A method and system for assessing carbon emissions from land use change

Abstract

This invention relates to the field of carbon emission technology, specifically to a method and system for assessing carbon emissions from land use change. The method and system include: a data preprocessing module, a carbon emission assessment module, and a carbon emission assessment correction module. This invention aligns remote sensing images of the actual area with planning and design maps, and uses a land type labeling network to label the rasterized images of the actual area with land types, constructing a land type change map. Further, it simulates and analyzes the carbon emission changes corresponding to land type transitions in historical records, and inputs this data into a carbon emission change distribution analysis network to generate a carbon emission change distribution map. By correcting the carbon emission change distribution map based on the spatial distribution relationships of land type change clusters, and using convolution operations to optimize the carbon emission change distribution map, the corrected image can more accurately reflect the carbon emission changes in the planning and design area.

Description

Technical Field

[0001] This invention relates to the field of carbon emission technology, and more specifically to a method and system for assessing carbon emissions from land use change. Background Technology

[0002] When conducting regional development planning, it is usually necessary to consider the environmental impact of carbon emissions caused by land use change. However, the current carbon emission assessment for planning and design relies heavily on the professional experience of professionals and a large amount of analysis work, which makes the carbon emission assessment process cumbersome and inefficient, and lacks intuitive assessment tools. Summary of the Invention

[0003] This invention aligns remote sensing images of the actual area with planning and design maps, and uses a land type labeling network to label the rasterized images of the actual area with land types, constructing a land type change map. It further simulates and analyzes the carbon emission changes corresponding to land type transitions in historical records, and inputs this data into a carbon emission change distribution analysis network to generate a carbon emission change distribution map. By correcting the carbon emission change distribution map based on the spatial distribution relationships within the land type transition clusters, and using convolution operations to optimize the carbon emission change distribution map, this invention ensures that the corrected image more accurately reflects the carbon emission changes in the planning and design area, providing operators with a more precise assessment of carbon emission impact.

[0004] This invention provides a method for assessing carbon emissions from land use change, comprising:

[0005] Acquire remote sensing images of the area to be analyzed, acquire the corresponding planning and design map of the area to be analyzed, and adjust the remote sensing images of the area to be analyzed to match the planning and design map. Figure 1 After the size is determined, it is rasterized simultaneously with the planning and design map to construct the corresponding actual area rasterized image and planning and design rasterized image. Each cell in the planning and design rasterized image is labeled with land type. The actual area rasterized image is sent to the land type labeling network for processing, and the land type is labeled for each cell in the actual area rasterized image.

[0006] A land type change map is constructed based on rasterized images of the actual area and rasterized images of the planning and design. Each cell in the land type change map stores the pixel value corresponding to the land type transformation from the corresponding cell in the actual area rasterized image to the corresponding cell in the planning and design rasterized image. The land type change map is then fed into a carbon emission change distribution analysis network for processing to obtain a carbon emission change distribution map. The size of the carbon emission change distribution map is consistent with that of the actual area rasterized image, and each cell in the carbon emission change distribution map stores the simulated carbon emission change corresponding to the land type transformation from the corresponding cell in the actual area rasterized image to the corresponding cell in the planning and design rasterized image.

[0007] Based on the land type change map, several land type change clusters are identified. Then, the carbon emission change distribution map is adjusted based on the spatial distribution relationship of all land type change clusters to construct a revised carbon emission change distribution map. Finally, using the identity matrix as the convolution kernel, a convolution operation is performed on the revised carbon emission change distribution map to obtain the carbon emission change amount.

[0008] As a preferred approach, the carbon emission change distribution map is adjusted based on spatial distribution relationships across all land type transformation clusters, specifically including the following steps:

[0009] Traverse all pixels in the carbon emission change distribution map. For each selected pixel, perform the following operations: obtain all pixels within the 8-neighborhood of the selected pixel and record them as reference pixels; use the proportion of reference pixels belonging to the land type change cluster area as the cluster weight; use the proportion of reference pixels not belonging to the land type change cluster area as the dispersion weight; associate the reference pixels belonging to the land type change cluster area with the cluster weight; associate the reference pixels not belonging to the land type change cluster area with the dispersion weight; and replace the simulated carbon emission change corresponding to the selected pixel with the output of the weighted average operation of the simulated carbon emission change corresponding to the selected pixel to construct a smooth carbon emission change distribution map.

[0010] Construct a spatial distribution map, the size of which is related to the distribution of carbon emission changes. Figure 1 In this process, the data corresponding to pixels belonging to the land type change cluster area in the spatial distribution relationship map is set to 1, and the data corresponding to pixels not belonging to the land type change cluster area in the spatial distribution relationship map is set to 0. Then, all pixels in the spatial distribution relationship map are traversed, and the following operations are performed on each selected pixel: All pixels within the 8-neighborhood of the selected pixel are obtained and recorded as reference relationship pixels (including the selected pixel itself). The proportion of reference relationship pixels belonging to the land type change cluster area is used as the clustering relationship weight, and the proportion of reference relationship pixels not belonging to the land type change cluster area is used as the dispersion relationship weight. Reference relationship pixels belonging to the land type change cluster area are associated with the clustering relationship weight, and reference relationship pixels not belonging to the land type change cluster area are associated with the dispersion relationship weight. The output of the weighted average operation performed on the data corresponding to all reference relationship pixels according to their respective weights is used to replace the data corresponding to the selected pixel, thus constructing a smooth spatial distribution relationship map.

[0011] The Hadamard product is calculated by smoothing the carbon emission change distribution map and the smooth spatial distribution relationship map to adjust the carbon emission change distribution map.

[0012] As a preferred approach, several land type change clusters are identified based on land type change maps, specifically including the following steps:

[0013] Perform an 8-neighborhood connectivity operation on the land type change map to construct several connectivity regions, each of which is a cluster of land type changes;

[0014] The connected component operation is as follows: For any cell in the land type change map, denoted as the target cell, select all cells within the 8-neighborhood of the target cell. If the pixel value of the selected cell is the same as the pixel value of the target cell, add the selected cell to the corresponding connected component. Each newly added cell is used as the target cell for the next round of query. This process continues until no new cells are added to the connected component, thus completing the construction of a connected component.

[0015] As a preferred approach, training the land type labeling network involves the following steps:

[0016] Obtain land type labeling training samples, which include rasterized images. Label all pixels in the rasterized images according to land type. Combine all labeled land type labeling training samples into a land type labeling training set. Train the land type labeling network using the land type labeling training set. Determine if the accuracy of the land type labeling network is higher than a set threshold. If the accuracy of the land type labeling network is higher than the set threshold, output the trained land type labeling network; otherwise, continue training the land type labeling network using the land type labeling training set.

[0017] As a preferred approach, training a carbon emission change distribution analysis network involves the following steps:

[0018] Acquire training samples for carbon emission change distribution analysis, including land type change maps. Label the training samples using these maps. Combine all labeled training samples into a training set. Train the carbon emission change distribution analysis network using this training set. Determine if the network's accuracy exceeds a set threshold. If it does, output the trained network; otherwise, continue training the network using the training set.

[0019] As a preferred aspect, both the land type labeling network and the carbon emission change distribution analysis network are based on the U-net model.

[0020] The present invention also provides a carbon emission assessment system for land use change, comprising:

[0021] The data preprocessing module is used to acquire remote sensing images of the area to be analyzed, acquire the corresponding planning and design map of the area to be analyzed, and adjust the remote sensing images of the area to be analyzed to match the planning and design map. Figure 1 After the size is determined, it is rasterized simultaneously with the planning and design map to construct the corresponding actual area rasterized image and planning and design rasterized image. Each cell in the planning and design rasterized image is labeled with land type. The actual area rasterized image is sent to the land type labeling network for processing, and the land type is labeled for each cell in the actual area rasterized image.

[0022] The carbon emission assessment module is used to construct a land type change map based on actual regional rasterized images and planning design rasterized images. Each cell in the land type change map stores the pixel value corresponding to the land type transformation from the corresponding cell in the actual regional rasterized image to the corresponding cell in the planning design rasterized image. The land type change map is then fed into the carbon emission change distribution analysis network for processing to obtain a carbon emission change distribution map. The size of the carbon emission change distribution map is consistent with that of the actual regional rasterized image, and each cell in the carbon emission change distribution map stores the simulated carbon emission change corresponding to the land type transformation from the corresponding cell in the actual regional rasterized image to the corresponding cell in the planning design rasterized image.

[0023] The carbon emission assessment correction module is used to identify several land type change clusters based on the land type change map, and adjust the carbon emission change distribution map based on the spatial distribution relationship of all land type change clusters to construct a corrected carbon emission change distribution map. Then, using the identity matrix as the convolution kernel, a convolution operation is performed on the corrected carbon emission change distribution map to obtain the carbon emission change amount.

[0024] The present invention has the following advantages:

[0025] This invention aligns remote sensing images of the actual area with planning and design maps, and uses a land type labeling network to label the rasterized images of the actual area with land types, constructing a land type change map. It further simulates and analyzes the carbon emission changes corresponding to land type transitions in historical records, and inputs this data into a carbon emission change distribution analysis network to generate a carbon emission change distribution map. By correcting the carbon emission change distribution map based on the spatial distribution relationships within the land type transition clusters, and using convolution operations to optimize the carbon emission change distribution map, this invention ensures that the corrected image more accurately reflects the carbon emission changes in the planning and design area, providing operators with a more precise assessment of carbon emission impact. Attached Figure Description

[0026] Figure 1 This is a schematic diagram of the carbon emission assessment system for land use change used in an embodiment of the present invention. Detailed Implementation

[0027] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of this invention.

[0028] Example 1: A method for assessing carbon emissions from land use change, comprising:

[0029] Acquire remote sensing images of the area to be analyzed. This area refers to the region planned and designed by the operators. Since this area will undergo construction planning in the future, it is necessary to assess the carbon emission changes of the planned area to guide the planning and design. Obtain the corresponding planning and design map of the area to be analyzed. This planning and design map refers to the image constructed by the operators after planning and designing the area to be analyzed. Adjust the remote sensing image of the area to be analyzed to match the planning and design. Figure 1 After the actual area is sized, it is simultaneously rasterized with the planning and design map to construct corresponding rasterized images of the actual area and the planning and design map. Each cell in the planning and design rasterized image is labeled with a land type, which includes forest, agricultural land, urban land, grassland, and wetland, etc. During the rasterization process, it is generally necessary to ensure that each grid corresponds to no more than one land type. The actual area rasterized image is then fed into a land type labeling network for processing, and each cell in the actual area rasterized image is labeled with a land type. The land type labeling network is built based on the U-net model.

[0030] A land type change map is constructed based on rasterized images of the actual area and rasterized images of the planning and design. Each pixel in the land type change map stores the pixel value corresponding to the land type transformation from the corresponding pixel in the actual area rasterized image to the corresponding pixel in the planning and design rasterized image. Here, land type transformation refers to forest-agricultural land, agricultural land-urbanized land, etc. The pixel value corresponding to each type of land type transformation is set in advance by the operator, and it is necessary to ensure that the pixel values ​​corresponding to different land type transformations can be clearly distinguished. The land type change map is then fed into a carbon emission change distribution analysis network for processing to obtain a carbon emission change distribution map. The size of the carbon emission change distribution map is the same as that of the actual area rasterized image. The carbon emission change distribution map is consistent with the actual area, and each cell stores the simulated carbon emission change corresponding to the land type transformation from the corresponding cell in the actual area's rasterized image to the corresponding cell in the planning and design rasterized image. The simulated carbon emission change here comes from learning the carbon emission change corresponding to the land type transformation in historical records. It can reflect the change in carbon emissions after the current land type of the area under analysis changes to the land type corresponding to the planning and design map. For example, the conversion of agricultural land to urban land will increase carbon emissions from transportation, construction, and energy consumption. The historical records here come from carbon emission monitoring in different areas. The carbon emission change distribution analysis network is built based on the U-net model.

[0031] Based on the land type change map, several land type transformation clusters are identified. Then, the carbon emission change distribution map is adjusted based on the spatial distribution relationships of all these clusters, constructing a revised carbon emission change distribution map. A convolution operation is then performed on the revised carbon emission change distribution map using an identity matrix as the convolution kernel to obtain the carbon emission change. This carbon emission change can provide operators with an environmental impact reference for planning and design in the area under analysis. It should be noted that the carbon emission change distribution map output by the carbon emission change distribution analysis network is affected by regional differences, leading to numerical fluctuations in the simulated carbon emission change and unreasonable continuity in spatial distribution. Therefore, the carbon emission change map needs adjustment to more accurately reflect the carbon emission situation, thereby enabling operators to more accurately assess the carbon emission situation in planning and design.

[0032] This application aligns remote sensing images of the actual area with planning and design maps, and uses a land type labeling network to label the rasterized images of the actual area with land types, constructing a land type change map. It further simulates and analyzes the carbon emission changes corresponding to land type transitions in historical records, and inputs this data into a carbon emission change distribution analysis network to generate a carbon emission change distribution map. By modifying the carbon emission change distribution map based on the spatial distribution relationships within the land type transition clusters, and using convolution operations to optimize the carbon emission change distribution map, the modified image more accurately reflects the carbon emission changes in the planning and design area, providing operators with a more precise assessment of carbon emission impacts.

[0033] Based on all areas of land type transformation clusters, the carbon emission change distribution map was adjusted in conjunction with spatial distribution relationships. The specific steps included are as follows:

[0034] The algorithm iterates through all pixels in the carbon emission change distribution map. For each selected pixel, the following operations are performed: All pixels within the 8-neighborhood of the selected pixel are obtained and designated as reference pixels (including the selected pixel itself). The proportion of reference pixels belonging to land type change clusters is used as the clustering weight, and the proportion of reference pixels not belonging to land type change clusters is used as the dispersion weight. This effectively distinguishes the differences in carbon emission impact between different land type change areas and surrounding areas, helping to better balance carbon emission changes in different regions. Reference pixels belonging to land type change clusters are associated with the clustering weight, and reference pixels not belonging to land type change clusters are associated with the dispersion weight. The simulated carbon emission changes corresponding to all reference pixels are weighted and averaged according to their respective weights, and the output of this weighted average is used to replace the simulated carbon emission changes corresponding to the selected pixel, constructing a smooth carbon emission change distribution map. The weighted average operation effectively corrects anomalies in carbon emission changes caused by regional differences, making the carbon emission change distribution map more continuous and smooth.

[0035] Construct a spatial distribution map, the size of which is related to the distribution of carbon emission changes. Figure 1In this process, the data corresponding to pixels belonging to the land type change cluster area in the spatial distribution relationship map is set to 1, and the data corresponding to pixels not belonging to the land type change cluster area in the spatial distribution relationship map is set to 0. Then, all pixels in the spatial distribution relationship map are traversed, and the following operations are performed on each selected pixel: All pixels within the 8-neighborhood of the selected pixel are obtained and recorded as reference relationship pixels (including the selected pixel itself). The proportion of reference relationship pixels belonging to the land type change cluster area is used as the clustering relationship weight, and the proportion of reference relationship pixels not belonging to the land type change cluster area is used as the dispersion relationship weight. Reference relationship pixels belonging to the land type change cluster area are associated with the clustering relationship weight, and reference relationship pixels not belonging to the land type change cluster area are associated with the dispersion relationship weight. The output of the weighted average operation performed on the data corresponding to all reference relationship pixels according to their respective weights is used to replace the data corresponding to the selected pixel, thus constructing a smooth spatial distribution relationship map.

[0036] The smoothed carbon emission change distribution map and the smoothed spatial distribution relationship map are subjected to Hadamard product calculation to adjust the carbon emission change distribution map. The Hadamard product operation ensures the spatial consistency between the smoothed carbon emission change distribution map and the spatial distribution relationship, making the spatial distribution of carbon emissions more consistent with the planning and design map.

[0037] Based on the land type change map, several land type transformation clusters were identified, specifically including the following steps:

[0038] Perform an 8-neighborhood connectivity operation on the land type change map to construct several connectivity regions, each of which is a cluster of land type changes;

[0039] The connected component operation is as follows: For any cell in the land type change map, denoted as the target cell, select all cells within the 8-neighborhood of the target cell. If the pixel value of the selected cell is the same as the pixel value of the target cell, add the selected cell to the corresponding connected component. Each newly added cell is used as the target cell for the next round of query. This process continues until no new cells are added to the connected component, thus completing the construction of a connected component.

[0040] Training the land type labeling network involves the following steps:

[0041] Obtain land type labeling training samples, which include rasterized images. These rasterized images are constructed by operators based on historical remote sensing images. All pixels in the rasterized images are labeled according to land type. This labeling work is generally done manually by operators. All labeled land type labeling training samples are combined into a land type labeling training set. The land type labeling network is trained using the land type labeling training set. It is then determined whether the accuracy of the land type labeling network is higher than a set threshold. If the accuracy of the land type labeling network is higher than the set threshold, the trained land type labeling network is output. Otherwise, the land type labeling network is trained again using the land type labeling training set.

[0042] Training a network for analyzing carbon emission change distribution involves the following steps:

[0043] Training samples for carbon emission change distribution analysis are obtained, including land type change maps. These land type change maps are derived from remote sensing images of the same area at adjacent detection time points in historical records. The training samples are labeled using these carbon emission change distribution maps, which are constructed by operators based on actual monitoring values ​​from historical records. All labeled training samples are combined into a training set for carbon emission change distribution analysis. The carbon emission change distribution analysis network is trained using this training set. The accuracy of the network is then assessed to determine if it exceeds a set threshold. If the accuracy exceeds the threshold, the trained network is output; otherwise, the network continues to be trained using the training set.

[0044] Example 2, a carbon emission assessment system for land use change, see [link to example]. Figure 1 ,include:

[0045] The data preprocessing module acquires remote sensing images of the area to be analyzed. This area refers to the region planned and designed by the operators, which will undergo construction planning in the future. Therefore, it is necessary to assess the carbon emission changes of the planned area to guide the planning and design. The module also acquires the corresponding planning and design map of the area to be analyzed. This planning and design map refers to the image constructed by the operators after planning and designing the area. The remote sensing image of the area to be analyzed is then adjusted to match the planning and design. Figure 1After the actual area is sized, it is simultaneously rasterized with the planning and design map to construct corresponding rasterized images of the actual area and the planning and design map. Each cell in the planning and design rasterized image is labeled with a land type, which includes forest, agricultural land, urban land, grassland, and wetland, etc. During the rasterization process, it is generally necessary to ensure that each grid corresponds to no more than one land type. The actual area rasterized image is then fed into a land type labeling network for processing, and each cell in the actual area rasterized image is labeled with a land type. The land type labeling network is built based on the U-net model.

[0046] The carbon emission assessment module is used to construct a land type change map based on rasterized images of the actual area and rasterized images of the planning and design. Each pixel in the land type change map stores the pixel value corresponding to the land type transformation from the corresponding pixel in the actual area rasterized image to the corresponding pixel in the planning and design rasterized image. Here, land type transformation refers to forest-agricultural land, agricultural land-urbanized land, etc. The pixel value corresponding to each land type transformation is preset by the operator, and it must be able to clearly distinguish the pixel values ​​corresponding to different land type transformations. The land type change map is then fed into the carbon emission change distribution analysis network for processing to obtain the carbon emission change distribution map. The size of the carbon emission change distribution map is consistent with the actual area rasterized image, and each cell in the carbon emission change distribution map stores the simulated carbon emission change corresponding to the land type transformation from the corresponding cell in the actual area rasterized image to the corresponding cell in the planning and design rasterized image. The simulated carbon emission change here comes from learning the carbon emission change corresponding to the land type transformation in the historical records. It can reflect the change in carbon emissions after the current land type of the area to be analyzed changes to the land type corresponding to the planning and design map. For example, the conversion of agricultural land to urban land will increase carbon emissions from transportation, construction and energy consumption. The historical records here come from carbon emission monitoring in different areas.

[0047] The carbon emission assessment correction module is used to identify several land type change clusters based on the land type change map. It then adjusts the carbon emission change distribution map based on the spatial distribution relationships of all these clusters, constructing a corrected carbon emission change distribution map. Finally, using an identity matrix as the convolution kernel, a convolution operation is performed on the corrected carbon emission change distribution map to obtain the carbon emission change. This carbon emission change can provide operators with an environmental impact reference for the planning and design of the area under analysis. It should be noted that the carbon emission change distribution map output by the carbon emission change distribution analysis network is affected by regional differences, leading to numerical fluctuations in the simulated carbon emission change and unreasonable continuity in spatial distribution. Therefore, the carbon emission change map needs to be adjusted so that the corrected map more accurately reflects the carbon emission situation, thereby enabling operators to more accurately assess the carbon emission situation in planning and design.

[0048] It should be understood that those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims. Parts not described in detail in this specification are prior art known to those skilled in the art.

Claims

1. A method for assessing carbon emissions from land use change, characterized in that, include: Acquire remote sensing images of the area to be analyzed, acquire the corresponding planning and design map of the area to be analyzed, adjust the remote sensing image of the area to be analyzed to the same size as the planning and design map, and rasterize both the remote sensing image of the area to be analyzed and the planning and design map at the same time, construct the corresponding rasterized image of the actual area and the rasterized image of the planning and design map, each cell in the planning and design rasterized image is labeled by land type, and the rasterized image of the actual area is sent to the land type labeling network for processing, and the land type is labeled for each cell in the rasterized image of the actual area; A land type change map is constructed based on rasterized images of the actual area and rasterized images of the planning and design. Each cell in the land type change map stores the pixel value corresponding to the land type transformation from the corresponding cell in the actual area rasterized image to the corresponding cell in the planning and design rasterized image. The land type change map is then fed into a carbon emission change distribution analysis network for processing to obtain a carbon emission change distribution map. The size of the carbon emission change distribution map is consistent with that of the actual area rasterized image, and each cell in the carbon emission change distribution map stores the simulated carbon emission change corresponding to the land type transformation from the corresponding cell in the actual area rasterized image to the corresponding cell in the planning and design rasterized image. Based on the land type change map, several land type change clusters are identified. Then, the carbon emission change distribution map is adjusted based on all land type change clusters and spatial distribution relationships to construct a revised carbon emission change distribution map. Finally, using the identity matrix as the convolution kernel, a convolution operation is performed on the revised carbon emission change map to obtain the carbon emission change amount. Based on all areas of land type transformation clusters, the carbon emission change distribution map was adjusted in conjunction with spatial distribution relationships. The specific steps included are as follows: Traverse all pixels in the carbon emission change distribution map. For each selected pixel, perform the following operations: obtain all pixels within the 8-neighborhood of the selected pixel and record them as reference pixels; use the proportion of reference pixels belonging to the land type change cluster area as the cluster weight; use the proportion of reference pixels not belonging to the land type change cluster area as the dispersion weight; associate the reference pixels belonging to the land type change cluster area with the cluster weight; associate the reference pixels not belonging to the land type change cluster area with the dispersion weight; and replace the simulated carbon emission change corresponding to the selected pixel with the output of the weighted average operation of the simulated carbon emission change corresponding to the selected pixel to construct a smooth carbon emission change distribution map. Construct a spatial distribution map with the same size as the carbon emission change distribution map. In the spatial distribution map, the data corresponding to pixels belonging to land type change clusters is set to 1, and the data corresponding to pixels not belonging to land type change clusters is set to 0. Iterate through all pixels in the spatial distribution map. For each selected pixel, perform the following operations: obtain all pixels within the 8-neighborhood of the selected pixel and record them as reference relationship pixels. Use the proportion of reference relationship pixels belonging to land type change clusters as the clustering relationship weight, and the proportion of reference relationship pixels not belonging to land type change clusters as the dispersion relationship weight. Associate the reference relationship pixels belonging to land type change clusters with the clustering relationship weight, and associate the reference relationship pixels not belonging to land type change clusters with the dispersion relationship weight. Replace the data corresponding to the selected pixel with the output of the weighted average operation performed on the data corresponding to all reference relationship pixels according to their respective weights, thus constructing a smooth spatial distribution map. The Hadamard product is calculated by smoothing the carbon emission change distribution map and the smooth spatial distribution relationship map to adjust the carbon emission change distribution map.

2. The method for assessing carbon emissions from land use change according to claim 1, characterized in that, Based on the land type change map, several land type transformation clusters were identified, specifically including the following steps: Perform an 8-neighborhood connectivity operation on the land type change map to construct several connectivity regions, each of which is a cluster of land type changes; The connected component operation is as follows: For any cell in the land type change map, which is denoted as the target cell, select all cells in the 8-neighborhood of the target cell. If the pixel value of the selected cell is the same as the pixel value of the target cell, then add the selected cell to the corresponding connected component, and each newly added cell is used as the target cell for the next round of query. The construction of a connected component is completed when no new cells are added to it.

3. The method for assessing carbon emissions from land use change according to claim 2, characterized in that, Training the land type labeling network involves the following steps: Obtain land type labeling training samples, which include rasterized images. Label all pixels in the rasterized images according to land type. Combine all labeled land type labeling training samples into a land type labeling training set. Train the land type labeling network using the land type labeling training set. Determine if the accuracy of the land type labeling network is higher than a set threshold. If the accuracy of the land type labeling network is higher than the set threshold, output the trained land type labeling network; otherwise, continue training the land type labeling network using the land type labeling training set.

4. The method for assessing carbon emissions from land use change according to claim 3, characterized in that, Training a network for analyzing carbon emission change distribution involves the following steps: Acquire training samples for carbon emission change distribution analysis, including land type change maps. Label the training samples using these maps. Combine all labeled training samples into a training set. Train the carbon emission change distribution analysis network using this training set. Determine if the network's accuracy exceeds a set threshold. If it does, output the trained network; otherwise, continue training the network using the training set.

5. The method for assessing carbon emissions from land use change according to claim 4, characterized in that, The land type labeling network and the carbon emission change distribution analysis network are both based on the U-net model.

6. A carbon emission assessment system for land use change, characterized in that, The system applies a carbon emission assessment method for land use change as described in any one of claims 1-5, comprising: The data preprocessing module is used to acquire remote sensing images of the area to be analyzed, acquire the corresponding planning and design map of the area to be analyzed, adjust the remote sensing image of the area to be analyzed to the same size as the planning and design map, and rasterize both the remote sensing image of the area to be analyzed and the planning and design map at the same time to construct the corresponding rasterized image of the actual area and the rasterized image of the planning and design map. Each cell in the planning and design rasterized image is labeled by land type. The rasterized image of the actual area is sent to the land type labeling network for processing, and the land type is labeled for each cell in the actual area rasterized image. The carbon emission assessment module is used to construct a land type change map based on actual regional rasterized images and planning design rasterized images. Each cell in the land type change map stores the pixel value corresponding to the land type transformation from the corresponding cell in the actual regional rasterized image to the corresponding cell in the planning design rasterized image. The land type change map is then fed into the carbon emission change distribution analysis network for processing to obtain a carbon emission change distribution map. The size of the carbon emission change distribution map is consistent with that of the actual regional rasterized image, and each cell in the carbon emission change distribution map stores the simulated carbon emission change corresponding to the land type transformation from the corresponding cell in the actual regional rasterized image to the corresponding cell in the planning design rasterized image. The carbon emission assessment correction module is used to identify several land type change clusters based on the land type change map, and adjust the carbon emission change distribution map based on the spatial distribution relationship of all land type change clusters to construct a corrected carbon emission change distribution map. Then, using the identity matrix as the convolution kernel, a convolution operation is performed on the corrected carbon emission change distribution map to obtain the carbon emission change amount.

Images