Disturbance land parcel identification method, device, equipment, medium and product
By collecting and processing remote sensing image data, training deep learning models to identify disturbed plots, solving the problem of insufficient identification accuracy in the existing technology, improving the identification accuracy of disturbed plots, and supporting soil erosion supervision.
Patent Information
- Application Number
- CN202410907030.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-08
- Publication Date
- 2024-12-06
AI Technical Summary
The existing technology has a relatively prominent phenomenon of misleading and missing lifting in the identification of disturbed plots, resulting in insufficient identification accuracy and it is difficult to effectively supervise artificial soil erosion.
By collecting multiple sets of sample data, including early and late remote sensing images, processing and screening, labeling disturbed plots, training deep learning models, identifying disturbed plots in remote sensing images to be identified, and improving sample data quality and identification accuracy.
It improves the accuracy of disturbed land identification, reduces manual intervention, and supports the soil erosion supervision work of production and construction projects.
Smart Images

Figure CN119091311A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of environmental monitoring technology, and in particular to a method, device, equipment, medium and product for identifying disturbed land. Background Art
[0002] The impact of production and construction projects on soil erosion appears to have limited space on the surface, but in fact the accumulation effect is very obvious. At the same time, different types of production and construction projects cause quite different disturbed geomorphic units, soil erosion characteristics and hazards. In addition, such projects cause frequent disturbances and are difficult to supervise.
[0003] In recent years, based on high-resolution remote sensing images, the remote sensing automatic extraction model combined with manual auxiliary correction can quickly and accurately detect the disturbance caused by production and construction projects, and then predict and evaluate the risks of man-made soil and water loss that may be caused. This is of great significance for effectively improving supervision efficiency, curbing man-made soil and water loss, and further improving the regional ecological environment. At present, pilot remote sensing supervision has begun for some plots, and artificial + intelligent means have been gradually applied to extract new interpretation spots and identify changed plots. The manual judgment method has problems such as low efficiency and different recognition standards. The use of deep learning artificial intelligence extraction methods can quickly identify and recognize the characteristics and natural changes of the plots, and then identify whether there are business changes based on learning experience, so as to improve the scientific and efficient decision-making and management. However, in the current automatic identification of disturbed plots, the phenomenon of false and missed mentions is still relatively prominent, so the accuracy of disturbed plot identification still needs to be improved. Summary of the invention
[0004] The purpose of this application is to provide a disturbed land parcel identification method, device, equipment, medium and product, which can improve the accuracy of disturbed land parcel identification.
[0005] To achieve the above objectives, this application provides the following solutions:
[0006] In a first aspect, the present application provides a disturbed land parcel identification method, comprising:
[0007] Collect multiple sets of sample data; each set of sample data includes early remote sensing images and late remote sensing images of the same area, and the disturbed plots in the late remote sensing images are buildings;
[0008] Processing and screening each group of sample data to obtain multiple groups of sample data that meet set requirements; the set requirements include that the difference in resolution between the early remote sensing image and the later remote sensing image in each group of sample data is less than a set value, the coordinate system is the same, and the area occupied by clouds does not exceed a set percentage;
[0009] Mark the disturbed plots in the later remote sensing images of each group of sample data that meets the set requirements;
[0010] Determine the training sample group according to each group of labeled sample data;
[0011] The deep learning model is trained using the training sample group to obtain a disturbed land identification model;
[0012] The disturbed land parcel recognition model is used to recognize disturbed land parcels in the later remote sensing images of the remote sensing image group to be recognized.
[0013] Optionally, the disturbed plots in the later remote sensing images of each group of sample data that meets the set requirements are marked, specifically including:
[0014] Rolling curtain analysis is performed on each group of sample data that meets the set requirements to determine the disturbed plots, and the disturbed plots are delineated to obtain each group of labeled sample data.
[0015] Optionally, determining a training sample group according to each group of labeled sample data specifically includes:
[0016] Perform vector-to-raster conversion on the early remote sensing images and the late remote sensing images in each group of labeled sample data;
[0017] The early remote sensing images and the late remote sensing images after vector-to-raster conversion are cropped according to the set size to obtain the early remote sensing image blocks and the late remote sensing image blocks after cropping. The late remote sensing image blocks containing the delineated disturbed plots and the early remote sensing image blocks corresponding to the late remote sensing image blocks constitute a sample block group, and each sample block group constitutes a training sample group.
[0018] Optionally, the set size is 256*256.
[0019] Optionally, the set percentage is 3%.
[0020] Optionally, the coordinate system is the National Geodetic Coordinate System 2000.
[0021] In a second aspect, the present application provides a disturbed land block identification device, comprising:
[0022] The sample data collection module is used to collect multiple sets of sample data; each set of sample data includes early remote sensing images and late remote sensing images of the same area, and the disturbed plots in the late remote sensing images are buildings;
[0023] The sample data processing module is used to process and screen each group of sample data to obtain multiple groups of sample data that meet the set requirements; the set requirements include that the difference in resolution between the early remote sensing image and the later remote sensing image in each group of sample data is less than a set value, the coordinate system is the same, and the area occupied by clouds does not exceed a set percentage;
[0024] The labeling module is used to label the disturbed plots in the later remote sensing images of each group of sample data that meets the set requirements;
[0025] A training sample group determination module is used to determine the training sample group according to each group of labeled sample data;
[0026] A deep learning model training module is used to train a deep learning model using a training sample group to obtain a disturbed land parcel recognition model;
[0027] The disturbed land parcel identification model application module is used to use the disturbed land parcel identification model to identify the disturbed land parcels in the later remote sensing images of the remote sensing image group to be identified.
[0028] In a third aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-described methods for identifying disturbed land parcels.
[0029] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for identifying disturbed land parcels.
[0030] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned methods for identifying disturbed land parcels.
[0031] According to the specific embodiments provided in this application, this application discloses the following technical effects:
[0032] The present application provides a method, device, equipment, medium and product for identifying disturbed plots, which uses newly built buildings in the later remote sensing images of two phases as disturbed plots, making sample data more targeted, processing and screening each group of sample data so that the sample data meets the set requirements, improving the quality of the sample data, and thus improving the accuracy of disturbed plot identification. BRIEF DESCRIPTION OF THE DRAWINGS
[0033] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. 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 paying creative work.
[0034] Figure 1 A schematic diagram of a flow chart of a disturbed land parcel identification method provided in one embodiment of the present application;
[0035] Figure 2 A schematic diagram of a remote sensing image processing process provided by an embodiment of the present application;
[0036] Figure 3 A schematic diagram of a disturbed plot provided in an embodiment of the present application;
[0037] Figure 4 A schematic diagram of a vector conversion grid provided in an embodiment of the present application;
[0038] Figure 5 A schematic diagram of a single sample provided for an embodiment of the present application;
[0039] Figure 6 A schematic diagram of functional modules of a disturbed land parcel identification device provided in another embodiment of the present application.
[0040] Figure 7 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. 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 order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.
[0043] The present application embodiment provides a method for identifying disturbed land parcels, such as Figure 1 As shown, the process includes the following steps.
[0044] Step 101: Collect multiple sets of sample data; each set of sample data includes an early remote sensing image and a late remote sensing image of the same area, and the disturbed land in the late remote sensing image is a building.
[0045] Step 102: Process and screen each group of sample data to obtain multiple groups of sample data that meet the set requirements; the set requirements include that the difference in resolution between the early remote sensing image and the later remote sensing image in each group of sample data is less than a set value, the coordinate system is the same, and the area occupied by clouds does not exceed a set percentage.
[0046] Step 103: marking the disturbed plots in the later remote sensing images of each group of sample data that meets the set requirements.
[0047] Step 104: Determine a training sample group based on the labeled sample data groups.
[0048] Step 105: Use the training sample group to train the deep learning model to obtain a disturbed land parcel recognition model.
[0049] Step 106: using the disturbed land parcel identification model to identify disturbed land parcels in the later remote sensing images in the remote sensing image group to be identified.
[0050] By implementing the above steps 101 to 106, the newly built buildings in the later remote sensing images of the two phases are used as disturbed plots, so that the sample data are more targeted, and each group of sample data is processed and screened so that the sample data meets the set requirements, thereby improving the quality of the sample data, thereby improving the accuracy of disturbed plot identification, reducing the degree of human intervention, and effectively supporting the monitoring and management of man-made soil and water loss in production and construction projects.
[0051] Before making samples, this application must first determine the rules for making samples and the requirements for high-quality samples. In this application, the main focus is on comparing and analyzing the remote sensing images of the previous and subsequent periods to extract disturbed plots. The plots involved in the samples are mainly new buildings, such as changes in land use types such as bare land, water bodies, forests and grasslands to the main body of buildings. The disturbance boundary should be clear and complete, and the shadow area should be minimized as much as possible. When outlining, the lines need to be clear, continuous, and fit the buildings.
[0052] The multiple groups of sample data collected in step 101 are specifically multiple groups of sample data from different regions, each group of sample data is a remote sensing image of a region in two periods, the remote sensing image earlier in time is the early remote sensing image, and the remote sensing image later in time is the late remote sensing image.
[0053] The disturbed land identification method described before step 101 further includes: remote sensing image processing process, such as Figure 2 As shown, it specifically includes: preprocessing, geometric correction, image enhancement, image cropping, image mosaicking and color grading of the collected remote sensing images in sequence, so as to obtain remote sensing information, that is, the remote sensing information of the remote sensing images in the sample data in step 101. The preprocessing includes noise reduction processing, thin cloud processing and shadow processing in sequence; the geometric correction includes image registration, rough geometric correction, fine geometric correction and orthorectification in sequence; the image enhancement includes color synthesis, histogram conversion, density segmentation, grayscale inversion, inter-image operation, domain enhancement and image fusion; the image mosaicking and color grading include image mosaicking and color grading; after the remote sensing information is obtained, it also includes the process of manual interpretation and image classification, and the sample data group that meets the requirements of the disturbed land in the later remote sensing image in step 101 as a building is screened out through image classification.
[0054] The percentage set in step 102 is 3%, and the coordinate system is the 2000 National Geodetic Coordinate System (CGCS2000). By setting the area occupied by clouds in remote sensing images to not exceed 3%, the remote sensing images are mainly clear sky images, reducing the impact of snow or cloud cover on the images.
[0055] In order to ensure the consistency of time and resolution of images of the same period, the medium resolution is set to be better than 2m, and the difference in resolution between the early remote sensing images and the later remote sensing images is within two pixels. The images in the sample data of this application are also required to be clear, with distinct layers of objects and uniform color, and the images must have no bad lines, missing bands, stripes, speckle noise and flare.
[0056] Among them, step 103 specifically includes: performing rolling curtain analysis on each set of sample data that meets the set requirements to determine the disturbed plots, delineating the disturbed plots, and obtaining each set of labeled sample data. The accuracy of delineating the disturbed plots is improved by rolling curtain analysis. More specifically, remote sensing processing software is used to perform rolling curtain analysis on each set of sample data that meets the set requirements, identify disturbance changes and delineate the disturbed plots, that is, determine the disturbed plots of newly added buildings in the later remote sensing images, and delineate the boundaries of the disturbed plots, such as Figure 3 As shown, be careful to avoid drawing large shadows and blurring the boundaries. Figure 3 The early images in (a) are early remote sensing images. Figure 3 (b) Late images are late remote sensing images. Figure 3 The area determined in (c) is the disturbed land area. Figure 3 In (d), the extraction of patches is to outline the boundaries of the disturbed plots.
[0057] Step 104 specifically includes: performing vector-to-raster conversion on the early remote sensing images and the late remote sensing images in each group of labeled sample data, such as Figure 4 As shown, Figure 4 (a) is the image before vector to raster conversion. Figure 4 (b) is the image after vector-to-raster conversion. The processed file format is tif or img. The number of image rows and columns of the early remote sensing image and the later remote sensing image after raster processing is guaranteed to be consistent.
[0058] The early remote sensing images and the late remote sensing images after the vector-to-raster conversion process are cropped according to the set size to obtain the early remote sensing image blocks and the late remote sensing image blocks after the cropping process. The late remote sensing image blocks containing the delineated disturbed plots and the early remote sensing image blocks corresponding to the late remote sensing image blocks constitute a sample block group, and each sample block group constitutes a training sample group. Step 104 specifically uses PYTHON to write software code to crop the raster data of the early and late remote sensing images and the disturbed patches. The set cropping size is 256*256 (the pixel is 256*256), the disturbed patches (disturbed plots) are retained, and the non-disturbed patches are deleted. A set of valid samples including the early remote sensing images, the late remote sensing images and the corresponding labels are generated to obtain a high-quality sample group (training sample group), wherein the labels are the delineation of the disturbed plots, as shown in the figure for a single sample. Figure 5 As shown, Figure 5 There are three samples in total, including sample 1, sample 2 and sample 3, where A represents the early remote sensing image and B represents the later remote sensing image.
[0059] Step 106 specifically includes: inputting the early remote sensing images and the late remote sensing images in the remote sensing image group to be identified into the disturbed land parcel identification model, and outputting the disturbed land parcel identification result. The disturbed land parcel identification result is the annotation of the disturbed land parcel in the late remote sensing image.
[0060] This application trains a deep learning model through a high-quality sample group to obtain a disturbed plot recognition model, which can improve the accuracy of the disturbed plot recognition model in identifying disturbed plots.
[0061] Based on the same inventive concept, the embodiment of the present application further provides a disturbed land parcel identification device for implementing the disturbed land parcel identification method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations of one or more disturbed land parcel identification device embodiments provided below can refer to the limitations of the disturbed land parcel identification method above, and will not be repeated here.
[0062] In an exemplary embodiment, Figure 6 As shown, a disturbed land parcel identification device is provided, including: a sample data acquisition module 201, used to collect multiple groups of sample data; each group of sample data includes an early remote sensing image and a later remote sensing image of the same area, and the disturbed land parcels in the later remote sensing image are buildings.
[0063] The sample data processing module 202 is used to process and screen each group of sample data to obtain multiple groups of sample data that meet the set requirements; the set requirements include that the difference in resolution between the early remote sensing image and the later remote sensing image in each group of sample data is less than a set value, the coordinate system is the same, and the area occupied by clouds does not exceed a set percentage.
[0064] The marking module 203 is used to mark the disturbed plots in the later remote sensing images of each group of sample data that meets the set requirements.
[0065] The training sample group determination module 204 is used to determine the training sample group according to each group of labeled sample data.
[0066] The deep learning model training module 205 is used to train the deep learning model using a training sample group to obtain a disturbed land parcel recognition model.
[0067] The disturbed land parcel identification model application module 206 is used to use the disturbed land parcel identification model to identify the disturbed land parcels in the later remote sensing images in the remote sensing image group to be identified.
[0068] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 7 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store video tag processing data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a video tag processing method is implemented.
[0069] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0070] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.
[0071] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.
[0072] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.
[0073] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.
[0074] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).
[0075] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, etc., but is not limited thereto.
[0076] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0077] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.
Claims
1. A disturbed land block identification method, characterized in that: The disturbed land block identification method comprises: Collect multiple sets of sample data; each set of sample data includes early remote sensing images and late remote sensing images of the same area, and the disturbed plots in the late remote sensing images are buildings; Processing and screening each group of sample data to obtain multiple groups of sample data that meet set requirements; the set requirements include that the difference in resolution between the early remote sensing image and the later remote sensing image in each group of sample data is less than a set value, the coordinate system is the same, and the area occupied by clouds does not exceed a set percentage; Mark the disturbed plots in the later remote sensing images of each group of sample data that meets the set requirements; Determine the training sample group according to each group of labeled sample data; The deep learning model is trained using the training sample group to obtain a disturbed land identification model; The disturbed land parcel recognition model is used to recognize disturbed land parcels in the later remote sensing images of the remote sensing image group to be recognized.
2. The disturbed land block identification method according to claim 1, characterized in that: The disturbed plots in the later remote sensing images of each set of sample data that meets the set requirements are marked, including: Rolling curtain analysis is performed on each group of sample data that meets the set requirements to determine the disturbed plots, and the disturbed plots are delineated to obtain each group of labeled sample data.
3. The disturbed land block identification method according to claim 1, characterized in that: The training sample groups are determined according to the labeled sample data groups, including: Perform vector-to-raster conversion on the early remote sensing images and the late remote sensing images in each group of labeled sample data; The early remote sensing images and the late remote sensing images after vector-to-raster conversion are cropped according to the set size to obtain the early remote sensing image blocks and the late remote sensing image blocks after cropping. The late remote sensing image blocks containing the delineated disturbed plots and the early remote sensing image blocks corresponding to the late remote sensing image blocks constitute a sample block group, and each sample block group constitutes a training sample group.
4. The disturbed land block identification method according to claim 3, characterized in that: The set size is 256*256.
5. The disturbed land block identification method according to claim 1, characterized in that: The set percentage is 3%.
6. The disturbed land identification method according to claim 1, characterized in that: The coordinate system is the 2000 National Geodetic Coordinate System.
7. A disturbed land block identification device, characterized in that: The disturbed land block identification device comprises: The sample data collection module is used to collect multiple sets of sample data; each set of sample data includes early remote sensing images and late remote sensing images of the same area, and the disturbed plots in the late remote sensing images are buildings; The sample data processing module is used to process and screen each group of sample data to obtain multiple groups of sample data that meet the set requirements; the set requirements include that the difference in resolution between the early remote sensing image and the later remote sensing image in each group of sample data is less than a set value, the coordinate system is the same, and the area occupied by clouds does not exceed a set percentage; The labeling module is used to label the disturbed plots in the later remote sensing images of each set of sample data that meets the set requirements; A training sample group determination module is used to determine the training sample group according to each group of labeled sample data; A deep learning model training module is used to train a deep learning model using a training sample group to obtain a disturbed land parcel recognition model; The disturbed land parcel identification model application module is used to use the disturbed land parcel identification model to identify the disturbed land parcels in the later remote sensing images of the remote sensing image group to be identified.
8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the disturbed land block identification method according to any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the disturbed land block identification method described in any one of claims 1 to 6 is implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the disturbed land block identification method described in any one of claims 1 to 6 is implemented.
