A dynamic management system and method for the full life cycle of spatiotemporal data of water administration law enforcement

Abstract

The embodiments of the present application relate to the technical field of water administration and law enforcement management, and in particular to a dynamic management system and method for the entire life cycle of spatiotemporal data of water administration and law enforcement. The system comprises: a storage module for reading input remote sensing image files; an identification module for interpreting information from the stored remote sensing image files, identifying water administration and law enforcement objects within the management scope, and forming suspected spots of each water administration and law enforcement object; a dynamic tracking module for monitoring the growth and decline of each water administration and law enforcement object through multi-temporal remote sensing image files; an extinction module for comparing each water administration and law enforcement object before and after rectification, and recording the extinction time when the comparison finds that the water administration and law enforcement object does not exist or disappears after rectification; and a full-cycle tracing module for tracing the entire life cycle of the first discovered water administration and law enforcement object, including its generation, increase, decrease, extinction, and reappearance. This system effectively solves the problem of insufficient flexibility and pertinence in tracing the source of water administration and law enforcement objects.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of water administration law enforcement management, and in particular to a dynamic management system and method for the full life cycle of water administration law enforcement spatiotemporal data. Background Art

[0002] The traditional method of identifying problems through manual inspections is limited by factors such as traffic conditions and is difficult to achieve full coverage. There are blind spots during inspections, and the number of problems is unclear. At the same time, manual inspections take a long time, and some water-related illegal activities develop very quickly. By the time problems are discovered, they are already very difficult to investigate and deal with. Manual inspection methods are backward and cannot promptly detect, prevent, and investigate problems. This can lead to the escalation of the situation, have a significant impact on flood control safety and ecological protection in the basin, and increase the difficulty of water administration law enforcement. How to fully record the spatial characteristics and temporal changes of river monitoring objects and the changes in their attributes throughout their life cycle, promptly detect and deal with water-related illegal activities in the river, prevent the situation from escalating, and strengthen the dynamic supervision of information on river law enforcement monitoring objects has become an urgent problem to be solved. Summary of the Invention

[0003] In view of this, the embodiments of the present application propose a dynamic management system and method for the entire life cycle of spatiotemporal data of water administration law enforcement, which can realize full-process, full-dimensional, full-cycle, and blind-spot-free dynamic management of water administration law enforcement objects with multi-temporal, multi-spatial scale, and multi-dimensional business attributes and relationships, effectively solving the current problems of insufficient flexibility and pertinence in the traceability of law enforcement objects, facilitating better correlation analysis of data between various time phases, and providing the most direct evidence support and disposal suggestions for the decision-making of water administration law enforcement departments.

[0004] In order to achieve the above-mentioned purpose, the embodiment of the present application proposes a dynamic management system for the full life cycle of water administration law enforcement spatiotemporal data, including: a storage module for reading the input remote sensing image file, the attribute information of the input remote sensing image file needs to include image number, image name, image shooting date, image type, image location description and storage time; an identification module for interpreting the information of the stored remote sensing image file, identifying the water administration law enforcement objects within the management scope, forming suspected spots of each water administration law enforcement object, and generating water administration law enforcement spatiotemporal data; a dynamic tracking module for using multi-temporal remote sensing image files and on-site data , monitor the growth and decline of each water administrative law enforcement object, and record the spatial nodes and time nodes of the growth and decline; the extinction module is used to display the rectification status of each water administrative law enforcement object, and compare the water administrative law enforcement objects before and after the rectification. When the comparison finds that the water administrative law enforcement object does not exist or disappears after the rectification, the extinction time is recorded; the full-cycle tracing module is used to trace the full life cycle of the generation, increase, decrease, extinction and reappearance of the water administrative law enforcement object discovered for the first time, and at the same time update the maximum target spatial range of the water administrative law enforcement object during its existence in real time, and continue to track the water administrative law enforcement objects that are subsequently generated within the said maximum target spatial range.

[0005] In order to achieve the above-mentioned purpose, the embodiment of the present application also proposes a dynamic management method for the entire life cycle of water administration law enforcement spatiotemporal data, including: reading the input remote sensing image file, the attribute information of the input remote sensing image file needs to include image number, image name, image shooting date, image type, image location description and storage time; interpreting the information of the stored remote sensing image file, identifying the water administration law enforcement objects within the management scope, forming suspected spots of each water administration law enforcement object, and generating water administration law enforcement spatiotemporal data; monitoring the growth and decline of each water administration law enforcement object through multi-temporal remote sensing image files and field data, and recording the spatial nodes and time nodes of the growth and decline; displaying the rectification status of each water administration law enforcement object, and comparing the water administration law enforcement objects before and after the rectification. When the comparison finds that the water administration law enforcement object does not exist or disappears after the rectification, the extinction time is recorded; tracing the entire life cycle of the first discovered water administration law enforcement object, including generation, increase, decrease, extinction and reappearance, and at the same time updating the maximum target spatial range of the water administration law enforcement object during its existence in real time, and continuously tracking the water administration law enforcement objects that are subsequently generated again within the said maximum target spatial range.

[0006] In order to achieve the above-mentioned purpose, an embodiment of the present application also proposes a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it can implement a dynamic management method for the entire life cycle of water administrative law enforcement spatiotemporal data as described above.

[0007] The embodiments of the present application propose a dynamic management system and method for the full life cycle of spatiotemporal data of water administration law enforcement, which constructs a dynamic management system for the full life cycle of spatiotemporal data of water administration law enforcement composed of a warehousing module, an identification module, a dynamic tracking module, an extinction module and a full-cycle tracing module. It can conduct full-process, full-dimensional, full-cycle and blind-spot dynamic management of water administration law enforcement objects with multi-temporal, multi-spatial scales, multi-dimensional business attributes and relationships, effectively solving the current problems of insufficient flexibility and pertinence in the traceability of water administration law enforcement objects, facilitating better correlation analysis of data between different time phases, and providing the most direct evidence support and disposal suggestions for the decision-making of water administration law enforcement departments.

[0008] Optionally, the recognition module is composed of an interpretation unit, a correction unit, a discrimination execution unit and an interpretation information library. The interpretation information library stores a first-time discovered object table and a suspected object interpretation result table. The first-time discovered object table is used to record the first-time discovered object and its basic data attributes, and the suspected object interpretation result table is used to record the basic data attributes, first-time discovered mark attributes and historical change attributes of each water administration law enforcement object; the interpretation unit is used to use a semantic segmentation model based on a hierarchical Transformer to interpret information on remote sensing image files to obtain interpretation results of identified water administration law enforcement objects. The interpretation results include the type and spatial range of the suspected patches of identified water administration law enforcement objects; the correction unit is used to obtain correction information from human experts through human-computer interaction, and based on the obtained correction information from human experts, to identify The interpretation results of the water administrative law enforcement objects identified are corrected; the judgment execution unit is used to compare the spatial range of the corrected suspected spots with the coverage of each water administrative law enforcement object in the first-discovered object table; if the spatial range of the corrected suspected spots does not overlap with the coverage of each water administrative law enforcement object in the first-discovered object table, the identified water administrative law enforcement object is determined to be the first-discovered object and is newly added to the first-discovered object table and the suspected object interpretation results table; if the spatial range of the corrected suspected spots overlaps with the coverage of each water administrative law enforcement object in the first-discovered object table, the overlapping water administrative law enforcement object is determined to be the target water administrative law enforcement object, and the interpretation results of the identified water administrative law enforcement object are entered into the historical change attributes of the target water administrative law enforcement object in the suspected object interpretation results table.

[0009] Optionally, a semantic segmentation model based on a hierarchical Transformer is composed of an encoder and a decoder. The encoder adopts a hierarchical Transformer structure, which consists of an image block embedding layer, four sequentially connected Transformer layers, and a feature fusion layer. The input of the first Transformer layer is the output of the image block embedding layer, the input of the second Transformer layer is the output of the first Transformer layer, the input of the third Transformer layer is the output of the second Transformer layer, the input of the fourth Transformer layer is the output of the third Transformer layer, and the input of the feature fusion layer is the output of each Transformer layer. Each Transformer layer contains N Transformer blocks, and each Transformer layer has N blocks. The Transformer blocks all contain multi-head self-attention blocks and hybrid forward propagation networks, and the decoder uses a multi-layer perceptron; the image block embedding layer is used to segment the input remote sensing image files, retaining only the image blocks of the river range that needs to be monitored, and input the retained image blocks into the first Transformer layer; each Transformer layer is used to extract features from its own input to obtain feature maps of different scales; the feature fusion layer is used to fuse the feature maps output by each Transformer layer, down-sample the feature maps whose resolution needs to be reduced through a 3×3 convolution operation, and up-sample the feature maps whose resolution needs to be increased using DUpsampling to obtain a fused feature map; the multi-layer perceptron is used to adjust the channels of the fused feature map to obtain the semantic segmentation result of the input remote sensing image file, that is, the interpretation result of the identified water administrative law enforcement object.

[0010] Optionally, the dynamic tracking module is composed of an object change monitoring unit, a change state analysis unit and an on-site inspection and tracking unit; the object change monitoring unit is used to compare and analyze the remote sensing images of the two phases before and after the same geographical location, and input the remote sensing images of the two phases before and after into the change detection model based on the double-branch Segformer and multi-scale attention respectively, and obtain the change detection results output by the change detection model based on the double-branch Segformer and multi-scale attention, and call the recognition module to interpret the remote sensing image of the later phase based on the change detection result, and obtain the interpretation results of the remote sensing image of the later phase; the change state analysis unit is used to interpret the remote sensing image of the later phase based on the change detection result. The interpretation results of the remote sensing images of the two phases before and after are used to analyze the change state of the water administration law enforcement object, and the change state analysis results are obtained and recorded. The attribute information of the change state analysis results includes the generation time, change type, change time, creation time and update time; the on-site inspection and tracking unit is used to use satellite positioning technology to capture and record the precise position and inspection path of the water administration law enforcement personnel in real time, and provide the suspected spots, change detection feature maps and change state analysis results of the two phases before and after the same water administration law enforcement object to the water administration law enforcement personnel, so that the water administration law enforcement personnel can conduct on-site investigation and confirm the change state of the water administration law enforcement object to correct the change state analysis results.

[0011] Optionally, a change detection model based on dual-branch Segformer and multi-scale attention consists of an encoder and a decoder. The encoder consists of two Segformer networks with shared weights and a change information extraction module. Each Segformer network consists of an overlapping block merging layer, four sequentially connected Transformer layers and four CCNet attention modules. For each Segformer network, the input of the first Transformer layer is the output of the overlapping block merging layer, the input of the second Transformer layer is the output of the first Transformer layer, the input of the third Transformer layer is the output of the second Transformer layer, and the input of the fourth Transformer layer is the output of the third Transformer layer. The inputs of the four CCNet attention modules are the outputs of the four Transformer layers respectively. Each Transformer layer contains N Transformer blocks, and each Transformer block contains a multi-head self-attention block and a hybrid forward propagation network; the front and back two The remote sensing images of each phase are respectively input into two Segformer networks based on the change detection model of dual-branch Segformer and multi-scale attention; the overlapping block merging layer is used to block the image of its own input, retaining only the image blocks of the river range that needs to be monitored, and the retained image blocks are input into the first layer of Transformer layer; each layer of Transformer layer is used to extract features from its own input to obtain semantic feature maps of different scales; the CCNet attention module is used to perform affinity operations and long-range context information aggregation operations on the input semantic feature map to obtain an enhanced semantic feature map. The enhanced semantic feature maps output by the four CCNet attention modules of the same Segformer network are spliced ​​and fused to obtain the final semantic feature map and input it into the change information extraction module; the change information extraction module is used to perform absolute difference operations on the final semantic feature maps output by the two Segformer networks to extract change features, and the extracted change features are input into the decoder; the decoder is used to perform binary change detection on the input change features, and after passing through a 3×3 convolution layer and then a 1×1 convolution layer, it outputs a change detection feature map.

[0012] Optionally, the extinction module is implemented and used by the superior water administration regulatory department, which is mainly responsible for displaying the rectification status of each water administration law enforcement object, the comparison before and after the rectification, and the historical suspected spots of each water administration law enforcement object, and feedbacking the water administration law enforcement results based on the comparison before and after the rectification, and evaluating whether the water administration law enforcement effect has achieved the expected goals. For water administration law enforcement objects that have achieved the expected goals, no longer exist or have disappeared after rectification, they will be accepted and confirmed to be extinct, and the extinction time will be recorded in the suspected object interpretation result library. Otherwise, they will be returned to the grassroots water administration law enforcement unit for further rectification.

[0013] Optionally, the full-cycle tracing module supports retrieval of first-discovered objects based on the first-discovered object codes input by the user, queries detailed information of the retrieved first-discovered objects in the interpretation information library, and traces the entire life cycle of the retrieved first-discovered objects, including their generation, increase, decrease, extinction, and reappearance. All collected information is sorted in chronological order, cleaned and formatted, and then presented to users through timelines, maps, status change lists, and interactive charts.

[0014] Optionally, the timeline display uses a timeline to display the entire life cycle of the water administration law enforcement object, and marks each key time point on the timeline, including the first discovery time, change time, extinction time and reappearance time, supporting users to view the corresponding detailed information by clicking on the key time points on the timeline; the map display displays the spatial position changes of the water administration law enforcement object on the map, uses different colors or icons to represent the different states of the water administration law enforcement object, and dynamically updates the display on the map as the timeline moves. For the maximum spatial range, a transparent polygonal area or circular area is marked on the map, and dynamically adjusted as the data is updated; the state change list display provides a list view that lists the detailed information of all state change events, supporting users to view and operate by clicking on the list items in the list view; the interactive chart display uses interactive charts to display the state change trends and statistical information of water administration law enforcement objects, including the increase and decrease trends of water administration law enforcement objects and the distribution of different state types. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the related technologies, the following is a brief introduction to the drawings required for use in the embodiments of the present application or the description of the related technologies. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0016] Figure 1 This is a schematic diagram of the structure of a dynamic management system for the entire life cycle of spatiotemporal data of water administration law enforcement provided in one embodiment of the present application;

[0017] Figure 2 This is a schematic diagram of the working principle of a dynamic management system for the full life cycle of spatiotemporal data of water administration law enforcement provided in one embodiment of the present application;

[0018] Figure 3 1 is a schematic diagram of the model structure of a semantic segmentation model based on a hierarchical Transformer provided in one embodiment of the present application;

[0019] Figure 4 1 is a schematic diagram of the structure of a feature fusion layer of a semantic segmentation model based on a hierarchical Transformer provided in one embodiment of the present application;

[0020] Figure 5 is a structural diagram of a dynamic tracking module provided in one embodiment of the present application;

[0021] Figure 6 1 is a schematic diagram of a model structure of a change detection model based on a dual-branch Segformer and multi-scale attention provided in one embodiment of the present application;

[0022] Figure 7 This is a schematic diagram of the change of the suspected pattern provided in an embodiment of the present application

[0023] Figure 8 This is a display diagram on a handheld mobile device of a water administration law enforcement officer provided in one embodiment of the present application;

[0024] Figure 9 1 is a schematic diagram showing the comparison before and after rectification provided in an embodiment of the present application;

[0025] Figure 10 This is a flowchart of a method for dynamic management of spatiotemporal data of water administration law enforcement throughout its life cycle provided in another embodiment of the present application. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, each embodiment of the present application will be described in detail below with reference to the accompanying drawings. However, it will be understood by those skilled in the art that in each embodiment of the present application, many technical details are proposed to enable the reader to better understand the present application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical solutions claimed in the present application can be implemented. The division of the following embodiments is for convenience of description and should not constitute any limitation on the specific implementation of the present application. The various embodiments can be combined and referenced with each other under the premise of no contradiction.

[0027] An embodiment of the present application proposes a dynamic management system for the entire life cycle of spatiotemporal data of water administrative law enforcement. The implementation details of the dynamic management system for the entire life cycle of spatiotemporal data of water administrative law enforcement proposed in this embodiment are described in detail below. The following content is only the implementation details provided for the convenience of understanding and is not necessary for the implementation of this solution.

[0028] The specific structure of a dynamic management system for the full life cycle of spatiotemporal data of water administration law enforcement proposed in this embodiment can be as follows: Figure 1 As shown, its working principle can be as follows Figure 2 As shown, the present embodiment proposes a dynamic management system for the entire life cycle of spatiotemporal data of water administration law enforcement, including: an warehousing module 101, an identification module 102, a dynamic tracking module 103, an extinction module 104 and a full-cycle tracing module 105.

[0029] The storage module 101 is used to read the input remote sensing image file. The attribute information of the input remote sensing image file needs to include image number, image name, image shooting date, image type, image location description and storage time.

[0030] The identification module 102 is used to interpret the information of the stored remote sensing image files, identify the water administration law enforcement objects within the management scope, form suspected spots of each water administration law enforcement object, and generate water administration law enforcement spatiotemporal data.

[0031] The dynamic tracking module 103 is used to monitor the changes in the water administration law enforcement objects through multi-temporal remote sensing image files and on-site data, and record the spatial nodes and time nodes of the changes.

[0032] The extinction module 104 is used to display the rectification status of each water administrative law enforcement object and compare the water administrative law enforcement objects before and after the rectification. When the comparison finds that the water administrative law enforcement object does not exist or disappears after the rectification, the extinction time is recorded.

[0033] The full-cycle tracing module 105 is used to trace the entire life cycle of the first discovered water administration law enforcement object, including its generation, increase, decrease, extinction, and reappearance, while updating and recording the maximum target spatial range of the water administration law enforcement object during its existence in real time, and continuously tracking the subsequent water administration law enforcement objects generated again within the said maximum target spatial range.

[0034] In one example, the identification module consists of an interpretation unit, a correction unit, a discrimination execution unit and an interpretation information library. The interpretation information library stores a table of first-discovered objects and a table of suspected object interpretation results. The table of first-discovered objects is used to record the first-discovered objects and their basic data attributes. The basic data attributes include but are not limited to the map object code, name, type, administrative division, regulatory unit, discovery time, spatial range, area and image number, etc. The table of suspected object interpretation results is used to record the basic data attributes, first-discovery mark attributes and historical change attributes of each water administrative law enforcement object. The first-discovery mark attributes include whether it is the first time to be discovered in the range and the object number of the first discovery. The historical change attributes include but are not limited to the change type (new addition, expansion, reduction, type change, extinction), generation time, elimination time, elimination type (completely cleared, meeting expectations), creation time and update time, etc.

[0035] Among them, the object code is the ID information in the database, and the recognition module will establish a unique code for each suspected patch, that is, a key-value item. For the object name, if it is a known object name (such as flood control projects, ecological parks, amusement parks, etc.), the known object name is used directly. If it is not a flood control project, the object type name is used as the name. Specifically, the object name is obtained by associating the object's geographical location with the existing object name and information library table. The object type is one of the 22 types. The 22 types are river regulation projects, wave breakwaters, water intake (drainage) projects, cross-river bridges, riverside roads, forests, ponds, enclosure farming, ecological parks and amusement parks, sand mining ships, sand mining sites, oil wells, well platforms, waste slag, river enclosure projects, production embankments, river-crossing pipelines, riverside factories (houses), brick and tile kiln factories, docks, photovoltaic power plants, and river dams (rubber dams). The spatial range refers to the vector data of the suspected patch outline. The attribute of the administrative division is obtained by comparing and associating with the administrative division vector map based on the geographical location and spatial scope of the water administrative law enforcement object. The attribute of the supervisory unit is obtained by comparing with the supervisory scope of the water administrative law enforcement unit based on the geographical location and coverage of the water administrative law enforcement object. The time of discovery / first discovery time is the time when the image is taken. The area is obtained by vector calculation based on the spatial scope of the suspected spot. Whether it is the first time to be discovered in this scope, the attribute value is yes or no. If the current spot does not overlap with the spatial scope of the object in the first discovery object library, it is considered to be the first time to be discovered in this scope. Otherwise, it is determined not to be the first time to be discovered in this scope. The first discovered object number, this attribute value is an object code in the first discovered object table.

[0036] The interpretation unit is used to interpret the information of remote sensing image files using a semantic segmentation model based on a hierarchical Transformer to obtain the interpretation results of the identified water administrative law enforcement objects. The interpretation results include the type and spatial range of the suspected patches of the identified water administrative law enforcement objects.

[0037] The correction unit is used to obtain the correction information of human experts through human-computer interaction, and to correct the interpretation results of the identified water administrative law enforcement objects based on the correction information obtained from the human experts.

[0038] The judgment execution unit is used to compare the spatial range of the corrected suspected map with the coverage of each water administrative law enforcement object in the first-discovered object table; if the spatial range of the corrected suspected map does not overlap with the coverage of each water administrative law enforcement object in the first-discovered object table, the identified water administrative law enforcement object is determined to be the first-discovered object and is newly added to the first-discovered object table and the suspected object interpretation result table; if the spatial range of the corrected suspected map overlaps with the coverage of each water administrative law enforcement object in the first-discovered object table, the overlapping water administrative law enforcement object is determined to be the target water administrative law enforcement object, and the interpretation result of the identified water administrative law enforcement object is entered into the historical change attribute of the target water administrative law enforcement object in the suspected object interpretation result table.

[0039] In one example, the semantic segmentation model based on the hierarchical Transformer divides the remote sensing image file into image blocks, retaining only the image blocks containing the river range that needs to be monitored, and performing model inference operations on these image blocks one by one. After obtaining the segmentation results, the segmentation results are merged according to the previous block principle, and after certain post-processing, the segmentation results of the entire image are formed. The model structure of the semantic segmentation model based on the hierarchical Transformer can be as follows: Figure 3 shown.

[0040] The semantic segmentation model based on the hierarchical Transformer consists of an encoder and a decoder. The encoder adopts a hierarchical Transformer structure, which consists of an image block embedding layer, four sequentially connected Transformer layers and a feature fusion layer. The input of the first Transformer layer is the output of the image block embedding layer, the input of the second Transformer layer is the output of the first Transformer layer, the input of the third Transformer layer is the output of the second Transformer layer, the input of the fourth Transformer layer is the output of the third Transformer layer, and the input of the feature fusion layer is the output of each Transformer layer. Each Transformer layer contains N (such as N = 2) Transformer blocks, each Transformer block contains a multi-head self-attention block and a hybrid forward propagation network, and the decoder adopts a multi-layer perceptron.

[0041] The Transformer block uses a window with a width and height of 3, a stride of 2, and an edge padding of 2. The sliding window partitions the image into blocks, preserving local continuity around the blocks through overlap while achieving a multi-scale effect similar to that of a convolutional neural network. A hybrid forward propagation network is also used, introducing 3×3 depthwise separable convolutions within the feedforward network to convey position information. This eliminates traditional position encoding, eliminating the need for interpolation of position encodings and the resulting loss of accuracy when the training and test resolutions differ.

[0042] The image block embedding layer is used to segment the input remote sensing image files into image blocks, retaining only the image blocks of the river range that needs to be monitored, and input the retained image blocks into the first Transformer layer.

[0043] Each Transformer layer is used to extract features from its own input to obtain feature maps of different scales.

[0044] Let h and w be the height and width of the input to the first Transformer layer, respectively. The scales of the feature maps output by the four Transformer layers are: h / 4×w / 4×C1, h / 8×w / 8×C2, h / 16×w / 16×C3, and h / 32×w / 32×C4. The values ​​of C1, C2, C3, and C4 can be 64, 128, 256, and 512, respectively.

[0045] The feature fusion layer is used to fuse the feature maps output by each Transformer layer. The feature maps whose resolution needs to be reduced are downsampled by 3×3 convolution operation, and the feature maps whose resolution needs to be increased are upsampled by DUpsampling to obtain the fused feature maps. The structure of the feature fusion layer can be as follows: Figure 4 shown.

[0046] DUpsampling uses a learnable upsampling method to replace traditional bilinear interpolation. First, feature map encoding is performed, applying N 1×1 convolutions to the input feature map to generate a new encoded feature map with N times the number of channels as the input feature map. Feature map decoding is then performed to reshape the encoded feature map into an output feature map with twice the width and height, reducing the number of channels to 1 / 4. Finally, the result is output, and the decoded feature map serves as the upsampled output.

[0047] The multi-layer perceptron is used to adjust the channels of the fused feature map to obtain the semantic segmentation results of the input remote sensing image file, that is, the interpretation results of the identified water administrative law enforcement objects.

[0048] The feature fusion layer unifies the number of channels of the feature maps of the four scales to C, and then superimposes the channel dimensions to obtain a feature map of scale h / 4×w / 4×C. The multi-layer perceptron adjusts the feature map of scale h / 4×w / 4×C to h / 4×w / 4×N. cls As the semantic segmentation result of the input remote sensing image file. cls is the number of categories to be semantically segmented.

[0049] The multi-layer perceptron consists of a layer of Conv2d and a layer of layer norm.

[0050] In one example, an image can be segmented by overlapping the image blocks by a certain percentage, such as 10%, with a block size of 512px × 512px. The resulting image blocks can be merged based on the position of each sub-block within the image block. A weighted average is used to obtain the interpretation of the overlapping portions. Let the distance from the pixel to be fused to the edge of the image block be m (in the range [0, 50]), and the number of overlapping pixels on a single edge be L (typically set to 51). The weight alpha is then calculated as (m + 1) / L. Post-processing primarily includes erosion, dilation, and hole filling, which are part of image morphology.

[0051] In one example, when training a hierarchical Transformer-based semantic segmentation model, existing water administration law enforcement background remote sensing monitoring image data (visible light 0.5m resolution) and river monitoring object background results data were annotated to generate training samples. For drone monitoring and interpretation, manually annotated background data can be directly used as training samples to train the semantic segmentation model. For spectral satellite remote sensing monitoring and interpretation, spectral satellite remote sensing images from the same period are downloaded, the background results data are downsampled to the resolution of the spectral satellite imagery, and the interpreted labels of the background results are directly mapped to the spectral satellite imagery as labels for training the semantic segmentation model.

[0052] In one example, the correction unit uses human-computer interaction to visualize the interpretation results in different colors to human experts. The human experts input their correction information based on the visualized interpretation results. The correction unit then corrects the interpretation results of the identified water administrative law enforcement objects based on the correction information obtained from the human experts.

[0053] In one example, the structure of the dynamic tracking module is as follows Figure 5 As shown, it is composed of an object change monitoring unit 1031, a change state analysis unit 1032 and an on-site inspection and tracking unit 1033.

[0054] The object change monitoring unit 1031 is used to compare and analyze the remote sensing images of the same geographical location in two phases. The remote sensing images of the two phases are respectively input into the change detection model based on the dual-branch Segformer and multi-scale attention, and the change detection results output by the change detection model based on the dual-branch Segformer and multi-scale attention are obtained. The recognition module is called to interpret the remote sensing image in the later phase based on the change detection results to obtain the interpretation results of the remote sensing image in the later phase. Among them, the change detection result is a binary image, where a point with a pixel value of 0 indicates no change, and a point with a pixel value of 1 indicates a change. The design of the object change monitoring unit effectively saves computing resources.

[0055] The change state analysis unit 1032 is used to perform a change state analysis on the water administrative law enforcement object based on the interpretation results of the remote sensing images of the previous and next time phases, obtain and record the change state analysis results, and the attribute information of the change state analysis results includes the generation time, change type, change time, creation time and update time.

[0056] The on-site inspection and tracking unit 1033 is used to use satellite positioning technology to capture and record the precise location and inspection path of water administration law enforcement personnel in real time, and provide the suspected image spots, change detection feature maps and change state analysis results of the same water administration law enforcement object before and after the two time phases to the water administration law enforcement personnel, so that the water administration law enforcement personnel can conduct on-site investigations and confirm the change state of the water administration law enforcement object to correct the change state analysis results.

[0057] In an example, the model structure of the change detection model based on dual-branch Segformer and multi-scale attention can be as follows Figure 6As shown in the figure, the change detection model based on dual-branch Segformer and multi-scale attention consists of a dual-branch encoder and a decoder. The encoder consists of two Segformer networks with shared weights and a change information extraction module. Each Segformer network consists of an overlapping block merging layer, four sequentially connected Transformer layers and four CCNet attention modules. For each Segformer network, the input of the first Transformer layer is the output of the overlapping block merging layer, the input of the second Transformer layer is the output of the first Transformer layer, the input of the third Transformer layer is the output of the second Transformer layer, and the input of the fourth Transformer layer is the output of the third Transformer layer. The inputs of the four CCNet attention modules are the outputs of the four Transformer layers respectively. Each Transformer layer contains N (such as N = 2) Transformer blocks, and each Transformer block contains a multi-head self-attention block and a hybrid forward propagation network.

[0058] The remote sensing images of the two phases are respectively input into two Segformer networks based on the dual-branch Segformer and multi-scale attention change detection models.

[0059] The overlapping block merging layer is used to divide its own input into image blocks, retaining only the image blocks of the river range that needs to be monitored, and input the retained image blocks into the first Transformer layer.

[0060] Each Transformer layer is used to extract features from its own input to obtain semantic feature maps of different scales. The semantic feature maps of different scales output by the four Transformer layers can be recorded as and

[0061] The CCNet attention module is used to perform affinity operations and long-range context information aggregation operations on the input semantic feature map to obtain an enhanced semantic feature map. The enhanced semantic feature maps output by the four CCNet attention modules of the same Segformer network are spliced ​​and fused to obtain the final semantic feature map and input it into the change information extraction module.

[0062] by For example, After being input into the corresponding CCNet attention module, the CCNet attention module first Perform two 1×1 convolution operations to obtain the query vector Q and the key vector K respectively. Then, the attention map A is obtained through affinity operation. For each position u in Q, a set Ω can be obtained. u , which comes from the eigenvalues ​​of K corresponding to the vertical and horizontal directions of the position u.

[0063] The affinity operation formula can be expressed as: where d iu It represents the characteristic Q u and By adding a softmax layer on d, we can get the final attention map A.

[0064] Then, in Perform a 1×1 convolution operation on the feature map V, and for each position u in V, you can also get a vector V u and a set Φ u , set Φ u is the set of eigenvectors in V that are located in the same row or column at position u.

[0065] Remote context information is collected by the aggregation operation and can be expressed as: It represents the feature vector of the output feature map H′ at position u, A i,u represents the scalar value at channel i and position u in the attention map A. Contextual information is added to the local feature H to enhance the local features and strengthen the pixel-wise representation, effectively extracting features.

[0066] The enhanced semantic feature maps output by the four CCNet attention modules of the same Segformer network are spliced ​​and fused to obtain the final semantic feature map. The outputs of the two Segformer networks are recorded as and

[0067] Change information extraction module, used to extract the final semantic feature map output by the two Segformer networks and Perform absolute difference operation to extract change feature f cd , and the extracted change features f cd Input to the decoder.

[0068] The decoder is used to transform the input feature f cd Perform binary change detection, pass through a 3×3 convolution layer, and then pass through a 1×1 convolution layer to output the change detection feature map.

[0069] In one example, when iteratively training a change detection model based on a dual-branch Segformer and multi-scale attention, the model training epoch is set to 100 and the batch size is set to 4. During the training process, AdamW is used as the optimizer with a weight decay coefficient of 0.0001. A polynomial learning rate schedule is used, with the initial learning rate for the encoder set to 1.5e-4 and the initial learning rate for the decoder set to 1.5e-3, 10 times that of the encoder.

[0070] In one example, an image can be segmented by overlapping sub-blocks at a certain ratio, such as 10%, with a block size of 512px × 512px. The decoded results of the image blocks are merged based on the position of each sub-block within the image block. A weighted average is used to obtain the decoded result for the overlapping portions. Let the distance from the pixel to be fused to the edge of the image block be m (in the range [0, 50]), and the number of overlapping pixels on a single edge be L (typically set to 51). The weight alpha then equals (m + 1) / L. Post-processing primarily includes erosion, dilation, and hole filling, which are part of image morphology.

[0071] In one example, the generation time is the time when the remote sensing image is generated. The change time is the time when the object type change is confirmed. The creation time is the time when the water administration law enforcement object is interpreted. The update time is the time when the water administration law enforcement object attribute information is filled in, and it is updated in real time as the information is filled in.

[0072] In one example, the change type of the change state analysis result is classified into addition, expansion, reduction, elimination, type conversion, etc., as follows:

[0073] 1) First, check whether the object found this time is within the coverage of the first found object. If not, it is a new type (such as Figure 7 (as shown in the first row), mark it as the first discovered object, enter the first discovered object table, and supplement the object code, name, type, administrative division, regulatory unit, discovery time, area and other information.

[0074] 2) If yes, it is necessary to subdivide and select the most recently discovered object associated with the first discovered object for spatial range comparison and analysis. Figure 7 If the spatial range is larger than the last discovered object, it is marked as an expansion type. At the same time, after supplementing the monitoring image, discovery time, first discovered object number and other information, it is entered into the suspected object interpretation results table.

[0075] 3) If there is a shrinking phenomenon (such as Figure 7 (as shown in the third row), it is marked as a narrowed type, and after supplementing the information, it is entered into the suspected object interpretation results table.

[0076] 4) If there is a type change phenomenon, that is, the object types of the two phases are different, such as from pond aquaculture to photovoltaic power plant (such as Figure 7 (as shown in the 4th row), it is marked as a type change type, and after supplementing the information, it is entered into the suspected object interpretation result table.

[0077] 5) Detection of objects that are not within the above scope and are discovered for the first time and have not been eliminated recently (such as Figure 7 As shown in the 5th row), this time it is marked as an elimination type in the form of a dot. After supplementing the information, it is entered into the suspected object interpretation result table.

[0078] In one example, the on-site inspection tracking module uses satellite positioning technology to capture and record the precise location and inspection path of water administration law enforcement personnel in real time, and displays the interpretation results to be determined on the satellite image, such as Figure 8 As shown, this information is provided to water administration law enforcement personnel for status confirmation. The results of monitoring changes in water administration law enforcement targets are combined with on-site inspection tracking. After extracting the change monitoring results, water administration law enforcement personnel, using a computer or mobile phone, will verify and mark the target change type based on the actual business target and on-site investigation. Ultimately, if a new, expanded, or type-changed enforcement target is confirmed, water administration law enforcement personnel will conduct on-site enforcement using the mobile app until the target is eliminated or the intended goal is determined to have been achieved, marking it with a dot.

[0079] In one example, the extinction module is implemented and used by the superior water administration regulatory department. It is mainly responsible for displaying the rectification status of each water administration law enforcement object, the comparison before and after the rectification, and the historical suspected spots of each water administration law enforcement object. Based on the comparison before and after the rectification, the water administration law enforcement results are fed back to evaluate whether the water administration law enforcement effect has achieved the expected goals. For water administration law enforcement objects that have achieved the expected goals, no longer exist or have disappeared after rectification, they will be accepted and confirmed to be extinct. The extinction time will be recorded in the suspected object interpretation result library. Otherwise, they will be returned to the grassroots water administration law enforcement unit for further rectification. Figure 9 A comparison before and after rectification is shown.

[0080] In one example, the full-cycle tracing module supports first-discovery object retrieval based on the first-discovery object code entered by the user. It then queries the interpretation database for detailed information about the retrieved first-discovery object, tracing the entire lifecycle of the retrieved first-discovery object, including its creation, addition, subtraction, extinction, and reappearance. All collected information is sorted chronologically, cleansed and formatted, and presented to the user via timelines, maps, status change lists, and interactive charts. In other words, it performs four functions: data retrieval and integration, status change sorting, data preprocessing, and multi-format display.

[0081] In one example, a timeline display uses a timeline to show the entire life cycle of a water administration law enforcement object. Key time points are marked on the timeline, including the time of first discovery, change time, extinction time, and recurrence time. Users can view the corresponding detailed information by clicking on key time points on the timeline.

[0082] The map display shows the spatial position changes of water administrative law enforcement objects on the map, uses different colors (such as green for generation, red for disappearance, yellow for reappearance, etc.) or icons to represent the different states of water administrative law enforcement objects, and dynamically updates the display on the map as the time axis moves. For the maximum spatial range, a transparent polygonal area or circular area is marked on the map, and dynamically adjusted as the data is updated.

[0083] The state change list display provides a list view that lists the detailed information of all state change events. Users can view and operate by clicking the list items in the list view, and expand or collapse the detailed information.

[0084] Interactive charts (such as bar charts and line charts) are used to display trends and statistical information about changes in the status of water administration enforcement targets, including the increase and decrease trends of water administration enforcement targets and the distribution of different status types. Users can interact with the mouse (such as hovering and clicking) to view more detailed data point information or filter the chart.

[0085] In this embodiment, a dynamic management system for the full life cycle of spatiotemporal data of water administration law enforcement is constructed and operated, which is composed of an warehousing module, an identification module, a dynamic tracking module, an extinction module and a full-cycle tracing module. This enables full-process, all-round, full-cycle and blind-spot dynamic management of water administration law enforcement objects with multi-temporal, multi-spatial scale, multi-dimensional business attributes and relationships, effectively solving the current problem of insufficient flexibility and pertinence in the traceability of water administration law enforcement objects, facilitating better correlation analysis of data between various time phases, and providing direct evidence support and disposal suggestions for the decision-making of water administration law enforcement departments.

[0086] It is worth mentioning that all modules involved in this embodiment are logical modules. In actual applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. In addition, to highlight the innovation of this application, this embodiment does not include units that are not closely related to solving the technical problem proposed by this application. However, this does not mean that other units do not exist in this embodiment.

[0087] Another embodiment of the present application proposes a method for dynamic management of spatiotemporal data of water administration law enforcement throughout its life cycle, which is applied to electronic devices, wherein the electronic device can be a terminal or a server. In this embodiment and the following embodiments, the electronic device is described using the server as an example. The implementation details of the method for dynamic management of spatiotemporal data of water administration law enforcement throughout its life cycle proposed in this embodiment are described in detail below. The following content is only the implementation details provided for easy understanding and is not necessary for the implementation of this solution.

[0088] The specific process of the dynamic management method for the entire life cycle of spatiotemporal data of water administration law enforcement proposed in this embodiment can be as follows: Figure 10 Shown, including:

[0089] Step 201 : read the input remote sensing image file. The attribute information of the input remote sensing image file needs to include image number, image name, image shooting date, image type, image location description and storage time.

[0090] Step 202 is to interpret the information of the stored remote sensing image files, identify the water administration law enforcement objects within the management scope, form suspected spots of each water administration law enforcement object, and generate water administration law enforcement spatiotemporal data.

[0091] Step 203: Monitor the changes in the water administration law enforcement targets through multi-temporal remote sensing image files and on-site data, and record the spatial nodes and time nodes of the changes.

[0092] Step 204 displays the rectification status of each water administration law enforcement object, and compares each water administration law enforcement object before and after the rectification. When the comparison finds that the water administration law enforcement object does not exist or disappears after the rectification, the disappearance time is recorded.

[0093] Step 205, traces back the entire life cycle of the first discovered water administration law enforcement object, including its generation, increase, decrease, extinction, and reappearance, while updating and recording the maximum target spatial range of the water administration law enforcement object during its existence in real time, and continuously tracks the subsequent water administration law enforcement objects generated again within the said maximum target spatial range.

[0094] The step division of the above various methods is only for the purpose of clear description. During implementation, they can be combined into one step or some steps can be split and decomposed into multiple steps. As long as they include the same logical relationship, they are all within the scope of protection of this application; adding insignificant modifications or introducing insignificant designs to the algorithm or process without changing the core design of the algorithm and process are all within the scope of protection of this application.

[0095] It is not difficult to find that this embodiment is a method embodiment corresponding to the above-mentioned system embodiment, and this embodiment can be implemented in conjunction with the above-mentioned system embodiment. The relevant technical details and technical effects mentioned in the above-mentioned system embodiments are still valid in this method embodiment, and to reduce repetition, they are not repeated here. Accordingly, the relevant technical details mentioned in this method embodiment can also be applied to the above-mentioned system embodiment.

[0096] Another embodiment of the present application proposes a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it can implement a method for dynamic management of the entire life cycle of spatiotemporal data of water administrative law enforcement as described in the above embodiments.

[0097] That is, those skilled in the art will understand that all or part of the steps in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a program, which is stored in a storage medium and includes a number of instructions for causing a device (which may be a single-chip microcomputer, chip, etc.) or a processor to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: a USB flash drive, a mobile hard drive, a ROM (Read-Only Memory), a RAM (Random Access Memory), a magnetic disk, or an optical disk, etc., various media that can store program code.

[0098] Those skilled in the art will appreciate that the above embodiments are specific embodiments for implementing the present application, and that in actual applications, various changes may be made thereto in form and detail without departing from the spirit and scope of the present application.

Claims

1. A dynamic management system for the entire life cycle of spatiotemporal data of water administration law enforcement, characterized by: include: The storage module is used to read the input remote sensing image file. The attribute information of the input remote sensing image file needs to include image number, image name, image shooting date, image type, image location description and storage time; The recognition module is used to interpret the information of the stored remote sensing image files, identify the water administration law enforcement objects within the management scope, form the suspected spots of each water administration law enforcement object, and generate water administration law enforcement spatiotemporal data; The dynamic tracking module is used to monitor the changes in the water administration law enforcement targets through multi-temporal remote sensing image files and field data, and record the spatial and temporal nodes of the changes; The extinction module is used to display the rectification status of each water administration law enforcement object and compare the water administration law enforcement objects before and after the rectification. If the comparison shows that the water administration law enforcement object no longer exists or disappears after the rectification, the extinction time is recorded; The full-cycle tracing module is used to trace the entire life cycle of the first discovered water administration law enforcement object, including its creation, increase, decrease, extinction, and reappearance. It also updates and records the maximum target spatial range of the water administration law enforcement object during its existence in real time, and continuously tracks the subsequent water administration law enforcement objects that appear again within the maximum target spatial range. The recognition module consists of an interpretation unit, a correction unit, a discrimination execution unit, and an interpretation information database. The interpretation information database stores a table of first-discovered objects and a table of suspected object interpretation results. The first-discovered object table is used to record the first-discovered objects and their basic data attributes. The suspected object interpretation results table is used to record the basic data attributes, first-discovery mark attributes, and historical change attributes of each water administration law enforcement object. The interpretation unit is used to interpret the remote sensing image files using a semantic segmentation model based on a hierarchical Transformer to obtain the interpretation results of the identified water administration law enforcement objects. The interpretation results include the type and spatial range of the suspected patches of the identified water administration law enforcement objects; A correction unit is used to obtain correction information from human experts through human-computer interaction, and to correct the interpretation results of the identified water administration law enforcement objects based on the correction information obtained from the human experts; The discrimination execution unit is used to compare the spatial range of the corrected suspected spots with the coverage of each water administration law enforcement object in the first-discovered object table; If the spatial range of the corrected suspected map spot does not overlap with the coverage of each water administration law enforcement object in the first-discovery object table, the identified water administration law enforcement object will be determined as the first-discovery object and newly added to the first-discovery object table and the suspected object interpretation result table; If the spatial range of the corrected suspected map overlaps with the coverage of each water administrative law enforcement object in the first-discovered object table, the overlapping water administrative law enforcement object will be determined as the target water administrative law enforcement object, and the interpretation results of the identified water administrative law enforcement object will be entered into the historical change attributes of the target water administrative law enforcement object in the suspected object interpretation results table.

2. A dynamic management system for the full life cycle of spatiotemporal data of water administration law enforcement according to claim 1, characterized in that: The semantic segmentation model based on the hierarchical Transformer consists of an encoder and a decoder. The encoder adopts a hierarchical Transformer structure, which consists of an image block embedding layer, four sequentially connected Transformer layers, and a feature fusion layer. The input of the first Transformer layer is the output of the image block embedding layer, the input of the second Transformer layer is the output of the first Transformer layer, the input of the third Transformer layer is the output of the second Transformer layer, the input of the fourth Transformer layer is the output of the third Transformer layer, and the input of the feature fusion layer is the output of each Transformer layer. Each Transformer layer contains N Transformer blocks, each of which contains a multi-head self-attention block and a hybrid forward propagation network. The decoder uses a multi-layer perceptron. The image block embedding layer is used to divide the input remote sensing image file into blocks, retaining only the image blocks of the river range that needs to be monitored, and input the retained image blocks into the first layer of Transformer layer; Each Transformer layer is used to extract features from its own input to obtain feature maps of different scales; The feature fusion layer is used to fuse the feature maps output by each Transformer layer. The feature maps whose resolution needs to be reduced are downsampled through 3×3 convolution operations, and the feature maps whose resolution needs to be increased are upsampled using DUpsampling to obtain fused feature maps. The multi-layer perceptron is used to adjust the channels of the fused feature map to obtain the semantic segmentation results of the input remote sensing image file, that is, the interpretation results of the identified water administrative law enforcement objects.

3. A dynamic management system for the full life cycle of spatiotemporal data of water administration law enforcement according to claim 1, characterized in that: The dynamic tracking module consists of an object change monitoring unit, a change state analysis unit, and an on-site inspection and tracking unit; The object change monitoring unit is used to compare and analyze the remote sensing images of the same geographical location in two phases. The remote sensing images of the two phases are respectively input into the change detection model based on the dual-branch Segformer and multi-scale attention, and the change detection results output by the change detection model based on the dual-branch Segformer and multi-scale attention are obtained. The recognition module is called to interpret the remote sensing image with the later phase based on the change detection results to obtain the interpretation results of the remote sensing image with the later phase; A change state analysis unit is used to perform change state analysis on water administration law enforcement objects based on the interpretation results of remote sensing images of two phases, obtain and record the change state analysis results, and the attribute information of the change state analysis results includes generation time, change type, change time, creation time and update time; The on-site inspection and tracking unit is used to use satellite positioning technology to capture and record the precise location and inspection route of water administration law enforcement personnel in real time, and provide water administration law enforcement personnel with suspected patches, change detection feature maps and change status analysis results of the same water administration law enforcement object in the two time phases before and after, so that water administration law enforcement personnel can conduct on-site investigations and confirm the change status of the water administration law enforcement object to correct the change status analysis results.

4. A dynamic management system for the full life cycle of spatiotemporal data of water administration law enforcement according to claim 3, characterized in that: The change detection model based on dual-branch Segformer and multi-scale attention consists of an encoder and a decoder. The encoder consists of two Segformer networks with shared weights and a change information extraction module. Each Segformer network consists of an overlapping block merging layer, four sequentially connected Transformer layers and four CCNet attention modules. For each Segformer network, the input of the first Transformer layer is the output of the overlapping block merging layer, the input of the second Transformer layer is the output of the first Transformer layer, the input of the third Transformer layer is the output of the second Transformer layer, and the input of the fourth Transformer layer is the output of the third Transformer layer. The inputs of the four CCNet attention modules are the outputs of the four Transformer layers respectively. Each Transformer layer contains N Transformer blocks, and each Transformer block contains a multi-head self-attention block and a hybrid forward propagation network. The remote sensing images of the two phases are respectively input into two Segformer networks based on the dual-branch Segformer and multi-scale attention change detection models; The overlapping block merging layer is used to divide its own input image into blocks, retaining only the image blocks of the river range that needs to be monitored, and input the retained image blocks into the first layer of Transformer layer; Each Transformer layer is used to extract features from its own input to obtain semantic feature maps of different scales; The CCNet attention module is used to perform affinity operations and long-range context information aggregation operations on the input semantic feature map to obtain an enhanced semantic feature map. The enhanced semantic feature maps output by the four CCNet attention modules of the same Segformer network are spliced ​​and fused to obtain the final semantic feature map and input it into the change information extraction module; The change information extraction module is used to perform absolute difference calculation on the final semantic feature maps output by the two Segformer networks to extract change features, and input the extracted change features into the decoder; The decoder is used to perform binary change detection on the input change features. After passing through a 3×3 convolution layer and then a 1×1 convolution layer, the decoder outputs the change detection result.

5. A dynamic management system for the full life cycle of spatiotemporal data of water administration law enforcement according to claim 1, characterized in that: The extinction module is implemented and used by the superior water administration supervision department. It is mainly responsible for displaying the rectification status of each water administration law enforcement object, the comparison before and after the rectification, and the historical suspected spots of each water administration law enforcement object. It feedbacks the water administration law enforcement results based on the comparison before and after the rectification, and evaluates whether the water administration law enforcement effect has achieved the expected goals. For water administration law enforcement objects that have achieved the expected goals, no longer exist or have disappeared after rectification, they will be accepted and confirmed to be extinct. The extinction time will be recorded in the suspected object interpretation result database. Otherwise, they will be returned to the grassroots water administration law enforcement unit for further rectification.

6. A dynamic management system for the full life cycle of spatiotemporal data of water administration law enforcement according to claim 1, characterized in that: The full-cycle tracing module supports the retrieval of first-discovered objects based on the first-discovered object codes input by the user, queries the detailed information of the retrieved first-discovered objects in the interpretation information library, and traces the entire life cycle of the retrieved first-discovered objects, including their generation, increase, decrease, extinction, and reappearance. All collected information is sorted in chronological order, and after cleaning and formatting, it is presented to users through timelines, maps, status change lists, and interactive charts.

7. A dynamic management system for the full life cycle of spatiotemporal data of water administration law enforcement according to claim 6, characterized in that: Timeline display uses a timeline to display the entire life cycle of water administration law enforcement objects. The timeline marks each key time point, including the first discovery time, change time, extinction time, and reappearance time. Users can view the corresponding detailed information by clicking on the key time point on the timeline; Map display: By showing the spatial position changes of water administration law enforcement objects on the map, different colors or icons are used to indicate the different states of water administration law enforcement objects, and the display on the map is dynamically updated as the time axis moves. For the maximum spatial range, a transparent polygonal area or circular area is used to mark it on the map, and it is dynamically adjusted as the data is updated; The state change list display provides a list view that lists the detailed information of all state change events, allowing users to view and operate by clicking on the list items in the list view; Interactive chart display uses interactive charts to display the trend and statistical information of the status changes of water administrative law enforcement objects, including the increase and decrease trends of water administrative law enforcement objects and the distribution of different status types.

8. A method for dynamic management of spatiotemporal data of water administration law enforcement throughout its life cycle, based on a dynamic management system for spatiotemporal data of water administration law enforcement throughout its life cycle according to any one of claims 1 to 7, characterized in that: include: Read the input remote sensing image file. The attribute information of the input remote sensing image file needs to include image number, image name, image shooting date, image type, image location description and storage time; Interpret the information of the stored remote sensing image files, identify the water administration law enforcement objects within the management scope, form the suspected spots of each water administration law enforcement object, and generate water administration law enforcement spatiotemporal data; Through multi-temporal remote sensing image files and on-site data, monitor the changes in the growth and decline of various water administration law enforcement targets and record the spatial and temporal nodes of the growth and decline; Display the rectification status of each water administration law enforcement target, and compare each water administration law enforcement target before and after the rectification. If the comparison shows that the water administration law enforcement target no longer exists or disappears after the rectification, record the time of disappearance; Trace the entire life cycle of the first discovered water administration law enforcement object, including its generation, increase, decrease, extinction, and reappearance, while updating and recording the maximum target spatial range of the water administration law enforcement object during its existence in real time, and continue to track the subsequent water administration law enforcement objects that are generated again within the said maximum target spatial range.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by the processor, it can implement a method for dynamic management of spatiotemporal data of water administrative law enforcement throughout its life cycle as described in claim 8.

Images