located;

[0013] a correction and time-series generation module, configured to perform correction processing on each of the remote sensing images according to the tidal height information corresponding to each of the collection time stamps and the water depth distribution information, and generate an island boundary profile change time series.

[0014] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0015] The present application provides a remote sensing information change monitoring method and system for sea area islands, which realizes accurate monitoring of remote sensing information changes of sea area islands through the cooperative operation of obtaining remote sensing image time series and environmental parameters, correcting remote sensing images by region, extracting boundary segments and splicing to generate island boundary profiles, and constructing island boundary profile change time series. First, the remote sensing image time series of the target sea area islands and the corresponding collection time stamps are obtained, and the tidal height information and the water depth distribution information of the monitoring sea area corresponding to each time stamp are collected synchronously. Then, the actual water depth distribution is calculated based on these environmental parameters, and the deep water and shallow water influence areas are divided accordingly. The first boundary segment is directly extracted in the deep water influence area, and the second boundary segment is extracted after the sea bottom reflection correction in the shallow water influence area. After splicing, a single-time island boundary profile is formed. Finally, all the profiles are arranged in chronological order to generate an island boundary profile change time series, and dynamic monitoring of island changes is realized.

[0016] The technical solutions of the present application solve the problems of large error and low precision caused by tidal offset, shallow water reflection interference and missing time series reference in traditional island monitoring through the steps of correction processing by fusing remote sensing image time series and environmental parameters, extracting boundaries by region, constructing and quantitative analyzing time series profiles, and realize high-precision dynamic monitoring of island boundary changes. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0018] Figure 1 A flowchart of a remote sensing information change monitoring method for sea area islands provided by an embodiment of the present application is shown in the figure.

[0019] Figure 2 A structural diagram of a remote sensing information change monitoring system for sea area islands provided by an embodiment of the present application is shown in the figure.

[0020] In the drawings, the components represented by the respective reference numerals are explained as follows:

[0021] Image time sequence acquisition module 01, time sequence tide water depth collection module 02, correction and time sequence generation module 03. DETAILED DESCRIPTION

[0022] The application provides a sea area and island remote sensing information change monitoring method and system, which is used for solving the technical problems that in the prior art, it is difficult to distinguish temporary deviation of island visual boundary and real coastline change due to tide influence, boundary extraction accuracy is reduced due to interference of shallow water area seabed reflection, and single time image analysis lacks time sequence monitoring consistency reference, thereby causing large error of island monitoring result and failing to meet application precision requirements such as sea area management.

[0023] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the application.

[0024] In the description of the application, the terms "first", "second" are used only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first", "second" can explicitly or implicitly include one or more of the features. In the description of the application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0025] In the description of the application, the term "for example" is used to indicate "as an example, illustration or explanation". Any embodiment described as "for example" in the application is not necessarily interpreted as more preferred or more advantageous than other embodiments. The following description is given in order to enable any person skilled in the art to implement and use the application. In the following description, details are listed for the purpose of explanation. It should be understood that those skilled in the art can realize the application without using these specific details. In other examples, well-known structures and processes will not be described in detail to avoid unnecessary details making the description of the application obscure. Therefore, the application is not intended to be limited to the shown embodiments, but is consistent with the broadest scope of principles and features disclosed in the application.

[0026] Embodiment one, as shown in the accompanying Figure 1 The application provides a sea area and island remote sensing information change monitoring method, which comprises the following steps:

[0027] S110: Obtain a remote sensing image time sequence of the target sea island, the remote sensing image time sequence comprising a plurality of remote sensing images, each remote sensing image having a collection time stamp;

[0028] In the embodiment of the present application, in the scenario of sea island remote sensing information change monitoring, in order to obtain continuous and accurate remote sensing image time sequence data, the image library is associated with the island identifier and filtered according to the time parameter, so as to establish the basic data sequence for time sequence analysis.

[0029] Specifically, first, the unique island identifier of the target sea island is obtained through the preset geographic information database, and the corresponding remote sensing image library is retrieved by taking the island identifier as an index, so as to ensure the accurate matching of data source and detection object.

[0030] At the same time, the user-set collection time interval and monitoring time period are received as the time reference for image filtering. The monitoring time period includes a start time and an end time.

[0031] Further, a plurality of remote sensing images meeting the conditions are filtered from the remote sensing image library according to the above-mentioned time parameters, and the collection time of each remote sensing image is extracted as a time stamp, forming a remote sensing image time sequence containing multi-time data.

[0032] This step provides consistent basic image data for subsequent remote sensing image correction combined with tide and water depth data by accurately associating the island identifier with the remote sensing image library and combining image filtering based on the time reference, thereby ensuring the continuity and accuracy of remote sensing information change monitoring.

[0033] The step S110 in the method provided by the embodiment of the present application comprises:

[0034] Obtain an island identifier of a target sea island, and retrieve a corresponding remote sensing image library based on the island identifier;

[0035] Receive a user-set collection time interval and monitoring time period, wherein the monitoring time period includes a start time and an end time;

[0036] According to the collection time interval, the start time and the end time, filter a plurality of remote sensing images from the remote sensing image library, and synchronously extract the collection time of each remote sensing image as the collection time stamp of each remote sensing image.

[0037] In the embodiment of the present application, in order to obtain remote sensing image time sequence meeting the monitoring requirements, the remote sensing image library is accurately associated with the island identifier, and the data is filtered in combination with the user-set time parameter, so as to establish the basic image sequence for time sequence analysis, and ensure the continuity and accuracy of subsequent correction processing.

[0038] Specifically, first, the island identifier of the target sea island in the target sea area is obtained through a preset geographic information database. The identifier is a unique number, geographic coordinate code, etc. of the island, and is the core index associated with the remote sensing image library.

[0039] Illustratively, the island identifier of a certain target sea island in a target sea area is "HD-2023-015". Through the identifier, all remote sensing images related to the island in the remote sensing image library can be directly retrieved, ensuring that the data source strictly matches the monitoring object and avoiding errors caused by confusion of island information.

[0040] The remote sensing image library is pre-constructed based on the association mapping relationship between the island identifier and the remote sensing image. The library stores remote sensing images of the island and the surrounding sea area collected by different satellites and different sensors at different times, and classifies and indexes them according to collection time, image resolution, etc.

[0041] Further, the collection time interval and the monitoring time period set by the user are received. The collection time interval can be set according to the monitoring accuracy requirement, such as every 3 days, every 1 week, or every month.

[0042] Secondly, the monitoring time period needs to specify the start time and the end time, for example, "January 1, 2024 to December 31, 2024". These time parameters are the key basis for filtering remote sensing images, and directly determine the time density and coverage range of the remote sensing image time sequence.

[0043] Illustratively, if the user sets the collection time interval to 10 days and the monitoring time period to March 1, 2024 to June 1, 2024, then the remote sensing images collected at time points such as March 1, March 11, March 21, April 1, etc. within the monitoring time period and the collection time interval of about 10 days need to be filtered from the remote sensing image library to ensure the regularity of the time sequence.

[0044] Further, based on the set collection time interval, monitoring start time and end time, multiple remote sensing images that meet the time parameter conditions are filtered from the retrieved remote sensing image library.

[0045] Specifically, in the image filtering process, first, the time retrieval interface of the remote sensing image library is called, and the monitoring start time and end time are input to preliminarily filter all remote sensing images within the monitoring time period.

[0046] Further, according to the set collection time interval, the preliminary filtering result is filtered again, for example, if the collection time interval is 7 days, then starting from the start time, images are selected at an interval of every 7 days to ensure that the images are evenly distributed in time and meet the monitoring frequency requirement.

[0047] Meanwhile, the integrity and availability of the remote sensing image are verified synchronously in the screening process, and the images with serious noise, cloud coverage exceeding the threshold or data damage are removed, and finally multiple remote sensing images meeting the conditions are retained, and the collection time of each image is extracted as a timestamp to arrange the continuous remote sensing image time sequence in chronological order.

[0048] For example, if the user sets the monitoring time period as May 1, 2024 to August 1, 2024 and the collection time interval is 15 days. First, the remote sensing images in the monitoring time period are retrieved from the remote sensing image library, and then the image on May 16, 2024, which is fully covered by clouds due to heavy rain, is removed after verification, and finally six effective images on May 1, 2024, May 31, 2024, June 15, 2024, June 30, 2024, July 15, 2024 and August 1, 2024 are screened out, and the timestamps of the images correspond to the collection time of each image, and the images are arranged in order to form a complete remote sensing image time sequence.

[0049] Finally, the remote sensing image time sequence obtained through the above steps contains multiple effective remote sensing images with accurate timestamps, which lays a solid data foundation for subsequent correction processing combined with tidal height and water depth distribution information to generate the island boundary profile change time sequence, and guarantees the time sequence consistency and data reliability of the entire monitoring process.

[0050] S120: Collecting tidal height information of the target sea island at each collection timestamp and water depth distribution information of the monitoring sea area where the target sea island is located;

[0051] In the embodiment of the present application, in the scenario of sea island remote sensing information change monitoring, in order to accurately obtain tidal and water depth data for image correction, the monitoring range needs to be determined in combination with island features and remote sensing image attributes, and then key information is collected for subsequent remote sensing correction processing to provide data support.

[0052] Specifically, first, the island feature information of the target sea island and the remote sensing image resolution of each remote sensing image are obtained.

[0053] The island feature information includes island coordinate position and island area, which is the basis for positioning the monitoring area and calculating the buffer range; the remote sensing image resolution directly affects the setting of the buffer distance to ensure that the collected water depth data is compatible with the accuracy of the remote sensing image.

[0054] Further, the buffer range is determined according to the island area and the remote sensing image resolution. This step quantitatively calculates the boundary of the monitoring sea area, which not only ensures that the area around the island that may be affected by the tide and water depth is covered, but also avoids data redundancy caused by an excessively large range.

[0055] On this basis, the tidal height information corresponding to each collection timestamp is called from the tidal database based on the island coordinate position, to ensure that the water level state of each time point can be accurately matched, and to provide data basis for eliminating the boundary offset caused by the tide.

[0056] At the same time, the monitoring sea area is determined in combination with the island coordinate position and the buffer range, the water depth distribution information of the monitoring sea area is obtained from the marine sounding database, and the foundation is laid for distinguishing deep water and shallow water areas and processing seabed reflection interference.

[0057] This step accurately obtains the key environmental parameters by systematically collecting the tidal height information and the water depth distribution information, provides reliable data support for subsequent elimination of tidal and shallow water reflection interference and improvement of boundary extraction precision, and guarantees the accuracy of the monitoring result.

[0058] The step S120 in the method provided in the embodiment of the application comprises:

[0059] Obtaining island feature information of the islands in the target sea area and remote sensing image resolutions of the remote sensing images, the island feature information comprising island coordinate positions and island areas;

[0060] Determining a buffer range according to the island areas and the remote sensing image resolutions;

[0061] Obtaining tidal height information corresponding to each collection timestamp based on the island coordinate positions;

[0062] Determining a monitoring sea area based on the island coordinate positions and the buffer range, and obtaining water depth distribution information of the monitoring sea area.

[0063] In the embodiment of the application, in order to accurately collect the tidal height and the water depth distribution information for remote sensing image correction, the spatial attributes and image precision features of the islands need to be determined first, to provide basis for demarcating a reasonable monitoring sea area range, to ensure the pertinence and effectiveness of subsequent data collection.

[0064] Specifically, first, the island feature information of the islands in the target sea area and the remote sensing image resolutions of the remote sensing images are obtained.

[0065] The coordinate position in the island feature information is a spatial reference for positioning the tidal data and the monitoring sea area, for example, the coordinate of an island is east longitude 120°30', north latitude 30°15', so the position of the island in the geographical space can be accurately locked.

[0066] In addition, the island area reflects the size of the island and is a basic parameter for calculating the buffer range. The remote sensing image resolution determines the ability of the remote sensing image to depict the details of the ground, and directly affects the setting of the buffer distance, to avoid data redundancy caused by too large buffer range or information loss caused by too small buffer range.

[0067] Further, the buffer range is determined according to the island area and the resolution of the remote sensing image, so as to accurately delineate the boundary of the monitoring sea area which needs to collect the water depth distribution information.

[0068] In the method provided by the embodiment of the application, the step of "determining the buffer range according to the island area and the resolution of the remote sensing image" comprises:

[0069] calculating the equivalent radius of the island based on the island area, and obtaining a first buffer distance in a buffer distance mapping table based on the equivalent radius of the island;

[0070] determining a second buffer distance based on the resolution of the remote sensing image, the second buffer distance being a preset multiple of the resolution of the remote sensing image;

[0071] comparing the first buffer distance with the second buffer distance, and taking the larger value as a basic buffer radius;

[0072] increasing a safety margin on the basis of the basic buffer radius to obtain the buffer range.

[0073] In the embodiment of the application, in order to scientifically delineate the buffer range which is suitable for the scale of the island and meets the accuracy of the remote sensing image, multi-dimensional calculation needs to be performed in combination with the island area and the resolution of the remote sensing image, so as to accurately delineate the monitoring sea area and provide a reasonable spatial boundary for the effective collection of the water depth distribution information.

[0074] Specifically, first, the equivalent radius of the island is calculated based on the island area. That is, the irregularly shaped island is converted into a standardized circular scale parameter, so as to facilitate the unified calculation of the buffer distance.

[0075] For example, if the area of an island is 1256 square meters (π is 3.14), according to the formula of the area of a circle , the corresponding equivalent radius of the island can be deduced as meters.

[0076] Further, the equivalent radius of the island is used as an index to match a first buffer distance in a preset buffer distance mapping table.

[0077] In the buffer distance mapping table, a standard buffer distance corresponding to different equivalent radii is pre-stored, and the value thereof is set based on the typical range statistics of the island periphery affected by the tide and water depth, for example, the first buffer distance of 100 meters corresponding to the equivalent radius of 20 meters, so as to ensure that the buffer distance matches the actual scale of the island.

[0078] Further, the second buffer distance is determined based on the resolution of the remote sensing image. The second buffer distance is set as a preset multiple (e.g., 300 times) of the resolution of the remote sensing image. This design aims to ensure that the buffer range covers the area corresponding to the minimum spatial details identifiable from the image.

[0079] For example, if the resolution of the remote sensing image is 5 meters, the second buffer distance is 5 x 300 = 1500 meters. This effectively avoids the problem of the buffer range being too small due to the precision limitation of the remote sensing image, thereby missing the key geomorphological information of the shallow water area.

[0080] Further, the first buffer distance and the second buffer distance are compared, and the larger value is taken as the basic buffer radius to balance the island scale and the precision of the remote sensing image.

[0081] Specifically, when the island area is small but the resolution of the remote sensing image is high, the second buffer distance corresponding to the resolution of the remote sensing image is used as the main factor to ensure that enough surrounding sea area is covered. When the island area is large but the resolution of the remote sensing image is low, the first buffer distance corresponding to the island scale is used as the main factor to avoid data redundancy caused by a too large range. For example, if the first buffer distance is 100 meters and the second buffer distance is 1500 meters, the basic buffer radius is taken as 1500 meters.

[0082] Finally, a safety margin is added to the basic buffer radius to obtain the final buffer range. The safety margin is set to cope with uncertain factors such as tidal extremes and measurement errors, to ensure that the area affecting the identification of the island boundary is completely covered in extreme cases.

[0083] For example, if 50 meters is taken as the safety margin, based on the basic buffer radius of 1500 meters, combined with the safety margin of 50 meters, the final buffer range is 1550 meters (1500 + 50 = 1550).

[0084] Further, after determining the buffer range, the tidal height information corresponding to each collection timestamp is accurately retrieved from the tidal database based on the island coordinate position.

[0085] Since the tidal height has significant spatiotemporal specificity, the water level change at different times in the same sea area will directly affect the visual boundary of the island, therefore, the island coordinate needs to be locked to the tidal monitoring station coverage area to which it belongs, and then the tidal height information at the corresponding time is extracted according to the collection timestamp of each remote sensing image.

[0086] Exemplarily, the tidal monitoring station record corresponding to the island coordinate shows that the tidal height is +0.8 meters at the remote sensing image timestamp "2024-07-10-08:30", which will be used as a key basis for subsequent correction of the boundary offset of the remote sensing image caused by tides to ensure that the remote sensing images of different timestamps remain consistent in the water level reference.

[0087] Meanwhile, in combination with the island coordinate position and the determined buffer range, the spatial boundary of the monitored sea area can be accurately delimited. That is, the circular area formed with the island coordinate as the center and the buffer range as the radius is the monitored sea area for which the water depth distribution information needs to be collected.

[0088] Further, the water depth distribution information of the monitored sea area is retrieved from the marine sounding database. These information usually contains the water depth values of different coordinate points in the sea area, which can clearly reflect the water depth gradient change from the island periphery to the buffer boundary.

[0089] Exemplarily, the water depth values corresponding to some point coordinates in the monitored sea area are 3 meters and 10 meters respectively, which can be used to distinguish the shallow water area (such as water depth ≤5 meters) and the deep water area (such as water depth >5 meters), to provide basic data for subsequent differentiated remote sensing image correction strategies for different water depth areas, effectively improving the accuracy of island boundary extraction.

[0090] This step realizes the accurate definition of different water depth areas by systematically obtaining the water depth distribution information of the monitored sea area, lays a data foundation for subsequent use of different correction strategies for shallow and deep water areas respectively, and helps to eliminate interference factors such as seabed reflection, thereby improving the accuracy of island boundary extraction.

[0091] S130: Correcting each remote sensing image according to the tidal height information and the water depth distribution information corresponding to each collection timestamp to generate an island boundary contour change time sequence.

[0092] In the embodiments of the present application, in the scenario of sea island remote sensing information change monitoring, in order to eliminate the interference of tides and shallow water reflection on remote sensing images, accurate correction needs to be made based on the time-stamped environmental parameters, and a time-consistent island boundary contour is constructed through regional processing to provide a reliable reference for dynamic change monitoring.

[0093] Specifically, first, the actual water depth distribution at each time is calculated based on the water depth distribution information and the tidal height information corresponding to each collection timestamp.

[0094] Among them, the water depth distribution information reflects the basic topographic features of the monitored sea area, while the tidal height information reflects the water level fluctuation at different time points. The combination of the two can obtain the actual submerged depth state of seawater at each timestamp.

[0095] Further, the remote sensing images are corrected according to the actual water depth distribution corresponding to each acquisition timestamp. That is, the image pixels are associated with the actual water depth to eliminate environmental interference in a targeted manner.

[0096] For deep water regions where the actual water depth exceeds the reflection threshold, the water-land separation threshold is directly used to divide the boundary; for shallow water regions where the actual water depth is within the reflection interference range, the influence of seabed reflection on the image pixel value needs to be eliminated by the water depth pixel corrector before the island boundary contour is extracted.

[0097] On this basis, each remote sensing image in the remote sensing image time sequence is traversed, the actual water depth distribution corresponding to the timestamp is extracted, and the deep water influence region and the shallow water influence region are determined.

[0098] Meanwhile, the first boundary segment is extracted in the deep water region, and the second boundary segment after reflection correction is extracted in the shallow water region, and a complete island boundary contour at a single time is formed by spatial splicing. All time sequence images are processed in this way, and finally the island boundary contour change time sequence is generated.

[0099] This step realizes the dual goals of tide correction and water depth correction by fusing environmental parameters with image processing depth, provides time-sequential data for accurately capturing the dynamic changes of the real island boundary, and effectively supports the demand for monitoring accuracy in sea area management scenarios.

[0100] The step S130 in the method provided in the embodiment of the present application includes:

[0101] Based on the water depth distribution information and the tide height information corresponding to each acquisition timestamp, the actual water depth distribution corresponding to each acquisition timestamp is calculated.

[0102] According to the actual water depth distribution corresponding to each acquisition timestamp, the remote sensing images are corrected to obtain the island boundary contour corresponding to each acquisition timestamp, and the island boundary contour change time sequence is generated.

[0103] In the embodiment of the present application, in order to accurately capture the real boundary changes of the island, solve the problems of boundary deviation caused by tides, shallow reflection interference, and missing time sequence reference, etc., the remote sensing images need to be accurately corrected, and then a time-sequential island boundary contour change sequence is constructed to eliminate the influence of environmental interference on boundary recognition, and provide a reliable basis for dynamic monitoring.

[0104] First, based on the water depth distribution information and the tide height information corresponding to each acquisition timestamp, the actual water depth distribution corresponding to each acquisition timestamp is calculated.

[0105] Among them, the water depth distribution information reflects the basic topographic water depth characteristics of the monitored sea area, for example, the reference water depth of a certain area is 4 meters; and the tidal height information embodies the water level fluctuation at different time points, for example, the tidal height at a certain time stamp is +0.6 meters, and the actual water depth distribution (4+0.6=4.6 meters) can be obtained by superimposing the two, which is the core basis for distinguishing the water-land boundary and dividing the influence area.

[0106] Further, the remote sensing image is corrected according to the actual water depth distribution corresponding to each acquisition time stamp to obtain a single-time island boundary contour.

[0107] The method provided in the embodiments of the present application comprises the following steps:

[0108] Iterate through the plurality of remote sensing images in the remote sensing image time sequence to obtain a first remote sensing image;

[0109] Extract a first acquisition time stamp of the first remote sensing image, and obtain a first actual water depth distribution according to the first acquisition time stamp;

[0110] Determine a deep water influence area and a shallow water influence area in the first remote sensing image according to the first actual water depth distribution;

[0111] For the first remote sensing image, extract a first boundary segment of the target sea island in the deep water influence area, and extract a second boundary segment of the target sea island in the shallow water influence area;

[0112] Spatially splice the first boundary segment and the second boundary segment to generate an island boundary contour corresponding to the first acquisition time stamp;

[0113] Obtain the island boundary contour corresponding to each acquisition time stamp in the same manner as obtaining the island boundary contour corresponding to the first acquisition time stamp.

[0114] In the embodiments of the present application, in order to accurately extract the island boundary contour at different time stamps, the remote sensing image needs to be processed in different areas in combination with the actual water depth distribution, and the environmental interference is eliminated through a differentiated boundary extraction method to ensure the accuracy of the single-time contour and the time sequence consistency of the multi-time contour, thereby laying a foundation for generating the time sequence of island boundary contour changes.

[0115] Specifically, first, iterate through the plurality of remote sensing images in the remote sensing image time sequence, and select any one as a first remote sensing image.

[0116] Further, a first acquisition time stamp of the first remote sensing image is extracted, and a corresponding first actual water depth distribution is retrieved according to the time stamp.

[0117] wherein the first actual water depth distribution is the key data fused with the tidal height information and the basic water depth distribution information at the moment, for example, the first actual water depth distribution corresponding to a first collection timestamp of "2024-08-01-14:00" has a water depth of 6 meters in the sea island surrounding area and a water depth of 2 meters in some areas, which provides a quantitative basis for subsequent regional division.

[0118] On this basis, the deep water influence area and the shallow water influence area are determined in the first remote sensing image according to the first actual water depth distribution.

[0119] wherein the deep water influence area refers to a sea area with a relatively deep actual water depth (such as > 5 meters), and the boundary recognition difficulty is low due to small seabed reflection interference; the shallow water influence area refers to a sea area with a relatively shallow actual water depth (such as ≤ 5 meters), and the reflection of seabed topography on remote sensing signals is significant, which easily leads to blurred water-land boundary.

[0120] For example, if the first actual water depth distribution shows that the water depth of the sea area east of the island is 7 meters, it is divided into a deep water influence area; the water depth of the sea area west of the island is 3 meters, and it is divided into a shallow water influence area.

[0121] Further, the boundary segments are extracted in different regions of the first remote sensing image to realize accurate processing of different environmental interference degree regions and ensure the pertinence and accuracy of boundary extraction.

[0122] The method provided in the embodiments of the present application comprises the following steps:

[0123] obtaining a water-land separation threshold value;

[0124] in the deep water influence area, performing pixel classification processing on the first remote sensing image based on the water-land separation threshold value, and extracting the demarcation line between seawater and the target sea island to obtain the first boundary segment;

[0125] in the shallow water influence area, after performing seabed reflection correction on the first remote sensing image, combining the water-land separation threshold value to perform pixel classification processing, and extracting the demarcation line between seawater and the target sea island to obtain the second boundary segment.

[0126] In the embodiments of the present application, in order to accurately extract the island boundary segment in different water depth environments, differential processing needs to be performed in combination with the regional interference characteristics to ensure the accuracy and compatibility of the two boundary segments, thereby laying a foundation for subsequent island boundary contour splicing.

[0127] Specifically, a water-land separation threshold is first obtained. The water-land separation threshold is a pixel value threshold obtained based on a large number of sample images, and is used to distinguish the pixel features of sea water and island land.

[0128] For example, the water-land separation threshold is set to 0.55 (pixel value range 0-1), and when the pixel value of the remote sensing image is greater than 0.55, it is determined to be island land, and when the pixel value is less than 0.55, it is determined to be sea water, thereby providing a unified classification standard for boundary extraction.

[0129] Further, a first boundary segment is extracted in the deep water influence area.

[0130] In the deep water influence area, the actual water depth is deep (actual water depth > 5 meters), the sea bottom reflection has little interference on the remote sensing image, and the pixel value difference between the island and the sea water is significant (the island pixel value is concentrated in 0.6-0.9, and the sea water pixel value is concentrated in 0.2-0.4).

[0131] Therefore, the first remote sensing image is directly classified based on the water-land separation threshold to quickly and accurately divide the island land and sea water pixels in the deep water influence area, thereby providing a basis for extracting a clear first boundary segment.

[0132] Specifically, by traversing all remote sensing image pixels in the deep water influence area, the pixels with a pixel value greater than 0.55 are marked as land pixels, and the pixels with a pixel value less than 0.55 are marked as sea water pixels, and then an edge detection algorithm in the prior art is used to extract the boundary line between the two types of pixels, thereby obtaining the first boundary segment.

[0133] For example, after pixel classification in a certain deep water influence area, the pixel distribution of the island land and the sea water presents a clear continuous boundary, the first boundary segment extracted by the edge detection algorithm is complete and has no burrs, and can accurately reflect the true island boundary of the deep water influence area.

[0134] Meanwhile, a second boundary segment is extracted in the shallow water influence area. In the shallow water area, the actual water depth is shallow (actual water depth ≤ 5 meters), the sea bottom topography has a significant reflection effect on the remote sensing signal, which can cause the island land pixel value to be lowered, and directly using the water-land separation threshold can cause misclassification. Therefore, the sea bottom reflection correction processing needs to be performed on the shallow water influence area.

[0135] In the method provided by the embodiments of the present application, the step of "in the shallow water influence area, after the sea bottom reflection correction of the first remote sensing image, performing pixel classification processing in combination with the water-land separation threshold, and extracting the boundary line between the sea water and the target sea island, to obtain the second boundary segment" includes:

[0136] In the first remote sensing image, the actual water depth value and the original pixel value of each pixel point in the shallow water influence area are obtained.

[0137] correcting each pixel point in the shallow water influence area by the water depth pixel corrector to obtain a corrected pixel value, and replacing the original pixel value with the corrected pixel value;

[0138] performing pixel classification processing on the first remote sensing image based on the corrected pixel value and the water-land separation threshold, and extracting a demarcation line between seawater and islands in the target sea area to obtain the second boundary segment.

[0139] In the embodiments of the present application, in order to eliminate the interference of seabed reflection in the shallow water area on the remote sensing image and ensure the accuracy of island boundary extraction in this area, the pixel value needs to be corrected through the process of "data acquisition-reflection correction-classification extraction" before the boundary is divided, so that the water-land demarcation features of the shallow water influence area remain consistent with the judgment standard of the deep water influence area, and compatible basic data is provided for boundary segment splicing.

[0140] Specifically, first, in the first remote sensing image, the actual water depth value and the original pixel value of each pixel point in the shallow water influence area are obtained to establish the basic data for reflection correction, which provides input parameters for eliminating seabed reflection interference in the subsequent water depth pixel corrector, and ensures that the corrected pixel value can truly reflect the ground properties.

[0141] The actual water depth value is derived from the actual water depth distribution of the shallow water influence area, such as 2.1 meters and 3.7 meters, which reflects the true depth of seawater coverage.

[0142] In addition, the original pixel value is the initial gray value of the pixel point in the remote sensing image, such as 0.48 and 0.53, but due to the influence of seabed reflection, these initial gray values cannot truly distinguish islands and land from seawater. For example, the original pixel value of the island and land pixels in a certain shallow water influence area is only 0.52 due to seabed reflection interference, which is extremely small compared to the seawater pixel value of 0.49, and it is difficult to directly divide.

[0143] Therefore, in order to solve the problem that the original pixel value in the shallow water influence area cannot truly distinguish islands and land from seawater due to the interference of seabed reflection, the seabed reflection of each pixel point in the shallow water influence area needs to be corrected by the water depth pixel corrector to eliminate the pixel value distortion caused by reflection interference, and lay a foundation for subsequent accurate classification based on the water-land separation threshold.

[0144] The construction steps of the "water depth pixel corrector" in the method provided by the embodiments of the present application include:

[0145] Collecting a sample actual water depth record set and a sample original pixel set, and labeling based on the sample actual water depth record set and the sample original pixel set to obtain a sample corrected water depth record set;

[0146] The sample actual water depth record set and the sample original pixel set are input, and the sample corrected water depth record set is used as a supervisory signal to train the water depth pixel corrector.

[0147] In the embodiments of the present application, in order to construct a correction tool that can accurately eliminate the interference of seabed reflection in shallow water areas, a systematic sample collection, labeling and model training process is needed to enable the water depth pixel corrector to have the ability to derive the true ground pixel characteristics according to the actual water depth and the original pixel value.

[0148] Specifically, first, a sample actual water depth record set and a sample original pixel set are collected.

[0149] The sample actual water depth record set is measured water depth data collected by high-precision marine depth measuring equipment (such as a multi-beam depth sounder) in different shallow water affected areas, containing specific water depth values corresponding to a large number of pixel points, such as 1.2 meters, 2.8 meters, 4.5 meters, etc. These data need to cover different tidal states and seabed topography types in shallow water environments to ensure the representativeness of the samples.

[0150] In addition, the sample original pixel set is the initial gray value of the water depth sampling point in the remote sensing image, such as 0.42, 0.51, 0.55, etc., which directly reflects the original image characteristics affected by seabed reflection interference.

[0151] For example, 10,000 sample points are collected in a certain sandy shoal area, of which 5,000 are island peripheral land pixels (actual water depth is 0, covered by very shallow seawater), and 5,000 are pure seawater pixels, and their actual water depth and original pixel value are recorded to form a basic sample data set.

[0152] Further, based on the sample actual water depth record set and the sample original pixel set, a sample corrected pixel set is obtained.

[0153] The labeling process is performed by professional personnel in combination with field investigation data to correct the original pixel value to obtain the ideal pixel value after eliminating the reflection interference.

[0154] For example, a sample point is actually an island land with an actual water depth of 0.3 meters. Due to being submerged during high tide, the original pixel value is affected by seabed reflection and is low at 0.45. Referring to the pixel characteristics of similar land without reflection interference, the corrected pixel value is labeled as 0.72.

[0155] In addition, a sample point is actually pure seawater with an actual water depth of 3.2 meters. The original pixel value is affected by seabed reflection and is high at 0.53. Referring to the pixel characteristics of seawater in deep water affected areas, the corrected pixel value is labeled as 0.41.

[0156] Finally, after the labeling is completed, each sample point forms a complete data chain of "actual water depth value-original pixel value-corrected pixel value", providing supervision for model training.

[0157] On this basis, the sample actual water depth record set and the sample original pixel set are taken as inputs, and the sample corrected pixel set is taken as a supervision signal to train the water depth pixel corrector.

[0158] Specifically, a deep neural network is used as the basic framework for model training. The model structure includes an input layer that receives actual water depth values and original pixel values, a hidden layer that extracts feature correlations through multiple neurons, such as the nonlinear relationship between water depth and reflection intensity, and an output layer that outputs the final corrected pixel value.

[0159] During the training process, the model parameters are continuously optimized through the backpropagation algorithm to control the mean square error of the output corrected pixel value and the sample corrected pixel set within a preset threshold (e.g., ≤0.02).

[0160] For example, using 100,000 sample data for training, after 50 iterations, the model's correction accuracy on the validation set reaches more than 96%, i.e., the output corrected pixel value has minimal deviation from the ideal value annotated by humans after inputting the actual water depth and original pixel value, which meets the accuracy requirements for boundary extraction. At this time, the construction of the water depth pixel corrector is completed.

[0161] Finally, by constructing a training framework of "measured data-original features-ideal features", the water depth pixel corrector can accurately capture the interference patterns of seabed reflections in shallow water areas, providing an automated and high-precision correction tool for subsequent batch processing of shallow water area pixels in remote sensing images.

[0162] For example, a sample set in a certain shallow water area contains a set of data, including an actual water depth of 1.5 meters and an original pixel value of 0.48. Due to the influence of seabed sandy topography reflection, this land pixel value is pulled down to near the sea water value. After manual annotation, the corresponding sample corrected pixel value is 0.73 to match the land pixel features without reflection interference.

[0163] Further, inputting this sample into the trained water depth pixel corrector, the corrector outputs a corrected pixel value of 0.72, which is consistent with the actual annotated value, i.e., the accuracy meets the requirements.

[0164] In addition, when processing another remote sensing image with an actual water depth of 1.4 meters and an original pixel value of 0.49 in a shallow water area, the water depth pixel corrector can automatically output a corrected pixel value of 0.71, effectively eliminating reflection interference, making the pixel significantly different from the surrounding sea water pixels, and providing a reliable basis for subsequent classification processing based on the water-land separation threshold.

[0165] On this basis, the trained water depth pixel corrector is applied to the shallow water affected area of the first remote sensing image, that is, all pixel points in the shallow water affected area are traversed, the actual water depth value and the original pixel value of each pixel are obtained one by one, and the corresponding corrected pixel value is obtained after the water depth pixel corrector is input.

[0166] For example, the actual water depth of a certain pixel point is 2.1 meters, and the original pixel value is 0.48. After the water depth pixel corrector processing, the corrected pixel value is 0.65. The actual water depth of another pixel point is 3.7 meters, and the original pixel value is 0.53. The output corrected pixel value is 0.51.

[0167] Further, the original pixel value is replaced by the corrected pixel value, so that the pixel characteristics of the shallow water area are restored, for example, the land pixel is concentrated in 0.6-0.8 after correction, and the seawater pixel is concentrated in 0.4-0.5, and the difference between the two is significant.

[0168] Further, based on the corrected pixel value, the first remote sensing image is classified by combining a preset water-land separation threshold (such as 0.55).

[0169] Specifically, the corrected pixel value greater than 0.55 is marked as an island land pixel, and the corrected pixel value less than 0.55 is marked as a seawater pixel. After classification is completed, the boundary line of the two types of pixels is extracted by the edge detection algorithm in the prior art, and the boundary line is the second boundary segment after elimination of the reflection interference.

[0170] For example, after correction and classification of a certain shallow water affected area, the extracted second boundary segment is continuous and smooth in space, and the deviation from the measured coastline is controlled within 1 pixel, and the first boundary segment of the deep water affected area is naturally transitioned at the junction, providing high-precision basic data for subsequent splicing of the complete island boundary profile.

[0171] Further, the first boundary segment and the second boundary segment are spliced in space to generate an island boundary profile corresponding to the first acquisition timestamp. That is, through the coordinate alignment and smooth transition algorithm in the prior art, the slight deviation of the two boundary segments at the junction is eliminated, and the continuity and integrity of the overall profile are ensured.

[0172] Specifically, first, the end point coordinate set of the first boundary segment and the start point coordinate set of the second boundary segment are extracted, and the nearest junction point pair is found through spatial coordinate matching, for example, the end point coordinate ( , ) of the first segment and the start point coordinate ( , ) of the second segment, and the straight line distance is less than 1 pixel unit.

[0173] Further, the transition path between the connection points is fitted by using the Bezier curve in the prior art, so that the two boundary segments are naturally fused, and the edges or breaks are avoided.

[0174] For example, the end point of the first boundary segment of a deep water influence area of an island and the start point of the second boundary segment of a shallow water influence area have a 0.3 pixel offset, and after the Bezier curve fitting, the spliced boundary forms a continuous and smooth curve, and the overall deviation is controlled within 0.5 pixels, which accurately reflects the actual contour shape of the island.

[0175] On this basis, the island boundary contour corresponding to each acquisition timestamp is obtained in the manner of obtaining the island boundary contour corresponding to the first acquisition timestamp. That is, for each image in the remote sensing image time sequence, the process of "extracting the timestamp-acquiring the actual water depth distribution-dividing the deep and shallow water influence areas-extracting the boundary segments in different regions-splicing" is repeated.

[0176] For example, for the second remote sensing image, the second actual water depth distribution corresponding to the second acquisition timestamp is extracted, the corresponding deep and shallow water influence areas are divided, the third and fourth boundary segments are extracted and spliced, and the island boundary contour corresponding to the second acquisition timestamp is obtained.

[0177] By analogy, the processing of all time sequence images is completed. The finally formed multi-time island boundary contour has time sequence consistency because it is based on the unified correction standard and splicing logic, and provides accurate time sequence data support for sea island remote sensing information change monitoring.

[0178] On this basis, the island boundary contours corresponding to all acquisition timestamps are arranged in chronological order to form a coherent island boundary contour change time sequence. This time sequence takes time as the axis and records the morphological characteristics of the island boundary at different times. By comparing the contour differences of adjacent timestamps, the dynamic change trend of the island boundary can be directly presented.

[0179] For example, in the time sequence contour of a certain island for three consecutive months, the boundary of the eastern shallow water area has advanced 5 meters towards the sea, and the boundary of the western deep water area has remained basically stable. This change can be directly quantified through the island boundary contour change time sequence, which is helpful for subsequent sea area management departments to conduct dynamic assessment of island resources and coastal zone protection planning to ensure the balance between island ecological system and resource development.

[0180] Through the specific implementation manner described above, the embodiments of the present application achieve the following technical effects:

[0181] This application proposes a method for monitoring changes in remote sensing information of islands in a sea area. First, it acquires the time series of remote sensing images of the target sea area and islands. Based on island identifiers, it retrieves the corresponding remote sensing image library and filters images with acquisition timestamps according to user-defined time parameters. Next, it collects tidal height information and water depth distribution information corresponding to each timestamp. The buffer zone is determined by island characteristics and image resolution to accurately delineate the monitoring area. Then, based on the water depth distribution information and tidal height information, it calculates the actual water depth distribution and performs regional correction on the images accordingly. For deep-water affected areas, the first boundary segment is extracted directly using a land-water separation threshold. For shallow-water affected areas, the second boundary segment is extracted after eliminating reflection interference using a water depth pixel corrector. These segments are then stitched together to form the island boundary contour at a single moment. Finally, all contours are arranged in chronological order to generate a time series of island boundary contour changes, achieving accurate capture and temporal presentation of dynamic changes in island boundaries.

[0182] The method provided in this application adopts the technical solution of "image time series acquisition - environmental parameter collection - regional correction and boundary extraction - time series contour construction", which integrates the precise correction of tidal and water depth data, the boundary extraction of deep and shallow water areas, and time series analysis based on unified standards. It solves the problems of large error and low accuracy caused by tidal offset, shallow water reflection interference, and lack of time series reference in traditional monitoring, and realizes high-precision dynamic monitoring of changes in remote sensing information of sea areas and islands.

[0183] Example 2, as shown in the appendix Figure 2 As shown, based on the inventive concept of a remote sensing information change monitoring method for sea areas and islands provided in Embodiment 1, this application also provides a remote sensing information change monitoring system for sea areas and islands, specifically including:

[0184] Image time series acquisition module 01 is used to acquire the time series of remote sensing images of islands in the target sea area. The remote sensing image time series includes multiple remote sensing images, and each remote sensing image has an acquisition timestamp.

[0185] The time-series tide level and water depth acquisition module 02 is used to acquire the tide height information of the target sea area islands at each acquisition time stamp, as well as the water depth distribution information of the monitoring sea area where the target sea area islands are located;

[0186] The correction and time series generation module 03 is used to correct each remote sensing image based on the tidal height information and water depth distribution information corresponding to each acquisition timestamp, and generate a time series of island boundary contour changes.

[0187] In one embodiment, the image timing acquisition module 01 is further configured to:

[0188] Obtain the island identifiers of the islands in the target sea area, and retrieve the corresponding remote sensing image database based on the island identifiers;

[0189] receiving a user-set acquisition time interval and a monitoring time period, wherein the monitoring time period comprises a start time and an end time;

[0190] filtering a plurality of remote sensing images from the remote sensing image library according to the acquisition time interval, the start time and the end time, and synchronously extracting an acquisition time of each of the remote sensing images as an acquisition time stamp of each of the remote sensing images.

[0191] In one embodiment, the time-series tide water depth acquisition module 02 is further configured to:

[0192] obtaining island feature information of the islands in the target sea area and a remote sensing image resolution of each of the remote sensing images, wherein the island feature information comprises island coordinate positions and island areas;

[0193] determining a buffer zone range according to the island areas and the remote sensing image resolutions;

[0194] obtaining tide height information corresponding to each of the acquisition time stamps based on the island coordinate positions;

[0195] determining a monitoring sea area based on the island coordinate positions and the buffer zone range, and obtaining water depth distribution information of the monitoring sea area.

[0196] In one embodiment, the correction and time-series generation module 03 is further configured to:

[0197] calculating actual water depth distribution corresponding to each of the acquisition time stamps based on the water depth distribution information and the tide height information corresponding to each of the acquisition time stamps;

[0198] performing correction processing on each of the remote sensing images according to the actual water depth distribution corresponding to each of the acquisition time stamps to obtain an island boundary contour corresponding to each of the acquisition time stamps, and generating a time series of the island boundary contour changes.

[0199] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. Moreover, the above-mentioned embodiments of the present application have been described. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

[0200] The above only describes the preferred embodiments of the present application and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0201] The specification and drawings are, of course, to be regarded in an illustrative rather than a restrictive sense. It is to be understood that any such modifications, variations, combinations or equivalents that fall within the scope of the application are intended to be embraced herein.

Claims

1. A method for monitoring changes in remote sensing information of sea islands in a sea area, characterized in that, The method comprises: acquiring a remote sensing image time sequence of islands in a target sea area, the remote sensing image time sequence comprising a plurality of remote sensing images, each remote sensing image having a collection time stamp; collecting tidal height information of the islands in the target sea area corresponding to each collection time stamp, and water depth distribution information of a monitored sea area in which the islands in the target sea area are located; correcting each remote sensing image according to the tidal height information corresponding to each collection time stamp and the water depth distribution information, to generate an island boundary profile change time sequence; collecting tidal height information of the islands in the target sea area corresponding to each collection time stamp, and water depth distribution information of a monitored sea area in which the islands in the target sea area are located, comprising: acquiring island feature information of the islands in the target sea area and remote sensing image resolutions of each remote sensing image, the island feature information comprising island coordinate positions and island areas; determining a buffer zone range according to the island areas and the remote sensing image resolutions; acquiring tidal height information corresponding to each collection time stamp based on the island coordinate positions; determining a monitored sea area based on the island coordinate positions and the buffer zone range, and acquiring water depth distribution information of the monitored sea area; determining a buffer zone range according to the island areas and the remote sensing image resolutions, comprising: calculating an island equivalent radius based on the island area, and acquiring a first buffer distance in a buffer distance mapping table based on the island equivalent radius; determining a second buffer distance based on the remote sensing image resolutions, the second buffer distance being a preset multiple of the remote sensing image resolutions; comparing the first buffer distance with the second buffer distance, and taking a larger value as a basic buffer radius; increasing a safety margin on the basis of the basic buffer radius to obtain the buffer zone range; correcting each remote sensing image according to the tidal height information corresponding to each collection time stamp and the water depth distribution information, to generate an island boundary profile change time sequence, comprising: calculating actual water depth distribution corresponding to each collection time stamp based on the water depth distribution information and the tidal height information corresponding to each collection time stamp; correcting each remote sensing image according to the actual water depth distribution corresponding to each collection time stamp to obtain an island boundary profile corresponding to each collection time stamp, and generating the island boundary profile change time sequence.

2. The method of claim 1, wherein, Acquiring a remote sensing image time sequence of islands in a target sea area, the remote sensing image time sequence comprising a plurality of remote sensing images, each remote sensing image having a collection time stamp, comprising: acquiring an island identifier of the islands in the target sea area, and calling a corresponding remote sensing image library based on the island identifier; receiving a collection time interval and a monitoring time period set by a user, wherein the monitoring time period comprises a start time and an end time; selecting a plurality of remote sensing images from the remote sensing image library according to the collection time interval, the start time and the end time, and synchronously extracting a collection time of each remote sensing image as a collection time stamp of each remote sensing image.

3. The method of claim 1, wherein, Correcting each remote sensing image according to the actual water depth distribution corresponding to each collection time stamp to obtain an island boundary profile corresponding to each collection time stamp, comprising: acquiring a first remote sensing image from a plurality of remote sensing images in a remote sensing image time sequence; extracting a first acquisition timestamp of the first remote sensing image, and obtaining a first actual water depth distribution according to the first acquisition timestamp; determining a deep water influence area and a shallow water influence area in the first remote sensing image according to the first actual water depth distribution; extracting a first boundary segment of the target sea island in the deep water influence area and a second boundary segment of the target sea island in the shallow water influence area for the first remote sensing image; spatially splicing the first boundary segment and the second boundary segment to generate an island boundary contour corresponding to the first acquisition timestamp; obtaining an island boundary contour corresponding to each acquisition timestamp in the manner of obtaining the island boundary contour corresponding to the first acquisition timestamp.

4. The method of claim 3, wherein, extracting a first boundary segment of the target sea island in the deep water influence area and a second boundary segment of the target sea island in the shallow water influence area for the first remote sensing image, comprising: obtaining a water-land separation threshold value; performing pixel classification processing on the first remote sensing image based on the water-land separation threshold value in the deep water influence area, and extracting a demarcation line between seawater and the target sea island to obtain the first boundary segment; after performing seabed reflection correction on the first remote sensing image in the shallow water influence area, performing pixel classification processing in combination with the water-land separation threshold value, and extracting a demarcation line between seawater and the target sea island to obtain the second boundary segment.

5. The method of claim 4, wherein, after performing seabed reflection correction on the first remote sensing image in the shallow water influence area, performing pixel classification processing in combination with the water-land separation threshold value, and extracting a demarcation line between seawater and the target sea island to obtain the second boundary segment, comprising: in the first remote sensing image, obtaining actual water depth values and original pixel values of each pixel point in the shallow water influence area; performing seabed reflection correction on each pixel point in the shallow water influence area by a water depth pixel corrector to obtain corrected pixel values, and replacing the original pixel values with the corrected pixel values; performing pixel classification processing on the first remote sensing image based on the corrected pixel values and the water-land separation threshold value, and extracting a demarcation line between seawater and the target sea island to obtain the second boundary segment.

6. The method of claim 5, wherein, The construction steps of the water depth pixel corrector include: collecting a sample actual water depth record set and a sample original pixel set, and labeling based on the sample actual water depth record set and the sample original pixel set to obtain a sample corrected water depth record set; taking the sample actual water depth record set and the sample original pixel set as inputs, and taking the sample corrected water depth record set as a supervision signal to train and generate the water depth pixel corrector.

7. A system for monitoring changes in remote sensing information of sea areas and islands, characterized in that it comprises: The system is used to execute the sea island remote sensing information change monitoring method of any one of claims 1-6, and the system comprises: an image time sequence acquisition module configured to acquire a remote sensing image time sequence of a target sea island, the remote sensing image time sequence comprising a plurality of remote sensing images, each remote sensing image having an acquisition timestamp; The time-series tidal level and depth collecting module is configured to collect tidal height information of the target sea area islands corresponding to each of the collecting time stamps, and water depth distribution information of a monitoring sea area where the target sea area islands are located. The correction and time-series generation module is configured to perform correction processing on each of the remote sensing images according to the tidal height information corresponding to each of the collecting time stamps and the water depth distribution information, and generate a time series of island boundary contour changes.

Images