A method for extracting construction land use range

Abstract

The present invention discloses a method for extracting the scope of construction land use, comprising: extracting project data in a project that has been started, the project data including project image data, project boundary data and construction boundary data; performing multi-scale conversion on the project boundary data and the construction boundary data to obtain boundary perception features; inputting the image data into a feature extraction network, and encoding the image data into multi-scale features using the feature extraction network; fusing the boundary perception features and the multi-scale features to obtain comprehensive features; and combining the contextual semantic information of the comprehensive features to determine the scope of construction land use in the project that has been started. The present invention determines the scope of construction land use in the project that has been started by combining the comprehensive features with the contextual semantic information, without the need for manual on-site survey, which not only reduces human subjective judgment and operation errors, but also can effectively improve the efficiency of extracting the scope of construction land use.

Description

Technical Field

[0001] The present invention relates to the technical field of construction management, and in particular to a method for extracting a construction land utilization range. Background Art

[0002] Determining the scope of construction land use is an important part of engineering project management, which involves many aspects such as project planning, design, budget and construction.

[0003] The existing methods for extracting the scope of construction land use are usually completed through manual on-site surveys, which requires a lot of manpower, material and financial resources, and the accuracy of the determined construction land use scope is low. Summary of the invention

[0004] The present invention provides a method for extracting the scope of construction land use, so as to solve the technical problem that the existing method for extracting the scope of construction land use is usually completed by manual on-site investigation, which requires a lot of manpower, material resources and financial resources, and the accuracy of the determined construction land use scope is low.

[0005] The present invention provides a method for extracting a construction land use range, comprising:

[0006] Extracting project data from a construction site project that has already started, the project data including project image data, project boundary data, and construction boundary data;

[0007] Performing multi-scale transformation on the project boundary data and the construction boundary data to obtain boundary perception features;

[0008] Inputting the image data into a feature extraction network, and encoding the image data into multi-scale features using the feature extraction network;

[0009] Fusing the boundary perception feature and the multi-scale feature to obtain a comprehensive feature;

[0010] In combination with the comprehensive feature context semantic information, the scope of construction land use in the project that has been started is determined.

[0011] Furthermore, the multi-scale conversion of the project boundary data and the construction boundary data to obtain boundary perception features includes:

[0012] Converting the project boundary data and the construction boundary data into binary images, and creating a channel based on each of the binary images;

[0013] According to the multi-channel fusion input of the binary image, the binary image is subjected to multi-scale transformation to obtain boundary perception features.

[0014] Furthermore, the combining of the comprehensive feature context semantic information to determine the scope of construction land use in the project that has been started includes:

[0015] Based on context semantic information, the context elements of the comprehensive features are divided into a plurality of soft target regions, each of the soft target regions corresponding to a class;

[0016] Determine the rough classification result of the construction land utilization range based on the soft target area;

[0017] The historical boundary information and prior information are incorporated into the rough classification results to obtain the scope of construction land use in the projects that have been started.

[0018] Furthermore, the rough classification result includes a determined area, an optional area and an excluded area.

[0019] Furthermore, the rough classification result is corrected by incorporating historical boundary information and prior information to obtain the scope of construction land use in the projects that have been started, including:

[0020] Structuring all pixels in the optional area into a plurality of soft object areas, and structuring all pixels in the determined area into a plurality of determined object areas;

[0021] After aggregating each of the soft object regions, determining the determined object region in the aggregated soft object regions according to the aggregated soft object regions and the determined object regions;

[0022] The scope of construction land use in the project that has been started is obtained according to the determined object area.

[0023] Furthermore, the step of fusing the boundary perception feature and the multi-scale feature to obtain a comprehensive feature includes:

[0024] The edge-aware features are obtained by fusing multi-level boundary-aware features in series;

[0025] The edge-sensing feature is fused with the multi-scale feature to obtain a comprehensive feature.

[0026] The present invention also provides a construction land use range extraction device, comprising:

[0027] A project data extraction module, used to extract project data from projects that have already started, wherein the project data includes project image data, project boundary data and construction boundary data;

[0028] A boundary perception feature extraction module, used for performing multi-scale conversion on the project boundary data and the construction boundary data to obtain boundary perception features;

[0029] A multi-scale feature extraction module, used for inputting the image data into a feature extraction network, and encoding the image data into multi-scale features using the feature extraction network;

[0030] A comprehensive feature fusion module, used for fusing the boundary perception feature and the multi-scale feature to obtain a comprehensive feature;

[0031] The construction land use scope extraction module is used to determine the construction land use scope in the project that has been started by combining the comprehensive feature context semantic information.

[0032] The present invention also provides a terminal device, comprising: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, the method for extracting the construction land utilization range as described above is implemented.

[0033] The present invention also provides a computer-readable storage medium, which includes a stored computer program; wherein, when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the construction land utilization range extraction method as described above.

[0034] The present invention combines multi-scale features and contextual semantic information, considers various influencing factors of the construction land use scope, and comprehensively analyzes the construction land use scope in the projects that have been started, so as to effectively improve the accuracy of the extraction of the construction land use scope. In addition, the construction land use scope in the projects that have been started is determined by combining the comprehensive features with the contextual semantic information without manual on-site survey, which not only reduces human subjective judgment and operational errors, but also effectively improves the efficiency of the extraction of the construction land use scope.

[0035] Furthermore, the present invention can provide accurate boundaries of previous construction stages through historical boundary information, and use this information to correct and fine-tune the initial coarse classification results, thereby improving the accuracy of construction land use range extraction; prior information can provide important context for construction projects, which can further improve the accuracy of construction land use range extraction. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a flow chart of a method for extracting a construction land utilization range provided by an embodiment of the present invention;

[0037] Figure 2 is another flow chart of the method for extracting the construction land utilization range provided by an embodiment of the present invention;

[0038] Figure 3 It is a schematic diagram of regional clustering with integrated historical boundaries provided by an embodiment of the present invention;

[0039] Figure 4 A schematic diagram of obtaining a new boundary by multi-scale boundary constraints provided by an embodiment of the present invention;

[0040] Figure 5 It is a structural schematic diagram of a construction land utilization range extraction device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0042] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the feature. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0043] In the description of this application, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0044] See also Figure 1 An embodiment of the present invention provides a method for extracting a construction land use range, comprising:

[0045] S1. Extracting project data from projects where construction has already started, the project data includes project image data, project boundary data, and construction boundary data;

[0046] In an embodiment of the present invention, by extracting project data stored in the archives of projects on land for construction that have already started, image data, project boundary data and construction boundary data can be obtained, wherein the image data can be obtained through remote sensing image analysis, and the project boundary data and construction boundary data can be determined in combination with a geographic information system.

[0047] S2, perform multi-scale transformation on the project boundary data and the construction boundary data to obtain boundary perception features;

[0048] In the embodiment of the present invention, construction projects often change over time, and multi-scale transformation can adapt to changes in the scope of construction land use, thereby capturing construction characteristics at different stages. In addition, multi-scale boundary perception features help the model better understand the context of the construction boundary, including surrounding terrain, existing buildings, and construction progress, thereby effectively improving the accuracy of construction land use scope extraction.

[0049] S3, inputting the image data into the feature extraction network, and using the feature extraction network to encode the image data into multi-scale features;

[0050] S4, integrating boundary-aware features and multi-scale features to obtain comprehensive features;

[0051] S5. Determine the scope of construction land use in projects that have already started by combining comprehensive feature context semantic information.

[0052] In an embodiment of the present invention, by combining multi-scale features and contextual semantic information, considering various influencing factors of the scope of construction land use, and comprehensively analyzing the scope of construction land use in projects that have already started, the accuracy of determining and extracting the scope of construction land use can be effectively improved. By combining comprehensive features with contextual semantic information to determine the scope of construction land use in projects that have already started, there is no need for manual on-site surveys, which not only reduces human subjective judgment and operational errors, but also effectively improves the efficiency of extracting the scope of construction land use.

[0053] See also Figure 2 , which is another flow chart of a construction land use range extraction method provided by an embodiment of the present invention. Among them, the feature extraction network is used to extract multi-scale features of image data, the multi-scale boundary constraint is used to obtain boundary perception features through multi-scale transformation, and the OCR attention mechanism includes SOR, ORR, CM, OCR and AR, among which ORR (Object representations) is used to obtain the feature representation of the target area by aggregating the feature representations of pixels in the corresponding target area; SOR (Soft Object Regi ons) is the region segmentation part, which divides the context pixels into a group of soft target areas, and each channel corresponds to a category result, which is used as a coarse classification result to calculate the feature representation of the subsequent target area; OCR (Object contextual representations) is to calculate the representation of the context of the target field; AR (Augmented representations) module is the output part, including the original pixel representation and the target context representation, and the two parts are spliced ​​together.

[0054] In one embodiment, step S2, performing multi-scale conversion on the project boundary data and the construction boundary data to obtain boundary perception features, includes:

[0055] S21, converting the project boundary data and the construction boundary data into binary images, and creating a channel based on each binary image;

[0056] S22. According to the multi-channel fusion input binary image, perform multi-scale transformation on the binary image to obtain boundary perception features.

[0057] In the embodiment of the present invention, the project boundary data and the construction boundary data are converted into a binary image, that is, an image with only two pixel values: pixels representing boundaries (which can be white or 1 value) and non-boundary pixels (which can be black or 0 value). In multi-channel image processing, each channel can store and process different information, similar to the red, green and blue channels in a color image.

[0058] The embodiment of the present invention fuses all created channels (ie, binary images) together to form a multi-channel input, and each channel can independently contribute its information to provide more comprehensive image information for subsequent processing.

[0059] The embodiment of the present invention obtains boundary perception information by performing multi-scale transformation on binary images. The binary images can provide clear boundary representation, thereby helping automatic detection and segmentation algorithms to more accurately identify the scope of construction land use.

[0060] In one embodiment, step S5, combining the comprehensive feature context semantic information to determine the scope of construction land use in the construction land project that has started, includes:

[0061] S51, based on context semantic information, dividing the context elements of the comprehensive features into a plurality of soft target regions, each soft target region corresponding to a class;

[0062] In the embodiment of the present invention, the soft target region is a region where the boundary of the object is allowed to be fuzzy. In the soft target region, each pixel is assigned a probability value, indicating the possibility that it belongs to a specific category (e.g., a construction area). This representation is not an absolute "yes" or "no", but provides a continuous range of possibilities.

[0063] S52, determining the rough classification result of the construction land use range based on the soft target area;

[0064] In one embodiment, the coarse classification result includes a determined area, an optional area, and an excluded area.

[0065] See also Figure 3 , where area A is the determined area, area B is the optional area, and area C is the excluded area.

[0066] S53. Incorporating historical boundary information and prior information to correct the coarse classification results, and obtain the scope of construction land use in the projects that have already started.

[0067] In an embodiment of the present invention, by incorporating historical boundary information and prior information to correct the coarse classification results, misjudgments caused by image noise, occlusion or complex background can be reduced, thereby reducing misclassified pixels and improving the accuracy and reliability of construction land use range extraction.

[0068] In one embodiment, step S53, integrating historical boundary information and prior information to correct the rough classification results, and obtaining the scope of construction land use in the projects that have been started, includes:

[0069] S531, structuring all pixels in the optional area into a plurality of soft object areas, and structuring all pixels in the determined area into a plurality of determined object areas;

[0070] In the embodiment of the present invention, the soft object area is a potential construction area. Based on the integration of historical boundary information and prior information, all pixels in the optional area are structured into several soft object areas, and all pixels in the determined area are structured into several determined object areas.

[0071] S532, after aggregating each soft object region, determine the determined object region in the aggregated soft object region according to the aggregated soft object region and the determined object region;

[0072] The embodiment of the present invention performs a clustering operation on each soft object region. The clustering may be based on similarity between pixels, spatial proximity or other relevant standards. According to the information of the aggregated soft object region and the determined object region, the determined object region included in the aggregated soft object region is determined. The determined object region included in the aggregated soft object region may be determined by comparison, overlap analysis or classification algorithm.

[0073] S533. Obtain the scope of construction land use in the project that has been started based on the determined target area.

[0074] The embodiments of the present invention can provide accurate boundaries of previous construction stages through historical boundary information. This information can be used to correct and fine-tune the initial coarse classification results, thereby improving the accuracy of construction land use range extraction; prior information can provide important context for construction projects, which can further improve the accuracy of construction land use range extraction.

[0075] In one embodiment, step S4, fusing the boundary perception feature and the multi-scale feature to obtain a comprehensive feature, includes:

[0076] S41, fusing multi-level boundary perception features in series to obtain edge perception features;

[0077] S42. The edge-aware features are integrated with the multi-scale features to obtain comprehensive features.

[0078] In the embodiment of the present invention, the historical boundary can be used as the approximate boundary of the ground truth to obtain a new boundary. Specifically, for each pixel, if its category label is different from the pixels in the approximate boundary, the pixel is regarded as an edge pixel. Figure 4 , we can find the new red dashed boundary based on the historical boundary in the middle blue. Under the supervision of the ground truth edge, the prediction of the object edge can be forced to learn based on the pixels of the features at each stage.

[0079] In an embodiment of the present invention, the multi-layer features of each stage effectively learn clues about the boundaries of objects, and the edge-aware features can be obtained by serially fusing multi-level boundary-aware features and following multiple 3X3 convolutional layers.

[0080] Therefore, the interactive fusion of edge-aware features with the multi-scale features of the backbone network can seamlessly integrate boundary guidance into the multi-level feature fusion operation, effectively combining multi-scale features while maintaining clear target boundaries.

[0081] In the embodiment of the present invention, since the objects in the remote sensing images are different in scale, and objects of different categories may show similar morphological features. In certain cases, over-reliance on edge information may reduce segmentation performance. Therefore, the edge-aware features of the embodiment of the present invention are re-weighted, and then multiplied with the original backbone features and summed to adaptively balance the weights between the two feature maps to generate the final output.

[0082] The implementation of the embodiments of the present invention has the following beneficial effects:

[0083] The embodiments of the present invention combine multi-scale features and contextual semantic information, consider various influencing factors of the construction land use scope, and comprehensively analyze the construction land use scope in the projects that have been started, so as to effectively improve the accuracy of the construction land use scope extraction. In addition, by combining the comprehensive features with the contextual semantic information to determine the construction land use scope in the projects that have been started, no manual on-site survey is required, which not only reduces human subjective judgment and operational errors, but also can effectively improve the efficiency of the construction land use scope extraction.

[0084] Furthermore, the embodiments of the present invention can provide accurate boundaries of previous construction stages through historical boundary information. This information can be used to correct and fine-tune the initial coarse classification results, thereby improving the accuracy of construction land use range extraction; prior information can provide important context for construction projects, which can further improve the accuracy of construction land use range extraction.

[0085] See also Figure 5 Based on the same inventive concept as the above embodiment, the present invention also provides a construction land use range extraction device, comprising:

[0086] The project data extraction module 10 is used to extract the project data in the construction land project that has been started, and the project data includes project image data, project boundary data and construction boundary data;

[0087] The boundary perception feature extraction module 20 is used to perform multi-scale conversion on the project boundary data and the construction boundary data to obtain boundary perception features;

[0088] The multi-scale feature extraction module 30 is used to input the image data into the feature extraction network and encode the image data into multi-scale features using the feature extraction network;

[0089] A comprehensive feature fusion module 40 is used to fuse the boundary perception feature and the multi-scale feature to obtain a comprehensive feature;

[0090] The construction land use range extraction module 50 is used to determine the construction land use range in the project that has been started by combining the comprehensive feature context semantic information.

[0091] In one embodiment, the boundary-aware feature extraction module 20 is further used to:

[0092] Convert the project boundary data and the construction boundary data into binary images, and create a channel based on each binary image;

[0093] According to the multi-channel fusion input binary image, the binary image is multi-scale transformed to obtain the boundary perception feature.

[0094] In one embodiment, the construction land use range extraction module 50 is further used to:

[0095] Based on the contextual semantic information, the contextual elements of the comprehensive features are divided into several soft target regions, each of which corresponds to a class;

[0096] The coarse classification results of the construction land use range are determined based on the soft target area;

[0097] The historical boundary information and prior information are incorporated into the coarse classification results to obtain the scope of construction land use in the projects that have been started.

[0098] In one embodiment, the coarse classification result includes a determined area, an optional area, and an excluded area.

[0099] In one embodiment, the rough classification results are corrected by integrating historical boundary information and prior information to obtain the scope of construction land use in the projects that have been started, including:

[0100] Structuring all pixels in the optional area into a plurality of soft object areas, and structuring all pixels in the determined area into a plurality of determined object areas;

[0101] After aggregating each soft object region, determining the determined object region in the aggregated soft object region according to the aggregated soft object region and the determined object region;

[0102] The scope of construction land use in the projects that have been started is obtained based on the determined object area.

[0103] In one embodiment, the comprehensive feature fusion module 40 is further used for:

[0104] The edge-aware features are obtained by fusing multi-level boundary-aware features in series;

[0105] The edge-aware features are fused with the multi-scale features to obtain comprehensive features.

[0106] Correspondingly, an embodiment of the present invention also provides a terminal device, including: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, it implements the construction land utilization range extraction method of any one of the above embodiments.

[0107] The terminal device of this embodiment includes: a processor, a memory, and a computer program and computer instructions stored in the memory and executable on the processor. When the processor executes the computer program, each step in the above embodiment 1 is implemented, for example Figure 1 Alternatively, when the processor executes the computer program, the functions of each module / unit in the above device embodiment, such as the multi-scale feature extraction module 30, are implemented.

[0108] Exemplarily, the computer program can be divided into one or more modules / units, one or more modules / units are stored in a memory and executed by a processor to complete the present invention. One or more modules / units can be a series of computer program instruction segments that can perform specific functions, and the instruction segments are used to describe the execution process of the computer program in a terminal device. For example, the multi-scale feature extraction module 30 is used to input image data into a feature extraction network, and use the feature extraction network to encode the image data into multi-scale features.

[0109] The terminal device may be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that the schematic diagram is merely an example of a terminal device and does not constitute a limitation on the terminal device. The terminal device may include more or fewer components than shown in the diagram, or may combine certain components, or different components. For example, the terminal device may also include an input / output device, a network access device, a bus, etc.

[0110] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, and uses various interfaces and lines to connect various parts of the entire terminal device.

[0111] The memory can be used to store computer programs and / or modules. The processor realizes various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the mobile terminal, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Med i aCard, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0112] Among them, if the module / unit integrated in the terminal device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, RandomAccess Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electric carrier signals and telecommunication signals.

[0113] Correspondingly, an embodiment of the present invention also provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the construction land utilization range extraction method as described in any one of the above embodiments.

[0114] The above specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for extracting the scope of construction land use, characterized in that: include: Extracting project data from a construction site project that has already started, the project data including project image data, project boundary data, and construction boundary data; The project boundary data and the construction boundary data are subjected to multi-scale transformation to obtain boundary perception features; the multi-scale transformation of the project boundary data and the construction boundary data to obtain boundary perception features includes: converting the project boundary data and the construction boundary data into binary images, and creating a channel based on each of the binary images; inputting the binary images according to multi-channel fusion, and performing multi-scale transformation on the binary images to obtain boundary perception features; Inputting the image data into a feature extraction network, and encoding the image data into multi-scale features using the feature extraction network; Fusing the boundary perception feature and the multi-scale feature to obtain a comprehensive feature; Combined with the comprehensive feature contextual semantic information, determine the scope of construction land use in the project that has started; combined with the comprehensive feature contextual semantic information, determine the scope of construction land use in the project that has started, including: based on the contextual semantic information, divide the context elements of the comprehensive feature into a number of soft target areas, each of the soft target areas corresponds to a class; wherein the soft target area is an area where fuzzy boundaries are allowed to exist; determine a coarse classification result of the construction land use scope based on the soft target area; incorporate historical boundary information and prior information to correct the coarse classification result, and obtain the scope of construction land use in the project that has started.

2. The method for extracting the construction land use range according to claim 1, characterized in that: The rough classification result includes a determined area, an optional area, and an excluded area.

3. The method for extracting the construction land use range according to claim 1, characterized in that: The rough classification result is corrected by incorporating historical boundary information and prior information to obtain the scope of construction land use in the projects that have been started, including: Structuring all pixels in the optional area into a plurality of soft object areas, and structuring all pixels in the determined area into a plurality of determined object areas; After aggregating each of the soft object regions, determining the determined object region in the aggregated soft object regions according to the aggregated soft object regions and the determined object regions; The scope of construction land use in the project that has been started is obtained according to the determined object area.

4. The method for extracting the construction land use range according to claim 1, characterized in that: The step of fusing the boundary perception feature and the multi-scale feature to obtain a comprehensive feature includes: The edge-aware features are obtained by fusing multi-level boundary-aware features in series; The edge-sensing feature is fused with the multi-scale feature to obtain a comprehensive feature.

5. A construction land use range extraction device, characterized in that: include: A project data extraction module, used to extract project data from projects on land for construction that have already started, the project data including project image data, project boundary data and construction boundary data; The boundary perception feature extraction module is used to perform multi-scale transformation on the project boundary data and the construction boundary data to obtain boundary perception features; specifically, it is used to: transform the project boundary data and the construction boundary data into binary images, and create a channel based on each binary image; input the binary image according to multi-channel fusion, perform multi-scale transformation on the binary image, and obtain boundary perception features; A multi-scale feature extraction module, used for inputting the image data into a feature extraction network, and encoding the image data into multi-scale features using the feature extraction network; A comprehensive feature fusion module, used for fusing the boundary perception feature and the multi-scale feature to obtain a comprehensive feature; The construction land use scope extraction module is used to determine the construction land use scope in the project that has been started in combination with the contextual semantic information of the comprehensive feature; specifically, it is used to: divide the context elements of the comprehensive feature into a number of soft target areas based on the contextual semantic information, and each soft target area corresponds to a class; wherein the soft target area is an area where fuzzy boundaries are allowed to exist; determine the coarse classification result of the construction land use scope based on the soft target area; and correct the coarse classification result by incorporating historical boundary information and prior information to obtain the construction land use scope in the project that has been started.

6. A terminal device, characterized in that: include: A processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method for extracting the construction land use range as described in any one of claims 1 to 4 is implemented.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program; wherein, when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the construction land utilization range extraction method as described in any one of claims 1-4.

Images