Remote sensing data identification method, device, electronic device and readable storage medium

Abstract

This application discloses a remote sensing data recognition method, apparatus, electronic device, and readable storage medium. The method includes obtaining target sample data from the remote sensing image data to be recognized in response to a user's sample selection instruction on a visualization page; calculating the spectral distribution characteristics of the target sample data; describing the spectral distribution characteristics using a preset functional relationship, and fitting the preset functional relationship based on the spectral distribution characteristics; calculating a classification threshold corresponding to the target sample data based on the fitting result; and identifying the remote sensing image data to be recognized based on the relationship between the spectral distribution characteristics and the classification threshold. This application can effectively improve the recognition accuracy of remote sensing image data.

Description

Technical Field

[0001] The present application relates to the field of remote sensing technology, and in particular to a remote sensing data identification method, device, electronic device and readable storage medium. Background Art

[0002] Remote sensing technology is widely used for monitoring tasks in agriculture, forestry, animal husbandry, and other fields due to its long-range, large-area detection capabilities and the lack of direct contact. Spectral data, rich in information about ground features, has become a primary data source for remote sensing identification. In spectral remote sensing identification, the spectral index is the primary classification feature, and calculating the spectral index threshold is a crucial issue in determining classification accuracy.

[0003] Related technologies can calculate the spectral index threshold through methods such as Otsu (maximum inter-class variance method) threshold, adaptive threshold, maximum entropy threshold, iterative threshold, etc. However, due to the complexity of remote sensing image scenes and the variety of ground objects, the spectral index threshold calculated by the above methods is difficult to meet the actual accuracy requirements.

[0004] In view of this, how to improve the recognition accuracy of remote sensing image data is a technical problem that technical personnel in the field need to solve. Summary of the Invention

[0005] The present application provides a remote sensing data recognition method, device, electronic device and readable storage medium, which effectively improve the recognition accuracy of remote sensing image data.

[0006] To solve the above technical problems, the embodiments of the present invention provide the following technical solutions:

[0007] An embodiment of the present invention provides a remote sensing data recognition method, comprising:

[0008] Responding to the user's sample selection instruction on the visualization page, obtaining target sample data from the remote sensing image data to be identified;

[0009] Calculating the spectral distribution characteristics of the target sample data;

[0010] Describing the spectral distribution characteristics using a preset functional relationship, and performing fitting processing on the preset functional relationship based on the spectral distribution characteristics;

[0011] Calculate the classification threshold corresponding to the target sample data according to the fitting result;

[0012] The remote sensing image data to be identified is identified according to the relationship between the spectral distribution characteristics and the classification threshold.

[0013] Optionally, before responding to the user's sample selection instruction on the visualization page, the method further includes:

[0014] Obtain original remote sensing image data;

[0015] Performing registration operations, splicing operations, and radiometric calibration operations on the original remote sensing image data to generate spliced ​​reflectance orthophoto data;

[0016] The reflectivity orthophoto data is displayed on a visualization page as the remote sensing image data to be identified.

[0017] Optionally, the step of responding to a sample selection instruction from a user on a visualization page and obtaining target sample data from the remote sensing image data to be identified includes:

[0018] In response to the user's layer selection instruction, the target layer obtained from each raster layer is displayed to the user;

[0019] In response to a user's survey area selection instruction, a target survey area is obtained from the target layer;

[0020] In response to a sample selection instruction from a user, target sample data is acquired from the target survey area, where each sample in the target sample data belongs to the same ground feature type and is the ground feature type identified by the user.

[0021] Optionally, calculating the spectral distribution characteristics of the target sample data includes:

[0022] Storing each sample of the target sample data in a corresponding sample grid, wherein the number of bands contained in each sample grid is the same as the number of bands contained in the remote sensing image data to be identified;

[0023] The characteristic index of each sample grid is calculated, and a sample histogram is generated according to the characteristic index of each sample grid to serve as the spectral distribution characteristics of the target sample data.

[0024] Optionally, calculating the classification threshold corresponding to the target sample data according to the fitting result includes:

[0025] Calculating a correlation factor according to the fitting result and the preset functional relationship;

[0026] If the correlation factor is greater than or equal to a preset correlation threshold, the target sample data is determined to be valid, and a classification threshold corresponding to the target sample data is calculated according to the fitting result;

[0027] If the correlation factor is less than a preset correlation threshold, the target sample data is determined to be invalid.

[0028] Optionally, the using a preset functional relationship to describe the spectral distribution characteristics, and performing fitting processing on the preset functional relationship based on the spectral distribution characteristics, includes:

[0029] The spectral distribution characteristics are described using a Gaussian function relationship; the spectral distribution characteristics are represented by a sample histogram, the independent variable of the Gaussian function relationship corresponds to the abscissa of the sample histogram, and the dependent variable of the Gaussian function relationship corresponds to the ordinate of the sample histogram;

[0030] Based on the maximum value of the ordinate in the sample histogram, performing a function transformation on the Gaussian function relationship according to a preset function transformation method to obtain a transformation relationship;

[0031] Obtaining parameters of the Gaussian function relationship by calculating the transformation relationship according to the sample histogram;

[0032] A fitted Gaussian function relationship is obtained according to the parameters of the Gaussian function relationship and the abscissa data of the sample histogram.

[0033] Optionally, calculating the classification threshold corresponding to the target sample data according to the fitting result includes:

[0034] Based on the parameters of the Gaussian function relationship, calling the classification threshold calculation relationship to calculate the classification threshold, the classification threshold calculation relationship is Thr_NDVI=q*min(μ-σ) and / or Thr_NDVI=q*max(μ+σ);

[0035] Wherein, Thr_NDVI is the classification threshold, μ is the mean of the Gaussian function relationship, σ is the variance of the Gaussian function relationship, and q is the adjustment factor, which is determined according to the recognition target and the recognition scene.

[0036] Another embodiment of the present invention provides a remote sensing data recognition device, comprising:

[0037] The target sample acquisition module is used to respond to the user's sample selection instruction on the visualization page and obtain target sample data from the remote sensing image data to be identified;

[0038] A feature calculation module, used to calculate the spectral distribution characteristics of the target sample data;

[0039] A fitting module, configured to describe the spectral distribution characteristics using a preset functional relationship, and perform fitting processing on the preset functional relationship based on the spectral distribution characteristics;

[0040] A threshold calculation module is used to calculate the classification threshold corresponding to the target sample data according to the fitting result;

[0041] The identification module is used to identify the remote sensing image data to be identified based on the relationship between the spectral distribution characteristics and the classification threshold.

[0042] An embodiment of the present invention further provides an electronic device, comprising a processor, wherein the processor is configured to implement the steps of the remote sensing data recognition method as described in any of the preceding items when executing a computer program stored in a memory.

[0043] Finally, an embodiment of the present invention further provides a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the remote sensing data recognition method as described in any of the above items are implemented.

[0044] The advantage of the technical solution provided by this application is that users can independently select samples on the visualization page and obtain the spectrum and environmental information of any target. It is universal for different remote sensing data and application scenarios, and is more suitable for engineering operations. It can ensure the accuracy of identification or inversion in engineering operations with large-scale and complex scenes. Selecting sample data according to one's own needs does not only not require any professional knowledge of the user, but is also simple to operate and can effectively eliminate targets with similar spectra, which is conducive to improving the accuracy of remote sensing data recognition. By calculating and analyzing a small amount of sample data to determine the classification threshold, the process-based operation of the entire remote sensing data recognition can greatly save users time in data conversion and function selection, not only improving the recognition accuracy of remote sensing data, but also increasing the calculation speed of the threshold while ensuring data reliability, improving the efficiency of the entire remote sensing data recognition, and improving the user experience.

[0045] In addition, the embodiments of the present invention also provide corresponding implementation devices, electronic devices and readable storage media for the remote sensing data recognition method, further making the method more practical, and the devices, electronic devices and readable storage media have corresponding advantages.

[0046] It should be understood that the foregoing general description and the following detailed description are exemplary only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention or related technologies, the following briefly introduces the drawings required for use in the embodiments or related technical descriptions. 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 work.

[0048] Figure 1 A flowchart of a remote sensing data recognition method provided by an embodiment of the present invention;

[0049] Figure 2 A schematic diagram of a flow chart of another remote sensing data recognition method provided by an embodiment of the present invention;

[0050] Figure 3 A structural diagram of a specific implementation of a remote sensing data recognition device provided by an embodiment of the present invention;

[0051] Figure 4 This is a structural diagram of a specific implementation of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0052] In order to enable those skilled in the art to better understand the present invention, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0053] In the specification, claims, and drawings of this application, the terms "first," "second," "third," "fourth," and so on are used to distinguish between different items, not to describe a specific order. Furthermore, the terms "including," "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or elements is not limited to the listed steps or elements and may include steps or elements that are not listed.

[0054] After investigation, the inventors of this application found that the threshold calculation process of a related technology is as follows: first, a series of large-scale grayscale images are read in matrix form, and Wiener filtering and Gaussian filtering are performed on the images that meet the requirements in sequence to obtain image samples; then, the grayscale value information of the image samples is statistically analyzed to obtain the grayscale histogram of the image samples, and the grayscale images of the series of images are read and the histograms are accumulated; then, the gradient is calculated based on the accumulated histogram to obtain different gradient distribution data of the image samples; based on the grayscale histogram of the image samples and their gradient distribution data results, the grayscale distribution eigenvalue is solved, and the segmentation threshold is calculated by an empirical formula; finally, the adjustable range of the segmentation threshold is calculated, and a range limit for manual adjustment is provided. This method is suitable for threshold calculation of scenes such as single sample images, but the scenes in remote sensing images are changeable, and it is difficult to simply take fixed gradient feature points and define empirical formulas based on the grayscale image.

[0055] Another related technique involves calculating threshold values ​​for remote sensing classification of aquatic vegetation. This method ranks the spectral indices of pixels within a defined area of ​​two images of the same feature, then fits a linear model of the spectral indices. Substituting the classification thresholds for the spectral indices of images with synchronized field measurements into the linear model, the classification thresholds for the spectral indices of images without synchronized field measurements are calculated. This method fits hyperspectral curves and uses a linear fitting method to calculate the threshold for the second characteristic index of principal component analysis, combined with empirically measured thresholds. This method only classifies aquatic plants in a specific area and does not fully encompass the diverse features of remote sensing imagery used in real-world applications, making it less universal. Furthermore, the threshold relationship determined by linear fitting relies heavily on empirical data, requiring extensive data support. This precondition makes it difficult to apply to scenarios where data availability is limited.

[0056] In view of this, this application realizes semi-automatic threshold calculation by selecting positive samples, histogram analysis and combining empirical formulas, which can meet the high-precision real-world needs in various application scenarios such as vegetation, water bodies, and artificial building extraction.

[0057] After introducing the technical solutions of the embodiments of the present invention, various non-limiting implementation methods of the present application are described in detail below.

[0058] See first Figure 1 , Figure 1 A flowchart of a remote sensing data recognition method provided in an embodiment of the present invention may include the following:

[0059] S101: Responding to a sample selection instruction from a user on a visualization page, obtaining target sample data from remote sensing image data to be identified.

[0060] In this embodiment, the visualization page is presented to the user by the system for displaying the remote sensing image data to be identified and providing a human-computer interaction interface. The user can make selections on the display through KVM (Keyboard Video Mouse) or directly through the touch screen, which does not affect the implementation of this application. The user selects the sample data that the system needs to identify on the remote sensing image data to be identified according to his or her own needs. After detecting the user's sample selection instruction, the system responds to the instruction and extracts the sample data selected by the user from the remote sensing data to be identified as the target sample data for executing subsequent steps. The remote sensing image data to be identified in this step can be the original remote sensing image data or the remote sensing image data after a series of data processing, which does not affect the implementation of this application. The target sample data may include multiple samples, which can be multiple samples from the same area or multiple samples from different areas. This application does not impose any restrictions on this.

[0061] S102: Calculate the spectral distribution characteristics of the target sample data.

[0062] The spectral distribution characteristics of this embodiment reflect the category to which the sample data belongs. Persons skilled in the art can flexibly adopt corresponding feature calculation methods based on the type of ground feature to which the target sample belongs. For example, taking vegetation as an example, the spectral distribution characteristics can be obtained by statistically analyzing the sample histogram based on the normalized vegetation index of each sample. Taking water bodies as an example, the spectral distribution characteristics can be obtained by statistically analyzing the sample histogram based on the water body characteristic index of each sample. The water body characteristic index may include, but is not limited to, a normalized water body index. Persons skilled in the art can add any index that can characterize water body characteristics based on actual needs, and this will not affect the implementation of this application.

[0063] S103: Using a preset functional relationship to describe the spectrum distribution characteristics, and performing fitting processing on the preset functional relationship based on the spectrum distribution characteristics.

[0064] The preset functional relationship of this embodiment can be any mathematical relationship that can be used to describe the spectral distribution characteristics of the target sample data, such as a Gaussian function relationship or other binary linear exponential function relationship, which does not affect the implementation of this application. It is understood that no functional relationship can completely match the numerical relationship corresponding to the spectral distribution characteristics. Therefore, after determining the functional relationship used to describe the spectral distribution characteristics, this step needs to use the spectral distribution characteristics to perform fitting processing to obtain a fitting functional relationship that is most relevant to the spectral distribution characteristics.

[0065] S104: Calculate the classification threshold corresponding to the target sample data according to the fitting result.

[0066] After fitting in the previous step, a fitting function relationship corresponding to the preset function relationship is obtained. When the independent variables of the fitting function relationship and the preset function relationship are the same, there is a difference in the dependent variable. The smaller the difference between the two, the more correlated the spectral distribution characteristics are with the preset function relationship. The larger the difference between the two, the less correlated the spectral distribution characteristics are with the preset function relationship. The greater the correlation between the two, the more valid the target sample data is. In this step, a classification threshold can be calculated based on the fitting function relationship to identify whether the target sample data is the target category the user wants to identify, and the target sample data can be identified and classified based on this classification threshold.

[0067] S105: Identify the remote sensing image data to be identified based on the relationship between the spectral distribution characteristics and the classification threshold.

[0068] After determining the classification threshold in the previous step, since the spectral distribution characteristics reflect the ground feature characteristics of the target sample data, and the classification threshold is used to identify whether it is the standard for the category of the ground feature target that the user wants to identify, the target sample data can be classified and identified based on the relationship between the classification threshold and the spectral classification characteristics.

[0069] In the technical solution provided by the embodiments of the present invention, users can independently select samples on the visualization page to obtain spectral and environmental information for any target. This solution is not only universally applicable to different remote sensing data and application scenarios, but also allows users to select sample data based on their needs, effectively eliminating targets with similar spectra, thereby improving the accuracy of remote sensing data recognition. By calculating and analyzing a small amount of sample data to determine the classification threshold, not only can the recognition accuracy of remote sensing data be improved, but the threshold calculation speed can also be increased while ensuring data reliability, thereby improving the overall efficiency of remote sensing data recognition.

[0070] In order to further improve the recognition accuracy of remote sensing data, before responding to the user's sample selection instruction on the visualization page, a series of data processing can be performed on the original remote sensing image data. The remote sensing data obtained after the data processing is used as the remote sensing image data to be recognized in the subsequent steps, which may include the following steps:

[0071] Obtain original remote sensing image data;

[0072] Perform registration, stitching and radiometric calibration operations on the original remote sensing image data to generate stitched reflectance orthophoto data;

[0073] The reflectance orthophoto data is displayed on the visualization page as the remote sensing image data to be identified.

[0074] In this embodiment, configuration, splicing and radiometric calibration are all conventional means in image processing. Those skilled in the art may refer to any method and step in the relevant art for implementing configuration, splicing and radiometric calibration to perform the second step of this embodiment, and this application does not impose any limitation on this. Since reflectivity can provide feedback on the real information of the ground object, and non-orthophoto data such as oblique data will have shadows, which is not conducive to subsequent sample identification, the remote sensing image data used in the embodiment is reflectivity orthophoto image data, which includes multiple bands, and its pixel value is the spectral reflectivity of the ground object after radiation correction. Of course, those skilled in the art may also use other types of reflectivity image data, which does not affect the implementation of this application.

[0075] In the above embodiment, there is no limitation on how to perform step S101. In this embodiment, a method for generating target sample data is provided, which may include the following steps:

[0076] In response to the user's layer selection instruction, the target layer obtained from each raster layer is displayed to the user; in response to the user's survey area range selection instruction, the target survey area is obtained from the target layer; in response to the user's sample selection instruction, the target sample data is obtained from the target survey area.

[0077] In this embodiment, after obtaining the remote sensing data to be identified, the user determines the type of land object of the target positive sample. The so-called positive sample refers to the category of the land object that the user wants to identify. In order to improve the recognition accuracy, each sample in the target sample data selected by the user belongs to the same land object type and is the land object type that the user needs to identify.

[0078] In this embodiment, the user only needs to select typical positive samples and does not need to draw multiple categories of negative samples; while improving the automation of the entire remote sensing data recognition, the flexibility of operation is also retained, and the user can adjust the process parameters and edit the results.

[0079] The above embodiment does not impose any limitation on how to calculate the spectral distribution characteristics of the target sample data. The present application also provides an optional method for calculating the spectral distribution characteristics of the target sample data, which may include:

[0080] Each sample of the target sample data is stored in a corresponding sample grid. The number of bands contained in each sample grid is the same as the number of bands contained in the remote sensing image data to be identified. The characteristic index of each sample grid is calculated, and a sample histogram is generated based on the characteristic index of each sample grid to serve as the spectral distribution characteristics of the target sample data.

[0081] The sample data format of this embodiment is a grid format. The threshold is determined by computing and analyzing a small amount of sample grid data, thereby improving the calculation speed of the threshold while ensuring data reliability.

[0082] Based on the above embodiment, in order to improve the recognition rate and accuracy of remote sensing data, this embodiment may also perform sample validity screening before calculating the classification threshold, which may include the following:

[0083] The correlation factor is calculated based on the fitting result and the preset functional relationship; if the correlation factor is greater than or equal to the preset correlation threshold, the target sample data is determined to be valid, and then the subsequent steps of calculating the classification threshold corresponding to the target sample data based on the fitting result are performed; if the correlation factor is less than the preset correlation threshold, the target sample data is determined to be invalid, and the sample data will not be processed, and the data can be directly discarded or an error will be reported.

[0084] The above embodiment does not limit S103 in any way. This application also provides an optional implementation of this step, which may include:

[0085] The spectral distribution characteristics are described using a Gaussian function relationship. The spectral distribution characteristics are represented by a sample histogram. The independent variable of the Gaussian function relationship corresponds to the abscissa of the sample histogram, and the dependent variable of the Gaussian function relationship corresponds to the ordinate of the sample histogram.

[0086] Based on the maximum value of the ordinate in the sample histogram, the Gaussian function relationship is transformed according to a preset function transformation method to obtain a transformation relationship;

[0087] According to the sample histogram, the parameters of the Gaussian function relationship are obtained by calculating the transformation relationship;

[0088] The fitted Gaussian function relationship is obtained according to the parameters of the Gaussian function relationship and the horizontal coordinate data of the sample histogram.

[0089] The parameters of the Gaussian function relationship in this embodiment are the mean and variance of the Gaussian function relationship. Accordingly, the implementation process of calculating the classification threshold corresponding to the target sample data based on the fitting result can be:

[0090] Based on the parameters of the Gaussian function relationship, the classification threshold calculation relationship is called to calculate the classification threshold. The classification threshold calculation relationship is Thr_NDVI = q*min(μ-σ) and / or Thr_NDVI = q*max(μ+σ); where Thr_NDVI is the classification threshold, μ is the mean of the Gaussian function relationship, σ is the variance of the Gaussian function relationship, and q is the adjustment factor, which is determined according to the recognition target and recognition scene.

[0091] Among them, when calculating the classification threshold, only Thr_NDVI=q*min(μ-σ) can be used for calculation, or only Thr_NDVI=q*max(μ+σ) can be used for calculation, or Thr_NDVI=q*min(μ-σ) and Thr_NDVI=q*max(μ+σ) can be used for calculation at the same time. Which calculation relationship to use can be determined according to the actual application scenario, and this application does not impose any restrictions on this.

[0092] In order to make the technical personnel in the related field more clearly understand the technical solution of this application, this application is also combined with Figure 2 An illustrative example is provided. In this embodiment, the computer program corresponding to the remote sensing data recognition method is encapsulated as a functional module, and combined with a user interface UI, etc. to generate remote sensing data recognition software, which is referred to as software in the following embodiments and may include the following contents:

[0093] A1: The software's data import module reads the stitched reflectance orthophoto remote sensing image data and displays it in the software's data view window. Users can perform basic operations on the displayed image, including but not limited to data browsing, zooming in and out, zooming to a layer, and setting layer display mode.

[0094] A2: The data import module automatically reads all raster layers of the reflectance orthophoto remote sensing image data opened in the software, and the user selects a layer.

[0095] A3: After selecting a layer, users can draw rectangles or polygons in the data view to precisely limit the measurement area. Users can also select the entire image as the measurement area and directly enter the sample data analysis module.

[0096] A4: After entering the sample analysis module, users can select samples by drawing multiple rectangles or polygons in the data view. Here, only the target positive sample range is required. The software will calibrate the algorithm model based solely on the positive sample information. After clicking "Next," the software automatically performs sample analysis.

[0097] A5: Calculate the spectral distribution characteristics of the sample and perform Gaussian fitting. Calculate the correlation between the distribution characteristics and the Gaussian function, and use the correlation to determine whether the sample is valid. If the sample is valid, calculate the classification threshold. If the sample is invalid, prompt the user to reselect a sample.

[0098] Taking vegetation as an example, this embodiment only needs to select positive samples, without selecting a large number of negative samples. Select vegetation samples at M (M>=1) and store M sample grids, each of which includes N bands. For each sample grid, calculate the NDVI (Normalized Difference Vegetation Index) index according to the following relationship 1:

[0099]

[0100] Where red represents the red band reflectance; nir represents the near-infrared band reflectance.

[0101] The NDVI of each sample was normalized to [0, 255] based on the following equation 2:

[0102]

[0103] Among them, min and max are the minimum and maximum values ​​of NDVI of each sample, respectively.

[0104] Each NDVInor is rounded to an integer, and then a histogram is generated. The X value of the histogram is NDVInor, where X is the set of x, X is the value of the characteristic index, and each y value of the ordinate Y is the number of NDVInor = x in the sample, where Y is the set of y, and Y is the total number of grids at a certain value x.

[0105] The preset function relationship defined in this embodiment is a Gaussian function relationship, that is:

[0106]

[0107] Where x is the abscissa of the sample index histogram; y is the ordinate of the sample index histogram; a, b, c are the parameters to be determined. Based on the maximum value of the ordinate of the histogram, mathematically transform the above relationship 3 to obtain Y nm =y / y max ;y max is the maximum value of y, Y nm is the intermediate parameter. nm Perform median filtering and mean filtering successively to obtain the filtering result. Normalize the y in the filtering result. Then normalize the y obtained above. max Perform mathematical transformation according to the following equation 4 to obtain another intermediate parameter Z:

[0108] Z=ln(Y nm )=b0+b1x+b2x 2 ; (4)

[0109] Where,

[0110] Taking into account all sample point data, the above equation 4 is further transformed into:

[0111]

[0112] According to the vertical coordinate data of the histogram, each Z, namely Z1…Z n Value, when the horizontal axis data of the histogram and Z1…Z n Substituting into equation 5, we can calculate b0, b1, and b2. The parameters a, b, and c of the Gaussian function equation can be calculated using equations 6, 7, and 8 below:

[0113]

[0114]

[0115]

[0116] After calculating the parameters a, b and c, This is the fitted Gaussian function relationship. After obtaining the fitted Gaussian function relationship, the correlation factor R can be calculated based on the ordinate data of the histogram and the y value obtained from the fitted Gaussian function relationship. 2 :

[0117]

[0118] Where i∈[0, 255], y is the vertical coordinate of the sample index histogram after normalization, y μ is the mean of y, which is calculated by summing the numerator with y and the denominator with 255.

[0119] The classification threshold Thr_NDVI can be calculated according to the following calculation formula 10:

[0120] Thr_NDVI=0.9*(bc); (10)

[0121] When the NDVI of a pixel in an image is greater than Thr_NDVI, it is considered vegetation. When the NDVI of a pixel in an image is less than or equal to Thr_NDVI, it is not vegetation. Equation 10 is an empirical formula and can be adjusted based on the specific recognition scenario and recognition target.

[0122] A6: When the user clicks the "Preview" button, the software automatically acquires image information, invokes the corresponding algorithm to calculate feature spaces such as indices, and extracts targets using the classification thresholds calculated by the sample data analysis module. The remote sensing recognition results are then overlaid on the data view. The user reviews the remote sensing recognition results. If they meet the user's requirements, the results are directly output. If not, the user can manually adjust the parameters and re-classify the recognition.

[0123] A7: The types of outputs from the cost output module of the calling software vary depending on the specific remote sensing application requirements. The main output forms may include: vector data, raster data, thematic maps, reports, etc.

[0124] As can be seen from the above, the process-based operation provided by this embodiment can significantly save users time in data conversion and function selection. The internal algorithm has been simplified, resulting in excellent computational speed. The algorithm's cohesion is improved by optimizing the data format. The semi-open threshold setting and sample selection mechanism greatly enhance the universality and robustness of the remote sensing data recognition method. By discriminating image file information, the corresponding algorithm is automatically invoked. This ensures recognition or inversion accuracy in engineering operations across large areas and complex scenarios. It can meet the needs of the main body of industrial application processing, providing a rich set of subject recognition and auxiliary operation functions. Furthermore, the subject recognition algorithm is encapsulated as a dynamic library, and the software automatically outputs results. The user-friendly interface and integrated process-based operation eliminate the need for specialized knowledge, lowering the user experience. Users only need to select typical positive samples, eliminating the need to draw multiple negative samples. While improving software automation, operational flexibility is retained, allowing users to adjust process parameters and edit results. The operation is simple and provides a good user experience.

[0125] It should be noted that there is no strict order in which the steps in this application are performed. As long as they comply with the logical order, these steps can be performed simultaneously or in a predetermined order. Figure 1 and Figure 2 This is just a schematic and does not mean that this is the only execution order.

[0126] The embodiments of the present invention also provide a corresponding device for the remote sensing data recognition method, further enhancing the practicality of the method. The device can be described from the perspective of functional modules and hardware. The remote sensing data recognition device provided by the embodiments of the present invention is described below. The remote sensing data recognition device described below can be used in conjunction with the remote sensing data recognition method described above.

[0127] From the perspective of functional modules, see Figure 3 , Figure 3 A structural diagram of a remote sensing data recognition device provided in an embodiment of the present invention in a specific implementation manner, the device may include:

[0128] The target sample acquisition module 301 is used to respond to the sample selection instruction of the user on the visualization page and acquire target sample data from the remote sensing image data to be identified.

[0129] The feature calculation module 302 is used to calculate the spectral distribution features of the target sample data.

[0130] The fitting module 303 is configured to describe the spectrum distribution characteristics using a preset functional relationship and perform fitting processing on the preset functional relationship based on the spectrum distribution characteristics.

[0131] The threshold calculation module 304 is used to calculate the classification threshold corresponding to the target sample data according to the fitting result.

[0132] The identification module 305 is used to identify the remote sensing image data to be identified based on the relationship between the spectral distribution characteristics and the classification threshold.

[0133] Optionally, in some implementations of this embodiment, the above-mentioned device may further include a data processing module for obtaining original remote sensing image data before responding to the user's sample selection instruction on the visualization page; performing registration operations, stitching operations and radiation calibration operations on the original remote sensing image data to generate stitched reflectance orthophoto data; and displaying the reflectance orthophoto image data as remote sensing image data to be identified on the visualization page.

[0134] As an optional implementation of this embodiment, the above-mentioned target sample acquisition module 301 can be further used to: respond to the user's layer selection instruction, display the target layer obtained from each raster layer to the user; respond to the user's survey area range selection instruction, obtain the target survey area from the target layer; respond to the user's sample selection instruction, obtain target sample data from the target survey area, and each sample in the target sample data belongs to the same ground feature type and is the ground feature type identified by the user's needs.

[0135] As another optional implementation of this embodiment, the above-mentioned feature calculation module 302 can be further used to: store samples of the target sample data at various locations in corresponding sample grids, and the number of bands contained in each sample grid is the same as the number of bands contained in the remote sensing image data to be identified; calculate the characteristic index of each sample grid, and generate a sample histogram based on the characteristic index of each sample grid as the spectral distribution feature of the target sample data.

[0136] Optionally, in some implementations of this embodiment, the above-mentioned device may also include a validity identification module, for example, for calculating a correlation factor based on the fitting results and a preset functional relationship; if the correlation factor is greater than or equal to a preset correlation threshold, the target sample data is determined to be valid; if the correlation factor is less than the preset correlation threshold, the target sample data is determined to be invalid.

[0137] As an optional implementation of this embodiment, the above-mentioned fitting module 303 can be further used to: use a Gaussian function relationship to describe the spectral distribution characteristics; the spectral distribution characteristics are represented by a sample histogram, the independent variable of the Gaussian function relationship corresponds to the horizontal coordinate of the sample histogram, and the dependent variable of the Gaussian function relationship corresponds to the vertical coordinate of the sample histogram; based on the maximum value of the vertical coordinate in the sample histogram, the Gaussian function relationship is transformed according to a preset function transformation method to obtain a transformation relationship; according to the sample histogram, the parameters of the Gaussian function relationship are obtained by calculating the transformation relationship; according to the parameters of the Gaussian function relationship and the horizontal coordinate data of the sample histogram, a fitted Gaussian function relationship is obtained.

[0138] As an optional implementation of this embodiment, the above-mentioned threshold calculation module 304 can be further used to: based on the parameters of the Gaussian function relationship, call the classification threshold calculation relationship to calculate the classification threshold, the classification threshold calculation relationship is Thr_NDVI = q*min(μ-σ) and / or Thr_NDVI = q*max(μ+σ); wherein, Thr_NDVI is the classification threshold, μ is the mean of the Gaussian function relationship, σ is the variance of the Gaussian function relationship, q is the adjustment factor, and the adjustment factor is determined according to the recognition target and the recognition scene.

[0139] The functions of the functional modules of the remote sensing data recognition device described in the embodiment of the present invention can be specifically implemented according to the method in the above method embodiment. The specific implementation process can refer to the relevant description of the above method embodiment and will not be repeated here.

[0140] It can be seen from the above that the embodiments of the present invention can effectively improve the recognition accuracy of remote sensing image data.

[0141] The remote sensing data recognition device mentioned above is described from the perspective of functional modules. Furthermore, the present application also provides an electronic device, which is described from the perspective of hardware. Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application in one embodiment. Figure 4 As shown, the electronic device includes a memory 40 for storing computer programs; a processor 41 for implementing the steps of the remote sensing data recognition method mentioned in any of the above embodiments when executing the computer program.

[0142] The processor 41 may include one or more processing cores, such as a quad-core processor or an octa-core processor. The processor 41 may also be a controller, a microcontroller, a microprocessor, or other data processing chip. The processor 41 may be implemented in at least one hardware form selected from the group consisting of a DSP (Digital Signal Processing), an FPGA (Field-Programmable Gate Array), and a PLA (Programmable Logic Array). The processor 41 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 41 may be integrated with a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 41 may also include an AI (Artificial Intelligence) processor, which is used to handle computing operations related to machine learning.

[0143] The memory 40 may include one or more computer-readable storage media, which may be non-transitory. The memory 40 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash memory storage devices. In some embodiments, the memory 40 may be an internal storage unit of the electronic device, such as a hard disk of a server. In other embodiments, the memory 40 may also be an external storage device of the electronic device, such as a plug-in hard disk equipped on a server, a smart media card (SMC), a secure digital (SD) card, a flash memory card, etc. Furthermore, the memory 40 may include both an internal storage unit of the electronic device and an external storage device. The memory 40 can be used not only to store application software installed in the electronic device and various types of data, such as the code of the program that executes the vulnerability handling method, but also to temporarily store data that has been output or is about to be output. In this embodiment, the memory 40 is used to store at least the following computer program 401, wherein, after being loaded and executed by the processor 41, the computer program can implement the relevant steps of the remote sensing data identification method disclosed in any of the aforementioned embodiments. In addition, the resources stored in memory 40 may also include an operating system 402 and data 403, which may be stored in a temporary or permanent manner. Operating system 402 may include Windows, Unix, Linux, etc. Data 403 may include, but is not limited to, data corresponding to remote sensing data recognition results.

[0144] In some embodiments, the electronic device may further include a display screen 42, an input / output interface 43, a communication interface 44 or a network interface, a power supply 45 and a communication bus 46. Among them, the display screen 42 and the input / output interface 43, such as a keyboard, are user interfaces, and the optional user interface may also include a standard wired interface, a wireless interface, etc. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode) touch device, etc. The display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device and to display a visual user interface. The communication interface 44 may optionally include a wired interface and / or a wireless interface, such as a WI-FI interface, a Bluetooth interface, etc., which is generally used to establish a communication connection between the electronic device and other electronic devices. The communication bus 46 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0145] Those skilled in the art will understand that Figure 4 The structure shown in the figure does not constitute a limitation on the electronic device, and may include more or fewer components than shown in the figure, for example, it may also include a sensor 47 to realize various functions.

[0146] The functions of the functional modules of the electronic device described in the embodiment of the present invention can be specifically implemented according to the method in the above method embodiment. The specific implementation process can refer to the relevant description of the above method embodiment and will not be repeated here.

[0147] It can be seen from the above that the embodiments of the present invention can effectively improve the recognition accuracy of remote sensing image data.

[0148] It is understandable that if the remote sensing data recognition method in the above embodiment 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 technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), an electrically erasable programmable ROM, a register, a hard disk, a multimedia card, a card-type memory (such as an SD or DX memory, etc.), a magnetic memory, a removable disk, a CD-ROM, a magnetic disk or an optical disk, and other media that can store program code.

[0149] Based on this, an embodiment of the present invention further provides a readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the remote sensing data recognition method described in any of the above embodiments.

[0150] The functions of the functional modules of the readable storage medium described in the embodiment of the present invention can be specifically implemented according to the method in the above method embodiment. The specific implementation process can refer to the relevant description of the above method embodiment and will not be repeated here.

[0151] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from the other embodiments. References to the same or similar parts between the various embodiments are sufficient. The hardware disclosed in the embodiments, including devices and electronic devices, is described briefly because it corresponds to the methods disclosed in the embodiments. For relevant details, refer to the method description.

[0152] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0153] The above is a detailed introduction to the remote sensing data recognition method, device, electronic device and readable storage medium provided by this application. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of the present invention, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the scope of protection of the claims of this application.

Claims

1. A remote sensing data recognition method, characterized in that: include: Responding to the user's sample selection instruction on the visualization page, obtaining target sample data from the remote sensing image data to be identified; Each sample in the target sample data belongs to the same ground feature type and is the ground feature type identified by the user; Calculating the spectral distribution characteristics of the target sample data; A Gaussian function relationship is used to describe spectral distribution characteristics; the spectral distribution characteristics are represented by a sample histogram, the independent variable of the Gaussian function relationship corresponds to the abscissa of the sample histogram, and the dependent variable of the Gaussian function relationship corresponds to the ordinate of the sample histogram; based on the maximum value of the ordinate in the sample histogram, the Gaussian function relationship is transformed according to a preset function transformation method to obtain a transformation relationship; according to the sample histogram, the parameters of the Gaussian function relationship are obtained by calculating the transformation relationship; and a fitted Gaussian function relationship is obtained according to the parameters of the Gaussian function relationship and the abscissa data of the sample histogram; Calculate a correlation factor based on the fitting result and a preset functional relationship; if the correlation factor is greater than or equal to a preset correlation threshold, determine that the target sample data is valid, and calculate a classification threshold corresponding to the target sample data based on the fitting result; if the correlation factor is less than the preset correlation threshold, determine that the target sample data is invalid; wherein, calculating the correlation factor based on the fitting result and the preset functional relationship is: calculating the correlation factor based on the ordinate data of the sample histogram and the ordinate value of the fitting Gaussian function relationship; When the sample is determined to be valid according to the correlation factor, the classification threshold is calculated by calling the classification threshold calculation formula according to the parameters of the Gaussian function relationship. The classification threshold calculation formula is Thr_NDVI=q*min(μ-σ) and / or Thr_NDVI=q*max(μ+σ); wherein Thr_NDVI is the classification threshold, μ is the mean of the Gaussian function relationship, σ is the variance of the Gaussian function relationship, and q is an adjustment factor, which is determined according to the recognition target and the recognition scene; The remote sensing image data to be identified is identified according to the relationship between the spectral distribution characteristics and the classification threshold.

2. The remote sensing data recognition method according to claim 1, characterized in that: Before responding to the user's sample selection instruction on the visualization page, the method further includes: Obtain original remote sensing image data; Performing registration operations, splicing operations, and radiometric calibration operations on the original remote sensing image data to generate spliced ​​reflectance orthophoto data; The reflectivity orthophoto data is displayed on a visualization page as the remote sensing image data to be identified.

3. The remote sensing data recognition method according to claim 1, characterized in that: The step of responding to a sample selection instruction from a user on a visualization page and obtaining target sample data from the remote sensing image data to be identified includes: In response to the user's layer selection instruction, the target layer obtained from each raster layer is displayed to the user; In response to a user's survey area selection instruction, a target survey area is obtained from the target layer; In response to a user's sample selection instruction, target sample data is acquired from the target measurement area.

4. The remote sensing data recognition method according to claim 1, characterized in that: The calculating the spectral distribution characteristics of the target sample data includes: Storing each sample of the target sample data in a corresponding sample grid, wherein the number of bands contained in each sample grid is the same as the number of bands contained in the remote sensing image data to be identified; The characteristic index of each sample grid is calculated, and a sample histogram is generated according to the characteristic index of each sample grid to serve as the spectral distribution characteristics of the target sample data.

5. A remote sensing data recognition device, characterized in that: include: The target sample acquisition module is used to respond to the user's sample selection instruction on the visualization page and obtain target sample data from the remote sensing image data to be identified; Each sample in the target sample data belongs to the same ground feature type and is the ground feature type identified by the user; A feature calculation module, used to calculate the spectral distribution characteristics of the target sample data; A fitting module, configured to describe spectral distribution characteristics using a Gaussian function relationship; the spectral distribution characteristics are represented by a sample histogram, the independent variable of the Gaussian function relationship corresponds to the abscissa of the sample histogram, and the dependent variable of the Gaussian function relationship corresponds to the ordinate of the sample histogram; based on the maximum value of the ordinate in the sample histogram, the Gaussian function relationship is transformed according to a preset function transformation method to obtain a transformation relationship; according to the sample histogram, the parameters of the Gaussian function relationship are obtained by calculating the transformation relationship; and a fitted Gaussian function relationship is obtained based on the parameters of the Gaussian function relationship and the abscissa data of the sample histogram; A threshold calculation module is used to calculate a correlation factor based on the fitting result and a preset function relationship; if the correlation factor is greater than or equal to the preset correlation threshold, the target sample data is determined to be valid, and the classification threshold corresponding to the target sample data is calculated based on the fitting result; if the correlation factor is less than the preset correlation threshold, the target sample data is determined to be invalid; the threshold calculation module is further used to: calculate the correlation factor based on the ordinate data of the sample histogram and the ordinate value of the fitted Gaussian function relationship, and when the sample is determined to be valid based on the correlation factor, call the classification threshold calculation relationship to calculate the classification threshold based on the parameters of the Gaussian function relationship, and the classification threshold calculation relationship is Thr_NDVI=q*min(μ-σ) and / or Thr_NDVI=q*max(μ+σ); wherein Thr_NDVI is the classification threshold, μ is the mean of the Gaussian function relationship, σ is the variance of the Gaussian function relationship, and q is an adjustment factor, which is determined according to the recognition target and the recognition scene; The identification module is used to identify the remote sensing image data to be identified based on the relationship between the spectral distribution characteristics and the classification threshold.

6. An electronic device, characterized in that: The system comprises a processor and a memory, wherein the processor is configured to implement the steps of the remote sensing data recognition method according to any one of claims 1 to 4 when executing a computer program stored in the memory.

7. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the remote sensing data recognition method according to any one of claims 1 to 4 are implemented.

Images