Land coverage classification method based on Sentinel-2AB time sequence data
Through Sentinel-2AB time series data, the NDVI and TCW characteristics of land objects were analyzed, combined with SVM and MLC methods, the misclassification phenomenon in single-time phase remote sensing image classification was solved, and the classification accuracy of land cover types was improved.
Patent Information
- Application Number
- CN202510316005.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-18
- Publication Date
- 2025-08-08
AI Technical Summary
In the prior art, traditional single-time phase remote sensing image classification method has the problem of inaccurate classification results in land cover type classification, especially when misalignment occurs between land types with similar spectral curves.
The land cover classification method based on Sentinel-2AB time series data was used to analyze the normalized difference vegetation index (NDVI) and the humidity component (TCW) time series characteristics of typical land objects, and combined with support vector machine method (SVM) and maximum likelihood method (MLC).
The classification accuracy of land cover types is improved and the misclassification phenomenon is significantly reduced, especially the distinction between land objects with similar spectral curves, achieving higher classification accuracy and accuracy.
Smart Images
Figure CN120451615A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of land cover type classification, and in particular to a land cover classification method based on Sentinel-2AB time series data. Background Art
[0002] To date, many researchers have used a variety of classification methods to classify land cover types in remote sensing images of various resolutions. It is generally believed that the choice of classification method significantly affects the results. Traditional single-temporal remote sensing image classification methods can be divided into supervised classification and unsupervised classification. Each method has its own advantages and disadvantages. Currently, supervised classification is the most widely used, and one of the most commonly used methods in supervised classification is the maximum likelihood method. Summary of the Invention
[0003] The technical problem to be solved by the present invention is to overcome the shortcomings of the existing technology and provide a land cover classification method based on Sentinel-2AB time series data.
[0004] In order to solve the above technical problems, the technical solutions of the present invention are as follows: A land cover classification method based on Sentinel-2AB time series data, including: Select the target research area; Analyze the time series characteristics of the normalized difference vegetation index of typical features in the target study area; Analyze the time series characteristics of the moisture component of the tasseled cap of typical landforms in the target study area; Select the best time series features; Support vector machine and maximum likelihood methods were used for land cover classification.
[0005] As a preferred embodiment of the land cover classification method based on Sentinel-2AB time series data of the present invention, the time series characteristics of the normalized difference vegetation index of typical landforms in the target study area are analyzed, including: The normalized vegetation index is calculated using formula 1: NDVI = (ρNIR-ρR) / (ρNIR+ρR), where ρNIR and ρR represent the reflectance values of the near-infrared band and the red light band, respectively.
[0006] As a preferred embodiment of the land cover classification method based on Sentinel-2AB time series data of the present invention, the time series characteristics of the moisture component of the tasseled cap changes of typical landforms in the target study area are analyzed, including: The humidity component of the tasseled cap change is calculated using Formula 2: TCW = 0.1509ρ2 + 0.1973ρ3 + 0.3279ρ4 + 0.3406ρ8 + 0.7112ρ11 + 0.4572ρ12, where ρ2, ρ3, ρ4, ρ8, ρ11, and ρ12 are the reflectances of bands 2, 3, 4, 8, 11, and 12 in the Sentinel-2 image, respectively.
[0007] As a preferred solution of the land cover classification method based on Sentinel-2AB time series data of the present invention, after the land cover classification is performed using the support vector machine method and the maximum likelihood method, the method further includes: Perform accuracy analysis on land cover classification.
[0008] As a preferred solution of the land cover classification method based on Sentinel-2AB time series data of the present invention, the accuracy analysis of land cover classification includes: Google Earth data was used to conduct accuracy analysis of land cover classification.
[0009] The present invention also provides a land cover classification device based on Sentinel-2AB time series data, comprising: Selection module, used to select the target research area; The first analysis module is used to analyze the time series characteristics of the normalized difference vegetation index of typical features in the target study area; The second analysis module is used to analyze the time series characteristics of the moisture component of the tasseled cap changes of typical landforms in the target study area; Selection module, used to select the best time series features; Classification module for land cover classification using support vector machine and maximum likelihood methods.
[0010] As a preferred solution of the land cover classification device based on Sentinel-2AB time series data of the present invention, it also includes: The accuracy analysis module is used to perform accuracy analysis on land cover classification.
[0011] The present invention also provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method described in any one of the above-mentioned land cover classification methods based on Sentinel-2AB time series data is implemented.
[0012] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any of the above land cover classification methods based on Sentinel-2AB time series data.
[0013] The beneficial effects of the present invention are: The Sentinel-2 NDVI time series data in this paper can well reflect the phenological information of different land cover vegetation. The Sentinel-2 TCW time series data can well reflect the characteristics of non-vegetated land features as well as the characteristics of vegetation. Then, based on the months with obvious land feature characteristics, the best time series data is selected. Combined with the high-resolution data of Sentinel-2, land feature samples can be accurately selected and supervised classification can be performed. The classification results are good and can fully meet the macro requirements of land cover classification. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0015] Figure 1 A schematic diagram of the process of land cover classification based on Sentinel-2AB time series data provided by the present invention; Figure 2 Schematic diagram of a land cover classification device based on Sentinel-2AB time series data provided by the present invention; Figure 3 A schematic diagram of a computer device provided by the present invention. DETAILED DESCRIPTION
[0016] In order to make the contents of the present invention more clearly understood, the present invention is further described below in detail based on specific implementation methods in conjunction with the accompanying drawings.
[0017] See also Figure 1 , this application provides a land cover classification method based on Sentinel-2AB time series data, which specifically includes the following steps: Step S101: Select a target research area.
[0018] Specifically, in this embodiment, Sheli Town, Da'an City in western Jilin Province and Wuning County in northwestern Jiangxi Province are targeted as research areas.
[0019] Step S102: Analyze the time series characteristics of the normalized difference vegetation index of typical features in the target study area.
[0020] Specifically, research by scholars both domestically and internationally indicates that NDVI is the best indicator of vegetation production and coverage. NDVI is calculated by taking the difference and ratio of the reflectance between the near-infrared and red bands, extracting vegetation information. The calculation formula is: NDVI = (ρNIR - ρR) / (ρNIR + ρR), where ρNIR and ρR represent the reflectance values of the near-infrared and red bands, respectively.
[0021] Step S103: Analyze the time series characteristics of the humidity component of the tasseled cap changes of typical landforms in the target study area.
[0022] Specifically, Nedkov derived the transformation coefficients for the KT transformation of Sentinel-2 images in 2017. The moisture component (TCW) is sensitive to soil and vegetation moisture. High soil moisture results in a large moisture component, while low soil moisture results in a small moisture component. The calculation formula for the Tasseled Cap Transformation varies depending on the sensor. This paper applies the Tasseled Cap Transformation to Sentinel-2A / B images using the following formula: TCW=0.1509ρ2+0.1973ρ3+0.3279ρ4+0.3406ρ8+0.7112ρ11+0.4572ρ12 (3-2) where ρ2, ρ3, ρ4, ρ8, ρ11, and ρ12 are the reflectances of the 2nd, 3rd, 4th, 8th, 11th, and 12th bands in the Sentinel-2 image, respectively.
[0023] Step S104: Select the best time series feature.
[0024] Step S105: using support vector machine method and maximum likelihood method to perform land cover classification.
[0025] Step S106: performing accuracy analysis on the land cover classification.
[0026] Specifically, Google Earth data is used for accuracy analysis.
[0027] Sheli Town: The classification results of MLC combined with single-temporal imagery show serious misclassification: the misclassification between forest land and cultivated land is due to the fact that they have similar spectral curves in the single-temporal imagery in July. The misclassification is also serious between construction land, bare land, and saline-alkali land, resulting in poor classification results. In contrast, it can be clearly observed that the classification results of MLC combined with time series data are better, and can well distinguish construction land, bare land, and saline-alkali land. This is because the TCW time series characteristics of the three are obvious. The misclassification between forest land and cultivated land is significantly reduced, and the classification accuracy is significantly improved. This is because the NDVI time series characteristics between forest land and cultivated land are more obvious. Comparing (c) and (d), it can be seen that in the classification results of SVM combined with single-temporal imagery, there is misclassification between bare land and grassland and saline-alkali land with similar spectral curves, and there is also misclassification between forest land and cultivated land. Comparing (b) and (d), it can be seen that the MLC combined with time series data classification results misclassified some construction land as saline-alkali land, and the SVM combined with time series data classification method can reduce the fragmented areas that appear in the MLC combined with time series data classification results. Therefore, this study area is more suitable for land cover classification research using the SVM combined with time series data classification method.
[0028] Comparing the classification results of land cover types of single-phase data and time series data, time series data can fully display the characteristics of land features, and the classification results are high. In the study area of Sheli Town, the overall accuracy of time series data combined with MLC is 82.39%, and the Kappa coefficient is 0.78. The overall accuracy of time series data combined with SVW reaches 88.24%, and the Kappa coefficient reaches 0.83. The overall accuracy of single-phase data combined with SVW is 78.52%, and the Kappa coefficient is only 0.71. The overall accuracy of single-phase data combined with MLC is 70.68%, and the Kappa coefficient is only 0.60. In the Wuning County study area, the overall accuracy of time series data combined with MLC was 83.26%, and the Kappa coefficient was 0.80. The overall accuracy of time series data combined with SVW reached 86.24%, and the Kappa coefficient reached 0.84. The overall accuracy of single-phase data combined with SVW was 75.38%, and the Kappa coefficient was only 0.69. The overall accuracy of single-phase data combined with SVW was 70.32%, and the Kappa coefficient was only 0.61.
[0029] In the study area of Sheli Town, by combining time series data with SVW classification and single-time SVW classification, it was found that the mapping accuracy of forest land, grassland and cultivated land increased by 10.22, 9.75 and 11.11 percentage points respectively; in the study area of Wuning County, by combining time series data with SVW classification and single-time SVW classification, it was found that the mapping accuracy of evergreen forest, deciduous forest, evergreen grassland and cultivated land increased by 10.62, 17.54, 9.9 and 14.77 percentage points respectively, which fully demonstrates that time series data can well distinguish vegetation features with phenological patterns and objectively reflect the differences between local features.
[0030] Wuning County: When using MLC combined with single-temporal imagery for feature classification, misclassification of bare land and built-up land, two features with similar spectral curves, was severe. Misclassification was also significant between evergreen forest and deciduous forest. Using time series data reduced misclassification of forest land because the NDVI temporal characteristics of evergreen and deciduous forests clearly differentiated. Time series data also reduced misclassification of bare land and built-up land because their TCW temporal characteristics were more distinct. Therefore, using time series data yielded better classification results. Comparing Figures 6-2(c) and (d), SVM combined with single-temporal imagery showed significant misclassification between evergreen and deciduous forests, while the classification results for cultivated land and grassland were patchy. SVM combined with time series data effectively addressed these issues, achieving good classification results for all features. Compared to maximum likelihood, support vector machines (SVMs) were able to more accurately distinguish between various features, most notably between evergreen and deciduous forests. Therefore, the method of combining time series data with SVM is more suitable for land cover type classification in this study area.
[0031] Comparing the study area of Sheli Town in Northeast China with a temperate monsoon climate and the study area of Wuning County in Jiangxi with a subtropical monsoon climate, the classification results of land cover types using time series data are better, indicating that time series data can be used to classify land cover types in various parts of the country and obtain good results.
[0032] Figure 2 Schematic diagram of a land cover classification device based on Sentinel-2AB time series data provided in an embodiment of the present application. The device includes: a selection module 201, a first analysis module 202, a second analysis module 203, a selection module 204, a classification module 205, and an accuracy analysis module 206.
[0033] Specifically, the selection module 201 is used to select a target research area.
[0034] The first analysis module 202 is used to analyze the time series characteristics of the Normalized Difference Vegetation Index of typical landforms in the target study area.
[0035] The second analysis module 203 is used to analyze the time series characteristics of the moisture component of the tasseled cap changes of typical landforms in the target research area.
[0036] The selection module 204 is used to select the best time series features.
[0037] The classification module 205 is used to perform land cover classification using support vector machine method and maximum likelihood method.
[0038] The accuracy analysis module 206 is used to perform accuracy analysis on the land cover classification.
[0039] See also Figure 3 This embodiment also provides a computer device, the components of which may include but are not limited to: one or more processors or processing units, a system memory, and a bus connecting different system components (including the system memory and the processing unit).
[0040] The term "bus" refers to one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus architectures. Examples of these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.
[0041] The computer system / server typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computer system / server, including volatile and non-volatile media, removable and non-removable media.
[0042] The system memory may include computer system readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory. The computer device may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system may be used to read and write non-removable, non-volatile magnetic media. A disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") may be provided, as well as an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media). In these cases, each drive may be connected to the bus via one or more data medium interfaces. The memory may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of various embodiments of the present invention.
[0043] A program / utility having a set (at least one) of program modules, which may be stored, for example, in a memory, includes, but is not limited to, an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may include an implementation of a network environment. The program modules generally perform the functions and / or methods of the embodiments described herein.
[0044] A computer device may also communicate with one or more external devices, such as a keyboard, pointing device, display, etc. Such communication may be performed via an input / output (I / O) interface. Furthermore, a computer device may also communicate with one or more networks, such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet, via a network adapter.
[0045] The processing unit executes the functions and / or methods described in the embodiments of the present invention by running the programs stored in the system memory.
[0046] The above-mentioned computer program can be set in a computer storage medium, that is, the computer storage medium is encoded with a computer program, and when the program is executed by one or more computers, it enables one or more computers to perform the method flow and / or device operation shown in the above-mentioned embodiments of the present invention.
[0047] As time goes by and technology develops, the meaning of medium becomes more and more extensive. The dissemination path of computer programs is no longer limited to tangible media, and can also be downloaded directly from the Internet. Any combination of one or more computer-readable media can be used. Computer-readable media can be computer-readable signal media or computer-readable storage media. Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or components, or any combination of the above. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device or device.
[0048] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including, but not limited to, electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.
[0049] Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
[0050] Computer program code for performing the operations of the present invention may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).
[0051] In addition to the above embodiments, the present invention may also have other implementation methods; any technical solutions formed by equivalent replacement or equivalent transformation fall within the protection scope required by the present invention.
Claims
1. A land cover classification method based on Sentinel-2AB time series data, characterized by: include: Select the target research area; Analyze the time series characteristics of the normalized difference vegetation index of typical features in the target study area; Analyze the time series characteristics of the moisture component of the tasseled cap of typical landforms in the target study area; Select the best time series features; Support vector machine and maximum likelihood methods were used for land cover classification.
2. The land cover classification method based on Sentinel-2AB time series data according to claim 1, characterized in that: The time series characteristics of the normalized difference vegetation index of typical features in the target research area include: The normalized vegetation index is calculated using formula 1: NDVI = (ρNIR-ρR) / (ρNIR+ρR), where ρNIR and ρR represent the reflectance values of the near-infrared band and the red light band, respectively.
3. The land cover classification method based on Sentinel-2AB time series data according to claim 1, characterized in that: The time series characteristics of the moisture component of the tasseled cap of typical landforms in the target study area include: The humidity component of the tasseled cap change is calculated using Formula 2: TCW = 0.1509ρ2 + 0.1973ρ3 + 0.3279ρ4 + 0.3406ρ8 + 0.7112ρ11 + 0.4572ρ12, where ρ2, ρ3, ρ4, ρ8, ρ11, and ρ12 are the reflectances of bands 2, 3, 4, 8, 11, and 12 in the Sentinel-2 image, respectively.
4. The land cover classification method based on Sentinel-2AB time series data according to claim 1, characterized in that: After the land cover classification is performed using the support vector machine method and the maximum likelihood method, the method further includes: Perform accuracy analysis on land cover classification.
5. The land cover classification method based on Sentinel-2AB time series data according to claim 4, characterized in that: The accuracy analysis of land cover classification includes: Google Earth data was used to conduct accuracy analysis of land cover classification.
6. A land cover classification device based on Sentinel-2AB time series data, characterized by: include: Selection module, used to select the target research area; The first analysis module is used to analyze the time series characteristics of the normalized difference vegetation index of typical features in the target study area; The second analysis module is used to analyze the time series characteristics of the moisture component of the tasseled cap changes of typical landforms in the target study area; Selection module, used to select the best time series features; Classification module for land cover classification using support vector machine and maximum likelihood methods.
7. The land cover classification method based on Sentinel-2AB time series data according to claim 6, characterized in that: Also includes: The accuracy analysis module is used to perform accuracy analysis on land cover classification.
8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method according to any one of claims 1 to 5 is implemented.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.