A multi-period remote sensing image automatic change detection processing method and device

Abstract

The application relates to a multi-period remote sensing image automatic change detection processing method and device, and relates to the technical field of surveying and mapping geographic information. The method comprises the following steps: acquiring multi-period remote sensing images including fixed base period remote sensing images, later time phase remote sensing images and rolling base period remote sensing images; change extraction is carried out on the multi-period remote sensing images to obtain fixed base period graph patches and rolling base period graph patches; on the basis of the base period graph patches, combined with previous period graph patches, a deduplication process is carried out to obtain fixed base period graph patches j and target rolling base period graph patches; normalization monitoring is carried out by using the extracted fixed base period graph patches j and target rolling base period graph patches, so that suspected change graph patches in a target range can be quickly and accurately extracted, redundant checking work caused by repeated graph patch extraction of multi-period images is avoided, the requirement of rapid response of normalization monitoring is met, the change detection fullness rate is improved, and active, intelligent and comprehensive natural resource investigation and monitoring basic data are provided for natural resource investigation, monitoring and management.

Description

Technical Field

[0001] This application relates to the field of surveying and mapping geographic information technology, and in particular to a method and apparatus for automatic change detection and processing of multi-period remote sensing images. Background Technology

[0002] In the past, the investigation and monitoring of natural resources such as land mostly relied on remote sensing imagery to identify geographic features. However, the identification of changes in multiple periods of remote sensing imagery was done using a "human wave" approach, extracting change patches manually one by one across the map. This resulted in problems such as low operational collaboration capabilities, outdated technology, long monitoring cycles, and low efficiency, failing to meet the needs of natural resource management operations such as farmland protection, early detection, early warning, and early prevention of illegal land use, and ecological protection and restoration. Under the new circumstances and requirements, natural resource management operations demand higher accuracy, more frequent monitoring, and a wider range of monitoring targets. Timely understanding of the development and utilization dynamics of natural resources is crucial to supporting full-business, full-process monitoring and supervision of natural resources; efficiency is paramount.

[0003] In natural resource surveys and monitoring, existing remote sensing image change detection methods mainly involve identifying and comparing fixed-base-period remote sensing images with the latest acquired remote sensing images to detect and extract changed patches. However, the fixed-base-period images used for extraction are changes from the previous year, and multi-source, multi-temporal change detection is prone to repeated patch extraction, resulting in a large workload and low efficiency, which cannot meet the requirements of rapid response in routine monitoring. Summary of the Invention

[0004] This application provides an automatic change detection and processing method and apparatus for multi-phase remote sensing images. It utilizes multi-temporal remote sensing images to quickly and accurately extract change patches within a target area, realize the detection of changes in land cover, obtain information on changes in important geographic elements, and thus enable routine monitoring. This avoids redundant verification work caused by extracting duplicate patches from multiple images, meets the requirements for rapid response in routine monitoring, and solves the problems of high workload and low efficiency caused by repeated patch extraction in existing technologies.

[0005] Firstly, this application provides an automatic change detection and processing method for multi-period remote sensing images, including:

[0006] Acquire fixed base period remote sensing images, rolling base period remote sensing images, and the latest monitored post-temporal remote sensing images of the target area;

[0007] Change extraction is performed based on the fixed base period remote sensing image and the subsequent time phase remote sensing image to obtain fixed base period patch a; and change extraction is performed based on the rolling base period remote sensing image and the subsequent time phase remote sensing image to obtain rolling base period patch n.

[0008] Based on the fixed base period patch a, and combined with previously verified patches, duplicates are removed to obtain the fixed base period patch j;

[0009] Based on the rolling base period patch n, and combined with the fixed base period patch j and previously online patches, deduplication is performed to obtain the target rolling base period patch;

[0010] The target area is routinely monitored based on the fixed base period patch j and the target rolling base period patch;

[0011] Wherein, both the fixed base period patch j and the target rolling base period patch are base period change patches.

[0012] Optionally, acquire fixed base-period remote sensing images, rolling base-period remote sensing images, and the most recently monitored post-temporal remote sensing images of the target area, including:

[0013] Land change survey images of the target area are extracted from the pre-set land change survey database to serve as fixed base period remote sensing images;

[0014] The initial post-temporal images from the previous batch are acquired as rolling base period remote sensing images, and the latest monitoring images are acquired as post-temporal remote sensing images.

[0015] Optionally, change extraction is performed based on the rolling base period remote sensing image and the subsequent time-phase remote sensing image to obtain the rolling base period patch n, including:

[0016] Based on the rolling base period remote sensing image and the subsequent time phase remote sensing image, image selection is performed to obtain the preferred monitoring image. The preferred monitoring image is the monitoring image whose latest monitoring image time phase is better than the previous monitoring image time phase by a first preset time.

[0017] Based on the preferred monitoring image and the rolling base period remote sensing image, patch extraction is performed to obtain the rolling base period patch n.

[0018] Optionally, based on the fixed base period patch a, and combined with previously verified patches, deduplication is performed to obtain the fixed base period patch j, including:

[0019] Obtain previously verified map features and extract the scope of change investigation from the preset change investigation database;

[0020] Extract fixed base period patch b, which is outside the scope of the change investigation, from the fixed base period patch a;

[0021] The extracted image range D is at least a second preset time better than the rolling base period remote sensing image.

[0022] Based on the image range D, extract fixed base period patch c and fixed base period patch d from the fixed base period patch b;

[0023] Based on the previously verified map patches, fixed base period map patch c, and fixed base period map patch d, deduplication is performed to obtain fixed base period map patch j.

[0024] Optionally, extracting fixed base period patch b outside the scope of the change survey from the fixed base period patch a includes:

[0025] Using a preset semantic segmentation model, the fixed base period map patch a is subjected to full-element land use semantic segmentation using an intelligent remote sensing interpretation algorithm to obtain image land use information, which includes image land use and image land use range.

[0026] The image land cover information is compared with the change survey scope, and the fixed base period map patch b outside the change survey scope is extracted from the fixed base period map patch a.

[0027] Optionally, based on the previously verified map patches, fixed base period map patch c, and fixed base period map patch d, deduplication is performed to obtain fixed base period map patch j, including:

[0028] Based on the previously verified map patches, a summary detection was performed to obtain verified reference map patches;

[0029] Based on the verified reference map patch, the fixed base period map patch c and the fixed base period map patch d are cross-compared to obtain the fixed base period map patch e corresponding to the fixed base period map patch c and the fixed base period map patch f corresponding to the fixed base period map patch d;

[0030] Extract the image range E whose time phase differs from the time phase of the latest monitoring image by a third preset time;

[0031] Based on the image range E, extract fixed base period patch g and fixed base period patch h from the fixed base period patch f;

[0032] Extract the fixed base period patch i that falls within the cultivated land area of ​​the image range E from the fixed base period patch g;

[0033] Based on the fixed base period patch e, the fixed base period patch h, and the fixed base period patch i, the patches are merged to obtain the deduplicated fixed base period patch j;

[0034] Wherein, the fixed base period patch e is a base period patch extracted from the fixed base period patch c whose area intersection ratio is less than a first preset threshold, and the fixed base period patch f is a base period patch extracted from the fixed base period patch d whose area intersection ratio is less than a second preset threshold.

[0035] Optionally, based on the rolling base period patch n, and combined with the fixed base period patch j and previously online patches, deduplication is performed to obtain the target rolling base period patch, including:

[0036] Semantic segmentation and extraction are performed based on the post-temporal remote sensing image and the fixed base period patch j to obtain fixed base period patch k and fixed base period patch l;

[0037] Based on the previously launched map patches, a list of launched reference map patches is obtained. Furthermore, based on the launched reference map patches and the fixed base period map patch l, an intersection comparison is performed to obtain the fixed base period map patch m.

[0038] Based on the fixed base period patch m, the fixed base period patch k, and the rolling base period patch n, a deduplicated target rolling base period change patch is obtained by merging and area filtering.

[0039] Optionally, semantic segmentation extraction is performed based on the later-phase remote sensing image and the fixed base period patch j to obtain fixed base period patch k and fixed base period patch l, including:

[0040] The post-temporal remote sensing image is semantically segmented using a preset semantic segmentation model to obtain bulldozing patches;

[0041] Extract fixed base period patch k that intersects with the bulldozing patch from the fixed base period patch j, and extract fixed base period patch l that does not intersect with the bulldozing patch from the fixed base period patch j.

[0042] Optionally, based on the online reference patch and the fixed base period patch l, an intersection comparison is performed to obtain the fixed base period patch m, including:

[0043] Based on the already online reference patches, easily changeable patches are removed to obtain the target reference patches;

[0044] The area intersection ratio is calculated based on the target reference patch and the fixed base period patch l to obtain the area intersection ratio value;

[0045] Extract fixed base period patches m from fixed base period patch l whose area intersection ratio is less than the third preset threshold.

[0046] Secondly, this application provides an automatic change detection and processing device for multi-period remote sensing images, comprising:

[0047] The multi-phase remote sensing image acquisition module is used to acquire fixed base-phase remote sensing images, rolling base-phase remote sensing images, and the latest monitored post-phase remote sensing images of the target area.

[0048] The change extraction module is used to extract changes based on the fixed base period remote sensing image and the subsequent time phase remote sensing image to obtain a fixed base period patch a, and to extract changes based on the rolling base period remote sensing image and the subsequent time phase remote sensing image to obtain a rolling base period patch n.

[0049] The first deduplication module is used to perform deduplication processing based on the fixed base period patch a and the previously verified patches to obtain the fixed base period patch j.

[0050] The second deduplication module is used to perform deduplication processing based on the rolling base period patch n, combined with the fixed base period patch j and previously online patches, to obtain the target rolling base period patch.

[0051] The routine monitoring module is used to perform routine monitoring of the target area based on the fixed base period patch j and the target rolling base period patch;

[0052] Wherein, both the fixed base period patch j and the target rolling base period patch are base period change patches.

[0053] In summary, this embodiment acquires multiple periods of remote sensing images, including fixed base period remote sensing images, the latest monitored post-temporal remote sensing images, and rolling base period remote sensing images. Changes are then extracted from these multiple periods of remote sensing images to obtain fixed base period patches and rolling base period patches. Based on the extracted base period patches, duplicates are removed by combining them with previous patches to obtain fixed base period patch j and the target rolling base period patch. These fixed base period patches j and the target rolling base period patch are then used for routine monitoring. This embodiment utilizes multi-temporal remote sensing images to quickly and accurately extract targets. By identifying suspected changes in land cover within the specified area, this method enables the detection of changes in important geographic elements, allowing for timely and accurate understanding of the current status of natural resource utilization. It avoids redundant verification work caused by extracting duplicate patches from multiple image periods, meets the requirements for rapid response in routine monitoring, and improves the recall rate of change detection. This provides proactive, intelligent, and comprehensive basic data for natural resource surveys, monitoring, and management. Based on this, it allows for continuous monitoring and updating of national land survey results, addressing the problems of high workload and low efficiency caused by repeated patch extraction in existing technologies. Attached Figure Description

[0054] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0055] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 A flowchart illustrating an automatic change detection and processing method for multi-period remote sensing images provided in this application embodiment;

[0057] Figure 2 This is a flowchart illustrating the steps of an optional embodiment of the present application for an automatic change detection and processing method for multi-period remote sensing images;

[0058] Figure 3 This is a technical circuit diagram of an automatic change detection and processing method for multi-period remote sensing images provided as an optional example of this application;

[0059] Figure 4 This is an optional example of a graph extraction flowchart provided in this application;

[0060] Figure 5 This is an optional example of a "phase-optimal selection" tool usage diagram provided in this application;

[0061] Figure 6 This is an optional example of a fixed base period patch deduplication flowchart provided in this application;

[0062] Figure 7 This is an optional example of a flowchart for deduplicating rolling base period patches provided in this application;

[0063] Figure 8 This is a diagram illustrating the use of a "deduplication model" tool, provided as an optional example in this application.

[0064] Figure 9 A structural block diagram of an automatic change detection and processing device for multi-period remote sensing images provided in this application embodiment;

[0065] Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0066] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0067] To facilitate understanding of the embodiments of this application, further explanations and descriptions will be provided below in conjunction with the accompanying drawings and specific embodiments. These embodiments do not constitute a limitation on the embodiments of this application.

[0068] Figure 1 This is a flowchart illustrating an automatic change detection and processing method for multi-period remote sensing images provided in an embodiment of this application. Figure 1 As shown in the embodiments of this application, the automatic change detection and processing method for multi-period remote sensing images may specifically include the following steps:

[0069] Step 110: Acquire the fixed base period remote sensing image, the rolling base period remote sensing image, and the latest monitored post-temporal remote sensing image of the target area.

[0070] In this embodiment, the fixed base period remote sensing image can be understood as: an image from the previous year's land change survey, based on the current date. For example, it could be an image from the 2023 land change survey; this embodiment does not impose any restrictions on this. The rolling base period remote sensing image is also called the previous monitoring image. For later-phase remote sensing images, this usually refers to remote sensing images acquired after the rolling base period remote sensing image. To improve the accuracy and efficiency of change detection, this embodiment uses the rolling base period remote sensing image as the standard and selects the latest monitoring image from the newly monitored later-phase remote sensing image set that is superior to the rolling base period remote sensing image as the later-phase remote sensing image.

[0071] It should be noted that in the embodiments of this application, each remote sensing image can also be simply referred to as an image, and the previous phase image can also be called the previous phase map, and the subsequent phase image can also be called the subsequent phase map, etc.

[0072] Step 120: Based on the fixed base period remote sensing image and the subsequent time-phase remote sensing image, perform change extraction to obtain fixed base period patch a; and based on the rolling base period remote sensing image and the subsequent time-phase remote sensing image, perform change extraction to obtain rolling base period patch n.

[0073] In practical implementation, the base period patch can also be called the base period change patch, which is divided into deduplicated base period patches and non-deduplicated base period patches, mainly including but not limited to: fixed base period patches and rolling base period patches. This embodiment mainly uses previous and subsequent time phase images to extract changes and obtain the corresponding base period patches.

[0074] Specifically, this embodiment can pre-train a change detection model and preset a corresponding intelligent remote sensing interpretation algorithm. For the extraction of fixed base period patch a, this embodiment uses the previous year's change survey image (i.e., fixed base period remote sensing image) as the previous time phase image, and combines it with the latest monitored subsequent time phase image to extract the fixed base period patch using the intelligent remote sensing interpretation algorithm and the change detection model.

[0075] For the extraction of the rolling base period patch n, this embodiment uses the previous period's imagery as the preceding temporal image (i.e., the rolling base period remote sensing image). A subsequent temporal image, one day better than the previous image, is extracted from the most recently monitored imagery and used as the subsequent temporal image for rolling base period patch extraction. Then, combining the preceding and subsequent temporal images, an intelligent remote sensing interpretation algorithm is used in conjunction with a "change detection model" to extract the rolling base period patch n.

[0076] Step 130: Based on the fixed base period patch a, deduplication is performed by combining previously checked patches to obtain the fixed base period patch j.

[0077] Step 140: Based on the rolling base period patch n, and combined with the fixed base period patch j and previously online patches, perform deduplication processing to obtain the target rolling base period patch.

[0078] Wherein, both the fixed base period patch j and the target rolling base period patch are base period change patches.

[0079] Steps 130-140 are described uniformly as follows:

[0080] In this embodiment, previously verified map features can be obtained by merging the database of previously verified map features into a single layer. The database of previously verified map features mainly stores map features that have been verified previously. Similarly, previously online map features can be obtained by merging the database of previously online map features into a single layer. The database of previously online map features mainly stores map features that have been online previously.

[0081] In practical implementation, after obtaining the base period map features, deduplication can be performed on the base period map features. The deduplication process can at least take into account previous map features, including previously online map features and previously verified map features. This embodiment achieves change detection by comparing multiple map features, such as the base period map features and previous map features, and fully considers factors such as the overlap rate between map features to obtain the target base period map feature.

[0082] In practical implementation, the deduplication process for base period patches in this embodiment can take into account not only previous patches but also various other deduplication factors. Specifically, for the deduplication of fixed base period patch a, factors such as previously verified patches, the DLTB database of change surveys, the acquisition time of monitoring images, and patch overlap rate can be taken into account. The software can perform one-click deduplication to obtain the deduplicated fixed base period patch j. This embodiment will not elaborate on this aspect.

[0083] For the deduplication of the rolling base period patch n, factors such as the fixed base period patch j after deduplication, the bulldozing patch in the semantic segmentation of the later phase image, the previously online patches, the patch overlap rate, and the minimum map area can be taken into account. The deduplication can be performed by software in one click to obtain the target rolling base period patch after deduplication. This embodiment will not be described in detail.

[0084] Therefore, this embodiment realizes a method for rapid and accurate extraction and deduplication of map features for routine monitoring and verification. By using information technology, it can quickly and accurately extract changed map features within the target area, avoid redundant verification work caused by extracting duplicate map features from multiple images, improve the recall rate of change detection, and provide proactive, intelligent and comprehensive basic data for natural resource surveys, monitoring and management, thus forming a routine monitoring system that serves natural resource management.

[0085] Step 150: Perform routine monitoring of the target area based on the fixed base period patch j and the target rolling base period patch.

[0086] In this embodiment, by combining fixed base period and rolling base period with post-phase image complete detection and other change patch extraction methods, it is possible to quickly support the routine monitoring of natural resources and achieve multiple monitoring objectives of routine monitoring, including but not limited to: real-time monitoring, monthly clearing and quarterly verification, annual summary, etc., thereby improving the ability of survey and monitoring to support the business management of natural resources and providing technical support for the comprehensive support of precise management and scientific decision-making of natural resources.

[0087] As can be seen, this embodiment acquires multiple periods of remote sensing images, including fixed base period remote sensing images, the latest monitored post-temporal remote sensing images, and rolling base period remote sensing images. Change extraction is then performed on these multiple periods of remote sensing images to obtain fixed base period patches and rolling base period patches. Based on the extracted base period patches, deduplication is performed by combining them with previous patches to obtain fixed base period patch j and target rolling base period patch j. These are then used for routine monitoring. This embodiment utilizes multi-temporal remote sensing images to quickly and accurately extract suspected change patches within the target area, achieving land cover change detection and obtaining information on changes in important geographic elements. This allows for timely and accurate understanding of the current status of natural resource utilization, avoiding redundant verification work caused by extracting duplicate patches from multiple periods of images. It meets the requirements for rapid response in routine monitoring, improves the recall rate of change detection, and solves the problems of high workload and low efficiency caused by repeated patch extraction in existing technologies.

[0088] The automatic change detection and processing method for multi-period remote sensing imagery proposed in this embodiment can continuously monitor and update land survey results, maintaining the accuracy and timeliness of basic geographic information. It achieves real-time changes in land surveys through the development of intelligent methods. By monitoring and analyzing changed land parcels using multi-period remote sensing imagery, land violations can be quickly identified, alleviating the difficulty of investigation. The method of removing overlaps solves the problems of repeated extraction of changed patches and low recall rate of online patches in multi-source, multi-temporal remote sensing imagery.

[0089] Reference Figure 2 This illustration shows a flowchart of an optional embodiment of an automatic change detection and processing method for multi-period remote sensing images. The method specifically includes the following steps:

[0090] Step 210: Acquire the fixed base period remote sensing image, the rolling base period remote sensing image, and the latest monitored post-temporal remote sensing image of the target area.

[0091] Optionally, this embodiment acquires fixed base period remote sensing images, rolling base period remote sensing images, and the latest monitored post-temporal remote sensing images of the target area. Specifically, it may include: extracting land change survey images of the target area from a preset land change survey database to serve as fixed base period remote sensing images; acquiring the initial post-temporal images of the previous batch as rolling base period remote sensing images; and acquiring the latest monitored images as post-temporal remote sensing images.

[0092] In related technologies, existing solutions, besides using a fixed base period change monitoring mode for repeated extraction of image patches, also partially employ change monitoring based on a rolling base period. However, the base period image selected for rolling base period image patch extraction is a later phase image from the previous period of routine monitoring. This can result in short intervals between the two image periods, making it easy to miss slowly changing image patches, thus leading to low recall.

[0093] To address the problems of high repetition rate in fixed base period extraction and low recall rate in rolling base period extraction used in existing routine monitoring change patch extraction schemes, this embodiment proposes a change patch extraction method that combines fixed base period and rolling base period for complete detection of post-phase images, in order to reduce the occurrence of patch omissions.

[0094] For example, refer to Figure 3 The technical flowchart shown in this embodiment indicates that the data to be acquired during the data preparation stage may include, but is not limited to: pre-roll base period phase maps (i.e., previous period image combination maps), previous period result image data, post-roll period maps (i.e., current period image combination maps), current period image result data, database of previously verified map features, database of previously online map features, DLTB layer of previous year's land change survey data, and previous year's land change survey images.

[0095] It should be noted that, Figure 3 In the accompanying figures, “23-year change survey image” refers to the fixed base period remote sensing image, “latest monitoring image” refers to the later time phase remote sensing image, and “previous monitoring image” refers to the rolling base period remote sensing image.

[0096] In practical implementation, the acquired image data, including combined image files of pre- and post-temporal phase images, can all adopt a fixed and consistent format. Preferably, shapefile format can be used to ensure consistency between the mathematical basis and the image. The data structure can be shown in Table 1 below:

[0097]

[0098] Table 1. Data Structure of Before / After Timing Image Combination Map

[0099] Step 220: Based on the fixed base period remote sensing image and the subsequent time-phase remote sensing image, perform change extraction to obtain fixed base period patch a; and based on the rolling base period remote sensing image and the subsequent time-phase remote sensing image, perform change extraction to obtain rolling base period patch n.

[0100] For example, in combination Figure 3 The technical roadmap shown is for reference only. Figure 4The flowchart shown illustrates the process of extracting map patches. In this embodiment, the "change detection model" of the intelligent remote sensing interpretation algorithm production and application platform is used to extract fixed base period map patches and rolling base period map patches.

[0101] Optionally, the above-mentioned change extraction based on the rolling base period remote sensing image and the subsequent time-phase remote sensing image to obtain the rolling base period patch n may specifically include: image selection based on the rolling base period remote sensing image and the subsequent time-phase remote sensing image to obtain a preferred monitoring image, wherein the preferred monitoring image is a monitoring image whose latest monitoring image time phase is better than the previous monitoring image time phase by a first preset time; and patch extraction based on the preferred monitoring image and the rolling base period remote sensing image to obtain the rolling base period patch n.

[0102] In this embodiment, preferably, the first preset time is 1 day. That is, in the process of extracting the rolling base period patch n, the previous monitoring image can be used as the base period image, and the latest monitoring image that is 1 day or more better than the previous monitoring image can be used as the subsequent time phase image to extract the rolling base period change patch.

[0103] In practical implementation, this embodiment can extract map features based on actual map feature extraction needs, combined with the "change detection model" of the intelligent remote sensing interpretation algorithm production and application platform. Specifically, map feature extraction includes: ① using the previous year's change survey image as the base period image and the latest monitoring image as the subsequent time phase image, and using the "change detection model" of the intelligent remote sensing interpretation algorithm production and application platform to extract "fixed base period map features"; ② using the previous monitoring image as the base period image, obtaining the latest monitoring image that is one day or more better than the previous monitoring image as the subsequent time phase image, and using the "change detection model" of the intelligent remote sensing interpretation algorithm production and application platform to extract "rolling base period map features", and extracting rolling base period change map features.

[0104] In practical implementation, this embodiment can be used to construct a "timing-optimal selection" tool. For example... Figure 5 As shown, by using the "Optimal Timing Selection" tool to select the corresponding relevant image and inputting parameters (including but not limited to the number of days for image settling), the optimal image can be extracted with a single click. For obtaining the post-temporal settling image from the rolling base period patch extraction: use the "Optimal Timing Selection" tool, the usage method of which is as follows... Figure 5 As shown. The steps for using it are as follows: ① Select "Previous Period Image Combination Map" in "Input Previous Phase"; ② Select "Current Period Image Combination Map" in "Input Later Phase"; ③ Enter "1" for the number of days for image placement, which means that the current period image combination map is 1 day or more better than the previous period image; ④ Output the latest image placement, which is "Later Phase Image Placement Map 1 Day Better Than Previous Image". The resulting image corresponding to this image placement is "Later Phase Image 1 Day Better Than Previous Image".

[0105] Furthermore, for one-click extraction of optimal imagery, this embodiment can utilize the model builder in ArcGIS (a geographic data processing software) to develop an image selection system as a "temporally optimal selection" tool, and embed it into the ArcGIS toolbox for execution. This allows for one-click selection of changed patches to extract the required image data during routine monitoring work.

[0106] Step 230: Obtain previously verified map patches and extract the scope of change surveys from the preset change survey database.

[0107] Step 240: Extract fixed base period patch b outside the change survey range from the fixed base period patch a, and extract the image range D of the later time phase remote sensing image with a time phase better than the rolling base period remote sensing image by at least a second preset time.

[0108] Steps 230-240 are described uniformly as follows:

[0109] In this embodiment, the deduplication process for base period patches requires not only preparing before / after temporal images, but also preparing separate layer data. The extraction of separate layer data includes, but is not limited to: merging the database of previously verified patches into one layer, such as "Previous Verified Patches"; merging the database of previously online patches into one layer, such as "Previous Online Verified Patches"; extracting the cultivated land layer from the change survey data DLTB, such as "DLTB_Cultivated Land"; and extracting the orchard and forest land layers from the change survey data DLTB, such as "DLTB_Orchard Land". This embodiment determines the scope of the change survey by extracting corresponding layers from the change survey database. Layer categories include, but are not limited to, cultivated land layers, forest land layers, and orchard land layers.

[0110] In specific implementation, for a fixed base period map patch a, this embodiment can identify the land type (such as cultivated land, forest land, and orchard) included in the base period map patch a, and then, in conjunction with the change survey scope, extract the fixed base period map patch b outside the change survey scope (such as outside the orchard or forest land scope), such as... Figure 3 and Figure 6 As shown.

[0111] In one optional embodiment, the present application embodiment extracts fixed base period map patch b outside the change survey range from the fixed base period map patch a. Specifically, this may include: performing full-element land use semantic segmentation on the fixed base period map patch a using a preset semantic segmentation model and an intelligent remote sensing interpretation algorithm to obtain image land use information, wherein the image land use information includes image land use and image land use range; comparing the image land use information with the change survey range, and extracting the fixed base period map patch b outside the change survey range from the fixed base period map patch a.

[0112] In this embodiment, the land category range can be understood as the range of the image land category in the fixed base period map patch a. For example, when the fixed base period map patch a contains a garden, the garden is the specific range of the fixed base period map patch a.

[0113] In its specific implementation, this embodiment can input a fixed base period map patch a into a semantic segmentation model, and combine this with an intelligent remote sensing interpretation algorithm to perform full-element land use semantic segmentation on the fixed base period map patch a, so as to extract the fixed base period map patch b outside the scope of the change survey from the fixed base period map patch a, such as... Figure 3 and Figure 6 As shown.

[0114] The semantic segmentation extraction mainly utilizes the "semantic segmentation model" of the intelligent remote sensing interpretation algorithm production and application platform to extract the semantics of all land cover types from the current image. The semantic segmentation code table involved in the deduplication of image patches in this application is shown in Table 2 below:

[0115] Serial Number Semantic code Semantic land class 1 1 arable land 2 2 Garden, Young Forest 3 3 water body 4 4 grassland 5 5 building 6 6 the way 7 7 Structures 8 9 Pushing fill soil, bare soil

[0116] Table 2 shows the semantic segmentation code table. The semantic segmentation results are in "GDB" database format, and the data structure is shown in Table 3.

[0117] Serial Number Field Name field code Field type Field length Decimal places Remark 1 post post Char 10

[0118] Table 3. Data Structure of Semantic Segmentation Results

[0119] Step 250: Based on the image range D, extract fixed base period patch c and fixed base period patch d from the fixed base period patch b.

[0120] In specific implementation, after extracting the image range D where the latest monitoring image phase is better than the previous monitoring image phase by a second preset time (preferably, the second preset time is 10 days), this embodiment can extract fixed base period patches c that fall within the image range D from fixed base period patches b, and extract fixed base period patches d that do not fall within the range D from fixed base period patches b, such as... Figure 3 and Figure 6 As shown.

[0121] Step 260: Based on the previously verified map patches, fixed base period map patch c, and fixed base period map patch d, perform deduplication to obtain fixed base period map patch j.

[0122] In its implementation, this embodiment primarily utilizes previously verified patches, fixed base period patch c, and fixed base period patch d for deduplication of fixed base period patches. For deduplication of fixed base period patches, this embodiment fully considers factors such as the change survey DLTB database, monitoring image acquisition time, and patch overlap rate to address issues such as increased verification workload, large number of false changes in patches, and redundancy in verification work caused by repeated patch extraction in existing technologies.

[0123] Optionally, the above-mentioned deduplication process based on the previously verified map patches, fixed base period map patch c, and fixed base period map patch d to obtain fixed base period map patch j may specifically include the following sub-steps:

[0124] Sub-step 2601: Based on the previously verified map patches, perform summary detection to obtain verified reference map patches.

[0125] Sub-step 2602: Based on the verified reference patch, perform cross-comparison between the fixed base period patch c and the fixed base period patch d to obtain the fixed base period patch e corresponding to the fixed base period patch c and the fixed base period patch f corresponding to the fixed base period patch d.

[0126] Wherein, the fixed base period patch e is a base period patch extracted from the fixed base period patch c whose area intersection ratio is less than a first preset threshold, and the fixed base period patch f is a base period patch extracted from the fixed base period patch d whose area intersection ratio is less than a second preset threshold.

[0127] A unified description is provided for sub-steps 2601 and 2602:

[0128] Preferably, the first preset threshold is 50%, and the second preset threshold is 20%.

[0129] In practice, previously verified map features refer to previously verified change detection reference map features. By summarizing all previously verified change detection reference map features, verified reference map features are obtained. Then, these change detection reference map features are used to perform intersection comparisons with map features from a fixed base period. The intersection comparison process mainly includes calculating the area intersection ratio of the two map features. Specifically, referring to… Figure 3 and Figure 6 As shown, this embodiment calculates the area intersection ratio between the reference patch and the fixed base period patch c, extracts the fixed base period patch e whose area intersection ratio is less than 50%, and calculates the area intersection ratio between the reference patch and the fixed base period patch d, extracts the fixed base period patch f whose area intersection ratio is less than 20%.

[0130] Sub-step 2603: Extract the image range E whose time phase of the latest monitoring image differs from the time phase of the previous monitoring image by a third preset time.

[0131] Sub-step 2604: Based on the image range E, extract fixed base period patch g and fixed base period patch h from the fixed base period patch f.

[0132] Sub-step 2605: Extract the fixed base period patch i from the fixed base period patch g that falls within the cultivated land area of ​​the image range E.

[0133] Sub-step 2606: Based on the fixed base period patch e, the fixed base period patch h, and the fixed base period patch i, patch merging is performed to obtain the deduplicated fixed base period patch j.

[0134] A unified description is provided for sub-steps 2603-2606:

[0135] Preferably, the third preset time can be 30 days.

[0136] Reference Figure 3 and Figure 6 As shown, this embodiment extracts the image range E, which is 30 days different from the previous monitoring image, to compare the fixed base period patch f with the image range E. Within the fixed base period patch f, fixed base period patches g that fall within the image range E and h that do not fall within the image range E are extracted. Then, fixed base period patches i within the cultivated land area of ​​the previous year's land use change are extracted from fixed base period patch g. Thus, this embodiment obtains fixed base period patches e, h, and i after deduplication based on factors such as monitoring image acquisition time and patch overlap rate. Finally, fixed base period patches e, h, and i are merged to obtain the deduplicated fixed base period change patch j.

[0137] In related technologies, existing technologies still suffer from outdated deduplication methods for change detection patches. Specifically, previous deduplication methods directly erase newly extracted patches from previously uploaded patches. This method tends to generate a large number of fragmented patches and easily misses continuously changing patches and key monitoring patches.

[0138] To address the problems caused by outdated deduplication methods in existing technologies, this embodiment proposes a deduplication method that takes into account factors such as existing online patches, image resolution, image acquisition time, and patch overlap rate. This significantly reduces patch overlap and solves the problems of repeated extraction of change patches from multi-source, multi-temporal remote sensing images and low recall rate of online patches. Compared to methods without deduplication and with a fixed base period, this embodiment improves efficiency by 2 times, and the recall rate of subsequent rolling base period patches is increased to 85%.

[0139] Step 270: Semantic segmentation and extraction are performed based on the later-phase remote sensing image and the fixed base period patch j to obtain the fixed base period patch k and the fixed base period patch l.

[0140] In its specific implementation, this embodiment takes into account both the fixed base period variation patch j and the deduplication process of the rolling base period patch. (Refer to...) Figure 3 and Figure 7 As shown, in this embodiment, semantic segmentation is performed on the latest monitoring image B to extract semantically segmented patches, such as semantically segmented bulldozing patches. Then, the semantically segmented bulldozing patches and fixed base period patches j are combined to extract patches, resulting in fixed base period patches k and l.

[0141] In an optional embodiment, this embodiment performs semantic segmentation extraction based on the later-phase remote sensing image and the fixed base period patch j to obtain fixed base period patch k and fixed base period patch l. Specifically, it may include: performing semantic segmentation on the later-phase remote sensing image using a preset semantic segmentation model to obtain bulldozing patches; extracting fixed base period patch k that intersects with the bulldozing patches from the fixed base period patch j, and extracting fixed base period patch l that does not intersect with the bulldozing patches from the fixed base period patch j.

[0142] Among them, bulldozing patches are also called semantic segmentation bulldozing patches.

[0143] Step 280: Summarize the previously online map patches to obtain online reference map patches, and perform an intersection comparison between the online reference map patches and the fixed base period map patch l to obtain the fixed base period map patch m.

[0144] In a specific implementation, this embodiment can summarize all previously online transformation detection reference patches to obtain online reference patches, and then perform intersection comparison between the online patches and the fixed base period patch l to obtain the fixed base period patch m.

[0145] Optionally, the above-mentioned cross-comparison between the online reference patch and the fixed base period patch l to obtain the fixed base period patch m may specifically include the following sub-steps:

[0146] Sub-step 2801: Based on the online reference patch, perform easy-to-change removal to obtain the target reference patch.

[0147] Sub-step 2802: Calculate the area intersection ratio based on the target reference patch and the fixed base period patch l to obtain the area intersection ratio value.

[0148] Sub-step 2803: Extract fixed base period patches m from fixed base period patch l whose area intersection ratio is less than the third preset threshold.

[0149] A unified description is provided for sub-steps 2801-2803:

[0150] Preferably, the third preset threshold is 50%.

[0151] In the specific implementation, refer to Figure 3 and Figure 7 As shown, this embodiment performs easy-to-change removal based on the online reference map patches. The easy-to-change removal process includes removing easy-to-change map patches such as bulldozing from the online reference map patches to obtain the target reference map patches. Then, the area intersection ratio between the target reference map patch and the fixed base period map patch l is calculated, and fixed base period map patches m with an area intersection ratio of less than 50% in map patch l are extracted.

[0152] Step 290: Based on the fixed base period patch m, the fixed base period patch k, and the rolling base period patch n, merge and filter by area to obtain the target rolling base period change patch after deduplication.

[0153] In this specific implementation, the fixed base period patch m, the fixed base period patch k, and the rolling base period change patch n are merged to obtain the change patch. Then, the merged change patch is screened by area. The area screening mainly removes patches with an area of ​​less than 200㎡. Finally, the deduplicated patch is obtained, which is the deduplicated rolling base period change patch.

[0154] In practical implementation, this embodiment can perform one-click deduplication using a pre-packaged "deduplication model," thereby achieving deduplication of fixed base period patches and rolling base period patches. For example, using the pre-packaged "deduplication model," relevant vector data is input into each input condition. The tool's usage method is as follows: Figure 8 As shown, the usage steps are as follows: ① In "Input Previous Phase Map_Rolling Base Period", select "Previous Period Image Combination Map"; ② In "Input Post-Phase Map", select "Current Period Image Combination Map"; ③ In "Input Fixed Base Period Patch", select "Fixed Base Period Patch"; ④ In "Input Rolling Base Period Patch", select "Rolling Base Period Patch"; ⑤ In "Input Previous BHJC Patch", select "Previous Period Verification Patch"; ⑥ In "Input Previous Online Patch", select "Previous Period Online Patch"; ⑦ In "Input YYFG Results", select "YYFG Patch"; ⑧ In "Input DLTB_NMK_01", select "DLTB_Farmland"; ⑨ In "Input DLTB_NMK_Garden", select "DLTB_Garden Land"; ⑩ In "Input Deduplicated Patch", input "Deduplicated BHJC Patch" and its storage location.

[0155] Furthermore, for one-click deduplication, this embodiment can utilize the model builder in ArcGIS to write a deduplication model system and embed it into the ArcGIS toolbox for execution. In routine monitoring work, a one-click operation removes suspected changed map features that overlap with previous map features and require manual verification, thus achieving one-click deduplication.

[0156] Therefore, this embodiment uses information technology to quickly and accurately extract changed patches within the target area, avoiding redundant verification work caused by extracting duplicate patches from multiple images, and improving the recall rate of change detection.

[0157] Step 300: Perform routine monitoring of the target area based on the fixed base period patch j and the target rolling base period patch.

[0158] This embodiment utilizes the deduplicated fixed base period patch j and the target rolling base period patch to conduct routine monitoring of the target area. This provides strong support for routine monitoring, achieving the monitoring objectives of "real-time monitoring, monthly clearing and quarterly verification, and annual summary." It enhances the ability of survey monitoring to support natural resource business management and provides technical support for comprehensive support of precise management and scientific decision-making in natural resources.

[0159] In summary, this application embodiment acquires multiple remote sensing images of the target area, including fixed base period remote sensing images, rolling base period remote sensing images, and the latest monitored post-temporal remote sensing images. These multiple images are then combined and their changes are extracted separately to obtain fixed base period patch a and rolling base period patch n. Next, fixed base period patch a is deduplicated by extracting fixed base period patches b outside the change survey area from fixed base period patch a, and by extracting an image range D whose temporal phase of the post-temporal remote sensing image is at least a second preset time ahead of the rolling base period remote sensing image. Based on image range D, fixed base period patches c that fall within image range D and fixed base period patches d that do not fall within image range D are extracted from fixed base period patch b. Deduplication is then performed based on the previously verified patches, fixed base period patch c, and fixed base period patch d to obtain the deduplicated fixed base period patch j. For deduplication of rolling base period patches, this embodiment performs semantic segmentation extraction based on the later-phase remote sensing image and the fixed base period patch j to obtain fixed base period patch k and fixed base period patch l. Then, it summarizes the previously online patches to obtain online reference patches. Furthermore, it performs intersection comparison between the online reference patches and the fixed base period patch l to obtain fixed base period patch m. Based on the fixed base period patch m, fixed base period patch k, and rolling base period patch n, it performs merging and area filtering to obtain the deduplicated target rolling base period change patch. This achieves deduplication of fixed base period change patches and rolling base period change patches, and thus realizes a method and system for rapid and accurate extraction and deduplication of routine monitoring and verification patches. It avoids redundant verification work caused by extracting duplicate patches from multiple images, improves the recall rate of change detection, and solves various problems existing in the existing technology for change detection of base period patches.

[0160] Furthermore, this application embodiment uses the fixed base period patch j extracted by deduplication and the target rolling base period patch to conduct routine monitoring of the target area, thereby providing strong support for routine monitoring and providing proactive, intelligent and comprehensive basic data for natural resource surveys, monitoring and management, thus forming a routine monitoring work system that serves natural resource management.

[0161] It should be noted that, for the sake of simplicity, the method embodiments are described as a series of actions. However, those skilled in the art should know that the embodiments of this application are not limited to the described order of actions, because according to the embodiments of this application, some steps may be performed in other orders or simultaneously.

[0162] like Figure 9 As shown in the illustration, this application also provides an automatic change detection and processing device 900 for multi-period remote sensing images, comprising:

[0163] The multi-phase remote sensing image acquisition module 910 is used to acquire fixed base-phase remote sensing images, rolling base-phase remote sensing images, and the latest monitored post-phase remote sensing images of the target area.

[0164] The change extraction module 920 is used to perform change extraction based on the fixed base period remote sensing image and the subsequent time phase remote sensing image to obtain a fixed base period patch a, and to perform change extraction based on the rolling base period remote sensing image and the subsequent time phase remote sensing image to obtain a rolling base period patch n.

[0165] The first deduplication module 930 is used to perform deduplication processing based on the fixed base period patch a and the previously checked patches to obtain the fixed base period patch j.

[0166] The second deduplication module 940 is used to perform deduplication processing based on the rolling base period patch n, combined with the fixed base period patch j and previously online patches, to obtain the target rolling base period patch.

[0167] The routine monitoring module 950 is used to perform routine monitoring of the target area based on the fixed base period patch j and the target rolling base period patch; wherein, the fixed base period patch j and the target rolling base period patch are both base period change patches.

[0168] Optionally, the multi-period remote sensing image acquisition module 910 includes:

[0169] The fixed base period remote sensing image extraction submodule is used to extract land change survey images of the target area from the preset land change survey database as fixed base period remote sensing images;

[0170] The rolling base period remote sensing image acquisition submodule is used to acquire the initial post-temporal images of the previous batch as rolling base period remote sensing images;

[0171] The post-temporal remote sensing image acquisition submodule is used to acquire the latest monitoring images as post-temporal remote sensing images.

[0172] Optionally, the change extraction module 920 includes:

[0173] The image selection submodule is used to select images based on the rolling base period remote sensing images and the subsequent time phase remote sensing images to obtain preferred monitoring images. The preferred monitoring images are monitoring images whose latest monitoring image time phase is better than the previous monitoring image time phase by a first preset time.

[0174] The patch extraction submodule is used to extract patches based on the preferred monitoring image and the rolling base period remote sensing image to obtain the rolling base period patch n.

[0175] Optionally, the first deduplication module 930 includes:

[0176] The "Verified Patch Acquisition" submodule is used to acquire previously verified patches.

[0177] The Change Survey Scope Extraction Submodule is used to extract the change survey scope from a preset change survey database;

[0178] The fixed base period patch b extraction submodule is used to extract the fixed base period patch b outside the scope of the change investigation from the fixed base period patch a;

[0179] The image range D extraction submodule is used to extract the image range D of the later-phase remote sensing image whose temporal phase is better than the rolling base period remote sensing image by at least a second preset time.

[0180] The image patch c and image patch d extraction submodule extracts fixed base period image patch c and fixed base period image patch d from the fixed base period image patch b according to the image range D;

[0181] The first deduplication submodule is used to perform deduplication based on the previously verified map patches, fixed base period map patch c, and fixed base period map patch d to obtain fixed base period map patch j.

[0182] Optionally, the fixed base period patch b extraction submodule includes:

[0183] The first semantic segmentation unit is used to perform full-element land use semantic segmentation on the fixed base period patch a using a preset semantic segmentation model and an intelligent remote sensing interpretation algorithm to obtain image land use information, wherein the image land use information includes image land use and image land use range;

[0184] The first comparison unit is used to compare the image land cover information with the change survey scope and extract the fixed base period map patch b outside the change survey scope from the fixed base period map patch a.

[0185] Optionally, the first deduplication submodule includes:

[0186] The summary detection unit is used to perform summary detection based on the previously verified map patches to obtain verified reference map patches;

[0187] The intersection comparison unit is used to perform intersection comparison on the fixed base period patch c and the fixed base period patch d respectively according to the verified reference patch, to obtain the fixed base period patch e corresponding to the fixed base period patch c and the fixed base period patch f corresponding to the fixed base period patch d;

[0188] The image range E extraction unit is used to extract the image range E that is three preset times different from the time phase of the latest monitoring image;

[0189] The image patch g and image patch h extraction unit is used to extract the fixed base period image patch g and the fixed base period image patch h from the fixed base period image patch f according to the image range E;

[0190] The patch i extraction unit is used to extract fixed base period patch i from the fixed base period patch g, which falls within the cultivated land range of the image range E;

[0191] The first deduplication unit is used to merge the fixed base period patch e, the fixed base period patch h, and the fixed base period patch i to obtain the deduplicated fixed base period patch j; wherein, the fixed base period patch e is a base period patch extracted from the fixed base period patch c whose area intersection ratio is less than a first preset threshold, and the fixed base period patch f is a base period patch extracted from the fixed base period patch d whose area intersection ratio is less than a second preset threshold.

[0192] Optionally, the second deduplication module 940 includes:

[0193] The semantic segmentation submodule is used to perform semantic segmentation extraction based on the post-temporal remote sensing image and the fixed base period patch j to obtain the fixed base period patch k and the fixed base period patch l.

[0194] The "Reference Map Summary Submodule" is used to summarize the previously launched reference maps to obtain the online reference maps.

[0195] The intersection comparison submodule is used to perform an intersection comparison based on the online reference patch and the fixed base period patch l to obtain the fixed base period patch m;

[0196] The merging and area filtering submodule is used to merge and filter the fixed base period patch m, the fixed base period patch k, and the rolling base period patch n to obtain the target rolling base period change patch after deduplication.

[0197] Optionally, the semantic segmentation submodule includes:

[0198] The second semantic segmentation unit is used to perform semantic segmentation on the post-temporal remote sensing image using a preset semantic segmentation model to obtain bulldozing patches.

[0199] The patch k and patch l extraction unit is used to extract the fixed base period patch k that intersects with the bulldozing patch from the fixed base period patch j, and to extract the fixed base period patch l that does not intersect with the bulldozing patch from the fixed base period patch j.

[0200] Optionally, the intersection comparison submodule includes:

[0201] The variable rejection unit is used to perform variable rejection based on the online reference patch to obtain the target reference patch;

[0202] The intersection ratio calculation unit is used to calculate the area intersection ratio based on the target reference patch and the fixed base period patch l, and obtain the area intersection ratio value.

[0203] The patch m extraction unit is used to extract fixed base period patches m from fixed base period patches l whose area intersection ratio is less than a third preset threshold.

[0204] It should be noted that the automatic change detection and processing device for multi-period remote sensing images provided in the embodiments of this application can execute the automatic change detection and processing method for multi-period remote sensing images provided in any embodiment of this application, and has the corresponding functions and beneficial effects of executing the automatic change detection and processing method for multi-period remote sensing images.

[0205] In practical implementation, the aforementioned automatic change detection and processing device for multi-period remote sensing images can be integrated into the equipment. This allows the equipment to use information technology to quickly and accurately extract changed patches within the target area from multi-period remote sensing images and perform deduplication, avoiding redundant verification work caused by extracting duplicate patches from multiple images. As an electronic device, it reduces the occurrence of missed patches and improves the recall rate of change detection. This electronic device can consist of two or more physical entities, or it can consist of a single physical entity. For example, the electronic device can be a personal computer (PC), a computer, a server, etc. This application embodiment does not impose specific limitations in this regard.

[0206] like Figure 10As shown, this application embodiment provides an electronic device, including a processor 111, a communication interface 112, a memory 113, and a communication bus 114. The processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114. The memory 113 is used to store computer programs. When the processor 111 executes the program stored in the memory 113, it implements the steps of the multi-period remote sensing image automatic change detection and processing method provided in any of the aforementioned method embodiments. For example, the steps of the automatic change detection and processing method for multi-period remote sensing images may include the following steps: acquiring a fixed base period remote sensing image, a rolling base period remote sensing image, and the latest monitored post-period remote sensing image of the target area; performing change extraction based on the fixed base period remote sensing image and the post-period remote sensing image to obtain a fixed base period patch a; and performing change extraction based on the rolling base period remote sensing image and the post-period remote sensing image to obtain a rolling base period patch n; performing deduplication processing based on the fixed base period patch a and previously verified patches to obtain a fixed base period patch j; performing deduplication processing based on the rolling base period patch n and the fixed base period patch j and previously online patches to obtain a target rolling base period patch; and performing routine monitoring of the target area based on the fixed base period patch j and the target rolling base period patch; wherein, both the fixed base period patch j and the target rolling base period patch are base period change patches.

[0207] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the automatic change detection and processing method for multi-period remote sensing images as provided in any of the foregoing method embodiments.

[0208] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0209] The above description is merely a specific embodiment of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method for automatic change detection and processing of multi-period remote sensing images, characterized in that, include: Acquire fixed base period remote sensing images, rolling base period remote sensing images, and the latest monitored post-temporal remote sensing images of the target area; Change extraction is performed based on the fixed base period remote sensing image and the subsequent time phase remote sensing image to obtain fixed base period patch a; and change extraction is performed based on the rolling base period remote sensing image and the subsequent time phase remote sensing image to obtain rolling base period patch n. Based on the fixed base period patch a, and combined with previously verified patches, duplicates are removed to obtain the fixed base period patch j; Based on the rolling base period patch n, and combined with the fixed base period patch j and previously online patches, deduplication is performed to obtain the target rolling base period patch; The target area is routinely monitored based on the fixed base period patch j and the target rolling base period patch; Wherein, both the fixed base period patch j and the target rolling base period patch are base period change patches; Based on the fixed base period patch a, and combined with previously verified patches, a deduplication process is performed to obtain fixed base period patch j. This includes: acquiring previously verified patches and extracting the change survey range from a preset change survey database; extracting fixed base period patches b outside the change survey range from the fixed base period patch a; and extracting an image range D whose temporal phase of the later-phase remote sensing image is at least a second preset time better than the rolling base period remote sensing image; extracting fixed base period patches c and d from the fixed base period patch b based on the image range D; and performing deduplication based on the previously verified patches, fixed base period patches c, and fixed base period patches d to obtain fixed base period patch j. Based on the rolling base period patch n, and combined with the fixed base period patch j and previously online patches, deduplication is performed to obtain the target rolling base period patch, including: semantic segmentation extraction based on the later-phase remote sensing image and the fixed base period patch j to obtain fixed base period patch k and fixed base period patch l; summarizing based on the previously online patches to obtain online reference patches; and performing intersection comparison based on the online reference patches and the fixed base period patch l to obtain fixed base period patch m; merging and area filtering based on the fixed base period patch m, the fixed base period patch k, and the rolling base period patch n to obtain the deduplicated target rolling base period change patch.

2. The method according to claim 1, characterized in that, Acquire fixed-base-period remote sensing images, rolling-base-period remote sensing images, and the latest monitored post-temporal remote sensing images of the target area, including: Land change survey images of the target area are extracted from the pre-set land change survey database to serve as fixed base period remote sensing images; The initial post-temporal images from the previous batch are acquired as rolling base period remote sensing images, and the latest monitoring images are acquired as post-temporal remote sensing images.

3. The method according to claim 1, characterized in that, Changes are extracted based on the rolling base period remote sensing image and the subsequent time-phase remote sensing image to obtain the rolling base period patch n, including: Based on the rolling base period remote sensing image and the subsequent time phase remote sensing image, image selection is performed to obtain the preferred monitoring image. The preferred monitoring image is the monitoring image whose latest monitoring image time phase is better than the previous monitoring image time phase by a first preset time. Based on the preferred monitoring image and the rolling base period remote sensing image, patch extraction is performed to obtain the rolling base period patch n.

4. The method according to claim 1, characterized in that, Extracting fixed base period map patches b outside the scope of the change survey from the fixed base period map patch a includes: Using a preset semantic segmentation model, the fixed base period map patch a is subjected to full-element land use semantic segmentation using an intelligent remote sensing interpretation algorithm to obtain image land use information, which includes image land use and image land use range. The image land cover information is compared with the change survey scope, and the fixed base period map patch b outside the change survey scope is extracted from the fixed base period map patch a.

5. The method according to claim 1, characterized in that, Based on the previously verified map patches, fixed base period map patch c, and fixed base period map patch d, deduplication is performed to obtain fixed base period map patch j, including: Based on the previously verified map patches, a summary detection was performed to obtain verified reference map patches; Based on the verified reference map patch, the fixed base period map patch c and the fixed base period map patch d are cross-compared to obtain the fixed base period map patch e corresponding to the fixed base period map patch c and the fixed base period map patch f corresponding to the fixed base period map patch d; Extract the image range E whose time phase differs from the time phase of the latest monitoring image by a third preset time; Based on the image range E, extract fixed base period patch g and fixed base period patch h from the fixed base period patch f; Extract the fixed base period patch i that falls within the cultivated land area of ​​the image range E from the fixed base period patch g; Based on the fixed base period patch e, the fixed base period patch h, and the fixed base period patch i, the patches are merged to obtain the deduplicated fixed base period patch j; Wherein, the fixed base period patch e is a base period patch extracted from the fixed base period patch c whose area intersection ratio is less than a first preset threshold, and the fixed base period patch f is a base period patch extracted from the fixed base period patch d whose area intersection ratio is less than a second preset threshold.

6. The method according to claim 1, characterized in that, Semantic segmentation and extraction are performed based on the later-phase remote sensing image and the fixed base period patch j to obtain fixed base period patch k and fixed base period patch l, including: The post-temporal remote sensing image is semantically segmented using a preset semantic segmentation model to obtain bulldozing patches; Extract fixed base period patch k that intersects with the bulldozing patch from the fixed base period patch j, and extract fixed base period patch l that does not intersect with the bulldozing patch from the fixed base period patch j.

7. The method according to claim 1, characterized in that, Based on the intersection comparison between the online reference patch and the fixed base period patch l, the fixed base period patch m is obtained, including: Based on the already online reference patches, easily changeable patches are removed to obtain the target reference patches; The area intersection ratio is calculated based on the target reference patch and the fixed base period patch l to obtain the area intersection ratio value; Extract fixed base period patches m from fixed base period patch l whose area intersection ratio is less than the third preset threshold.

8. An automatic change detection and processing device for multi-period remote sensing images, characterized in that, include: The multi-phase remote sensing image acquisition module is used to acquire fixed base-phase remote sensing images, rolling base-phase remote sensing images, and the latest monitored post-phase remote sensing images of the target area. The change extraction module is used to extract changes based on the fixed base period remote sensing image and the subsequent time phase remote sensing image to obtain a fixed base period patch a, and to extract changes based on the rolling base period remote sensing image and the subsequent time phase remote sensing image to obtain a rolling base period patch n. The first deduplication module is used to perform deduplication processing based on the fixed base period patch a and the previously verified patches to obtain the fixed base period patch j. The second deduplication module is used to perform deduplication processing based on the rolling base period patch n, combined with the fixed base period patch j and previously online patches, to obtain the target rolling base period patch. The routine monitoring module is used to perform routine monitoring of the target area based on the fixed base period patch j and the target rolling base period patch; Wherein, both the fixed base period patch j and the target rolling base period patch are base period change patches; The first deduplication processing module includes: a verified patch acquisition submodule, used to acquire previously verified patches; a change survey range extraction submodule, used to extract change survey ranges from a preset change survey database; a fixed base period patch b extraction submodule, used to extract fixed base period patches b outside the change survey range from the fixed base period patch a; an image range D extraction submodule, used to extract the image range D of the later-phase remote sensing image whose temporal phase is at least a second preset time better than the rolling base period remote sensing image; a patch c and patch d extraction submodule, used to extract fixed base period patches c and d from the fixed base period patch b based on the image range D; and a first deduplication processing submodule, used to perform deduplication processing based on the previously verified patches, fixed base period patches c and d to obtain fixed base period patch j; The second deduplication processing module includes: a semantic segmentation submodule, used to perform semantic segmentation extraction based on the later-phase remote sensing image and the fixed base period patch j to obtain fixed base period patch k and fixed base period patch l; an online reference patch aggregation submodule, used to aggregate online reference patches based on the previous online patches to obtain online reference patches; an intersection comparison submodule, used to perform intersection comparison based on the online reference patches and the fixed base period patch l to obtain fixed base period patch m; and a merging and area filtering submodule, used to merge and filter based on the fixed base period patch m, the fixed base period patch k, and the rolling base period patch n to obtain the deduplicated target rolling base period change patch.

Images