Satellite hyperspectral sample library construction method, device and equipment and readable storage medium

Abstract

The application provides a satellite hyperspectral sample library construction method, device and equipment and a readable storage medium, relates to the technical field of satellite hyperspectral sample libraries, and comprises the following steps: obtaining at least two pieces of satellite hyperspectral data and basic data corresponding to the satellite hyperspectral data; performing slope segmentation on the elevation data corresponding to each piece of satellite hyperspectral data to obtain at least two pieces of elevation sub-data; performing weathered vegetation factor zoning calculation based on each piece of satellite hyperspectral data to obtain a zoning result of the satellite hyperspectral data; and labeling each pixel in each piece of satellite hyperspectral data according to geological data, the elevation sub-data and the zoning result to obtain labeled satellite hyperspectral data, wherein each pixel point in the labeled satellite hyperspectral data comprises classification attributes in three dimensions of weathered vegetation, lithological types and slope. The samples of the application meet the advantages of accurate interpretation requirements of different levels of scenes, targets and pixels.

Description

Technical Field

[0001] This invention relates to the field of hyperspectral image processing technology, and more specifically, to a method, apparatus, device, and readable storage medium for constructing a satellite hyperspectral sample library. Background Technology

[0002] Current remote sensing interpretation sample libraries mainly consist of spectral curve sample sets and RGB (visible light) images. However, these sample sets suffer from inconsistencies in spectral data, inconsistent classification systems, and a lack of hyperspectral lithology sample libraries specifically for engineering geological rock assemblages. Existing deep learning samples are limited in quantity and type, and insufficiently reflect the multi-scale, multi-sensor, and multi-temporal characteristics of remote sensing imagery. Therefore, there is an urgent need for a satellite hyperspectral sample library for engineering geological rock assemblages designed for intelligent interpretation, in order to improve the accuracy of intelligent interpretation of engineering geological lithology. Summary of the Invention

[0003] The purpose of this invention is to provide a method, apparatus, device, and readable storage medium for constructing a satellite hyperspectral sample library, thereby improving the aforementioned problems. To achieve the above objective, the technical solution adopted by this invention is as follows:

[0004] In a first aspect, this application provides a method for constructing a satellite hyperspectral sample library, comprising: acquiring at least two sets of satellite hyperspectral data and corresponding basic data, wherein the basic data includes elevation data and geological data; performing slope segmentation based on the elevation data corresponding to each set of satellite hyperspectral data to obtain at least two sets of elevation sub-data, wherein each set of elevation sub-data contains a different slope; performing weathering and vegetation factor partitioning calculation based on each set of satellite hyperspectral data to obtain partitioning results of the satellite hyperspectral data; labeling each pixel in each set of satellite hyperspectral data according to the geological data, elevation sub-data, and partitioning results to obtain labeled satellite hyperspectral data, wherein each pixel in the labeled satellite hyperspectral data includes classification attributes in three dimensions: weathering and vegetation, lithology type, and slope; sequentially performing sample cropping, merging of samples with the same attributes, and sample encoding on the labeled satellite hyperspectral data, and using the sample-encoded satellite hyperspectral data as a sample in the satellite hyperspectral sample library.

[0005] Secondly, this application also provides a satellite hyperspectral sample library construction device, comprising: a data acquisition unit for acquiring at least two sets of satellite hyperspectral data and corresponding basic data, the basic data including elevation data and geological data; a segmentation calculation unit for performing slope segmentation based on the elevation data corresponding to each set of satellite hyperspectral data to obtain at least two sets of elevation sub-data, each set of elevation sub-data containing a different slope; a partitioning calculation unit for performing weathering and vegetation factor partitioning calculation based on each set of satellite hyperspectral data to obtain partitioning results of the satellite hyperspectral data; a labeling unit for labeling each pixel in each set of satellite hyperspectral data according to the geological data, elevation sub-data, and partitioning results to obtain labeled satellite hyperspectral data, each pixel in the labeled satellite hyperspectral data including classification attributes in three dimensions: weathering and vegetation, lithology type, and slope; and an encoding unit for sequentially performing sample cropping, merging of samples with the same attributes, and sample encoding on the labeled satellite hyperspectral data, and using the encoded satellite hyperspectral data as a sample in the satellite hyperspectral sample library.

[0006] Thirdly, this application also provides a satellite hyperspectral sample library construction device, comprising:

[0007] Memory, used to store computer programs;

[0008] A processor is used to implement the steps of the satellite hyperspectral sample library construction method when executing the computer program.

[0009] Fourthly, this application also provides a readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described method for constructing a satellite hyperspectral sample library.

[0010] The beneficial effects of this invention are as follows:

[0011] The sample presented in this application is a comprehensive sample integrating scene, target, and pixel; it has the advantage of meeting the accurate interpretation requirements of different levels such as scene, target, and pixel; and it supports the processing of multi-source heterogeneous remote sensing images. The sample produced by this method can obtain the range of all lithologies under the same image by searching image samples. Therefore, it has the advantage of serving as a comprehensive sample for deep learning models, including multiple lithologies and ages.

[0012] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing embodiments of the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0013] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a schematic diagram of the satellite hyperspectral sample library construction method described in this embodiment of the invention;

[0015] Figure 2 This is a schematic diagram of the satellite hyperspectral sample library construction device described in an embodiment of the present invention;

[0016] Figure 3 This is a schematic diagram of the satellite hyperspectral sample library construction device described in an embodiment of the present invention.

[0017] The diagram is labeled as follows: 1. Data acquisition unit; 2. Correction unit; 21. First correction unit; 22. Second correction unit; 23. Third correction unit; 3. Segmented calculation unit; 4. Partitioned calculation unit; 41. Image acquisition unit; 42. Image calculation unit; 43. Region division unit; 44. Depth interpretation unit; 5. Annotation unit; 6. Encoding unit; 7. Filtering unit; 71. Image processing unit; 72. Absorption peak filtering unit; 73. Evaluation value calculation unit; 731. Normalization unit. 732, Matrix Construction Unit; 7321, Fourth Calculation Unit; 7322, Fifth Calculation Unit; 7323, Sixth Calculation Unit; 7324, Seventh Calculation Unit; 733, Second Calculation Unit; 734, Third Calculation Unit; 74, Marking Unit; 75, First Calculation Unit; 76, Marking Judgment Unit; 800, Satellite Hyperspectral Sample Library Construction Equipment; 801, Processor; 802, Memory; 803, Multimedia Component; 804, I / O Interface; 805, Communication Component. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0019] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0020] Example 1:

[0021] This embodiment provides a method for constructing a satellite hyperspectral sample library.

[0022] See Figure 1 The figure shows that the method includes steps S100, S200, S300, S400 and S500.

[0023] S100. Obtain at least two sets of satellite hyperspectral data and corresponding basic data, including elevation data and geological data.

[0024] S200. Correct all satellite hyperspectral data and update the satellite hyperspectral data to the corrected data.

[0025] S300. Based on the elevation data corresponding to each satellite hyperspectral data, the slope is segmented to obtain at least two elevation sub-data sets, each containing a different slope.

[0026] It should be noted that in this step, the elevation data is generally represented by a digital elevation model, and those skilled in the art can calculate the slope to obtain the slope value, and use this as a basis to divide the elevation data into segments. For example, three segments: 0–30°, 30–60°, and greater than 60°.

[0027] S400. Based on each satellite hyperspectral data, weathering vegetation factor zoning is calculated to obtain the zoning results of the satellite hyperspectral data.

[0028] It should be noted that in this application, the determination of the partition type of satellite hyperspectral data is achieved through the following steps S410, S420, S430, and S440.

[0029] S410. Obtain the corresponding optical image for each satellite hyperspectral data set from the same period.

[0030] S420. Calculate the vegetation index for each pixel in the satellite hyperspectral image based on the NDVI or RENDVI calculation method.

[0031] It should be noted that methods such as NDVI or RENDV for calculating vegetation indices are existing technologies and will not be elaborated upon in this application.

[0032] S430. Based on the vegetation index of each pixel in the satellite hyperspectral image, the satellite hyperspectral data is divided into four regions, including a region without vegetation.

[0033] Specifically, in this application, the satellite hyperspectral data is divided into four regions based on the calculated vegetation index: water and snow zone, no vegetation zone, low vegetation zone (vegetation coverage < 50%), and high vegetation zone (vegetation coverage ≥ 50%).

[0034] S440. Based on the satellite hyperspectral data and corresponding optical images from the same period, the unvegetated areas are interpreted, and based on the interpretation results, the unvegetated areas are divided into the original rock outcrop area and the weathered area.

[0035] This step involves further subdividing the unvegetated areas based on the content of the optical images. This subdivision can be done manually or using a neural network model. The aim is to refine the zoning of weathered vegetation.

[0036] S500. Based on geological data, elevation sub-data and zoning results, each pixel in each satellite hyperspectral data is labeled to obtain labeled satellite hyperspectral data. Each pixel in the labeled satellite hyperspectral data includes classification attributes in three dimensions: weathering vegetation, lithology and slope.

[0037] It should be noted that this application can further refine the geological data, such as dividing the region into geological groups based on its age. Furthermore, the region can be subdivided into segments based on factors such as the mechanical properties of the lithology and whether it poses a threat to engineering construction and operation (coal-bearing strata, gypsum-bearing strata, oil and gas-bearing strata). This will increase the richness of the sample database.

[0038] Then, in this step, the geological data corresponding to each pixel in the hyperspectral image is divided into three dimensions: weathering vegetation type and slope.

[0039] S600. The labeled satellite hyperspectral data are sequentially processed by sample cropping, merging of samples with the same attributes, and sample encoding. The satellite hyperspectral data with the encoded samples are then used as a sample in the satellite hyperspectral sample library.

[0040] This step, to facilitate deep learning, involves cropping the labeled satellite hyperspectral data according to preset length, width, and height requirements. Intersection operations are then performed based on weathering vegetation, lithology, and slope. This process results in each region possessing classification attributes across three dimensions. Regions with the same classification attributes are then merged to obtain a single sample region with the same classification attribute.

[0041] Simultaneously, this step also involves coding the samples. A sample code is a combination of the sample data source, lithology, vegetation, weathering, slope, stratigraphic code, and sample number. Each sample is assigned a unique sample code as an identifier. To clarify, in this application, the sample coding is divided into a 5-level classification system, as shown in Table 1 below, which illustrates the hierarchical division and coding method of this classification system.

[0042] Table 1 Classification System and Coding Hierarchy

[0043]

[0044] It should be noted that the codes Q and E mentioned in Table 1... 2 r 3 Each of these is a code name, and each represents a type of stratigraphy in the stratigraphic code column. This application will not elaborate on its specific meaning, but it can be defined differently according to actual needs.

[0045] Meanwhile, in this application, corresponding basic information data of each satellite hyperspectral image sample can also be added, including a series of data such as the time of image acquisition, the person who collected the sample, and the person who reviewed the sample. This is existing technology and will not be described in detail in this application.

[0046] In this application, by setting up as described above, a comprehensive sample can be obtained by combining the sample basic information data, the sample area, and the image sample, forming an integrated sample that incorporates scene, target, and geographic information.

[0047] Furthermore, in order to achieve the purpose of screening out some unqualified samples in the satellite hyperspectral sample library in this embodiment, this application also includes a sample screening step in the S600 satellite hyperspectral sample library. Specifically, in the application, step S600 includes step S610.

[0048] S610. All the original hyperspectral curves in the sample are first enveloped and then smoothed to obtain smooth hyperspectral curves after envelope removal.

[0049] Specifically, in this step, envelope elimination can be performed first, and then the curve after envelope elimination can be smoothed or filtered to form a new curve. In this step, the influence of noise can be eliminated through smoothing or filtering.

[0050] S620. Based on all the preprocessed original hyperspectral curves, the absorption peak parameters corresponding to each smooth hyperspectral curve are obtained by screening. The absorption peak parameters include the number of absorption peaks and the wavelength of each absorption peak.

[0051] Specifically, in this application, a smooth hyperspectral curve can be obtained by vector comparison according to the wavelengths of all absorption peaks in the smooth hyperspectral curve and the wavelength corresponding to each absorption peak.

[0052] S630. Based on all the preprocessed original hyperspectral curves, calculate the difference evaluation value corresponding to each smooth hyperspectral curve.

[0053] It should be noted that the difference evaluation value in this application can be implemented using a prior art classification method, such as clustering algorithms, where the difference evaluation value represents the distance value in the clustering algorithm. Furthermore, it also provides a new evaluation method to better suit curve classification.

[0054] Specifically, step S630 includes steps S631, S632, S633 and S634.

[0055] S631. All the preprocessed original hyperspectral curves are normalized to obtain the normalized hyperspectral curve corresponding to each hyperspectral curve.

[0056] For ease of understanding, in this step, for any smooth hyperspectral curve A1=[a1,a2,a3,…,a…] i …,a n ], where n is the total number of channels, i is less than or equal to n, a i Let represent the value corresponding to channel i in a smooth hyperspectral curve. Find the maximum value among all channels, a. max =max(A1), the new spectral curve after normalization is

[0057] S632. Construct an evaluation matrix corresponding to each original hyperspectral curve based on all normalized hyperspectral curves. Each row of the evaluation matrix contains the comparison value between a reference hyperspectral curve and one of the curves in all normalized hyperspectral curves. The reference hyperspectral curve is one of the curves in all normalized hyperspectral curves. The evaluation matrix is ​​an asymmetric matrix.

[0058] Specifically, in this application, the D-values ​​are calculated pairwise for all spectral curves in a sample, where the D-values ​​are the comparison values. Assuming there are m spectral curves in the sample, then there are a total of... There are several arrangement methods. Selecting one as a reference and adjusting the other yields P. min The values ​​are then used to obtain the D value. The final D value will form a matrix, which is a matrix with 0s on the diagonal, i.e., D. ii =0,D im ≠D mi Where i is greater than or equal to 1 and less than or equal to m.

[0059]

[0060] It should be noted that this application also provides a method for calculating the D value, including steps S6321, S6322, S6323 and S6321.

[0061] S6321. An adjustment factor matrix is ​​calculated based on the reference hyperspectral curve and the comparison hyperspectral curve. Each element in the adjustment factor matrix is ​​the quotient of the values ​​of the reference hyperspectral curve and the comparison hyperspectral curve in the same channel. Assume a spectral curve is X = [x1, x2, x3, ..., x...]. i ,…,x n The other spectral curve is Y = [y1, y2, y3, ..., y]. i ,…,y n The adjustment factor matrix λ is then: λ = [λ1, λ2, λ3, ..., λ i …,λ n ], where λ i =y i / x i .

[0062] S6322. The evaluation value for each channel is calculated based on the evaluation value calculation formula, referring to the hyperspectral curve, comparing the hyperspectral curve and the adjustment factor matrix.

[0063] The calculation formula for each channel is as follows:

[0064]

[0065] function f(λ) iThe range of x is [0, 1*n), and k is the adjustment coefficient, whose main function is to adjust x. j -λ i y j For sensitivity to the magnitude of the difference, a value of 2 is generally appropriate, while a value of 1 can be used when the matching is strict.

[0066] S6323. Filter the evaluation values ​​under all channels to obtain the minimum evaluation value and the wavelength corresponding to the minimum evaluation value.

[0067] S6324. Divide the minimum evaluation value by the total number of channels to obtain the comparison value.

[0068] Specifically, when D approaches 0, the spectral curve is similar in shape to the average spectral curve; when D approaches 1, the spectral curve differs significantly in shape from the average spectral curve.

[0069] S633. The average of all elements in the evaluation matrix corresponding to each original hyperspectral curve.

[0070] S634. Calculate the difference value based on the evaluation matrix corresponding to each original hyperspectral curve, and use the average and the difference value as the difference evaluation value.

[0071] In this step, the difference between the i-th curve and all other spectral curves is represented by D. i express.

[0072]

[0073] In the calculation, it is necessary to iterate through and calculate the adjustment factor λ. i The adjustment factor is generally located in the channel of the non-absorption peak. Using the adjustment factor of the non-absorption peak can reduce the number of calculations.

[0074] S640. Mark pixels based on the difference evaluation value and absorption peak parameters corresponding to each smooth hyperspectral curve.

[0075] S650. The marking rate is calculated based on the marked pixels and the total number of pixels in the sample.

[0076] S660. Based on the labeling rate, the result of the current sample screening is obtained.

[0077] It should be noted that in this application, if the percentage of labeled pixels to total pixels is less than 0.1% for a given sample, that sample is discarded.

[0078] It should be noted that this application also provides a band selection method. In band difference analysis, the main purpose is to initially determine the characteristic bands of ground features, reduce the number of channels in the original image, reduce deep learning time, improve the learning and recognition efficiency of deep learning, and better extract the digital features of ground features.

[0079] Two types of ground feature samples were retrieved, namely, ground feature A and ground feature B. Band selection was determined using two methods, mainly through comparison of absorption peak bands and band calculation.

[0080] First, the absorption peaks of ground features A and B are extracted, including the number of channels of the absorption peaks. Then, the difference between the absorption peaks of ground features A and B is judged and compared by the spectral distance method.

[0081] The calculation formula for the spectral distance method is as follows:

[0082]

[0083]

[0084] S = S1 * (1 - S2)

[0085] Let t be the number of bands selected. The number of combinations of selecting t channels from n-channel spectral data is: The value of t can be determined manually, or using a sequential forward search (SFS) or a sequential backward search (SBS) algorithm. The larger the S value, the greater the difference between the two land features.

[0086] Compare the absorption peaks and wavelengths with larger S-values ​​between ground features A and B. Using these wavelengths for deep learning can achieve the best learning results.

[0087] The satellite hyperspectral sample library created using the method provided in this application integrates scene, target, pixel, and geographic information into a comprehensive sample. The image samples within the library are hyperspectral images, and different stratigraphic lithologies exhibit different hyperspectral spectral curves. Therefore, the samples used in this application are comprehensive samples integrating scene, target, and pixel information; they have the advantage of meeting the requirements for accurate interpretation at different levels such as scene, target, and pixel; and they support the processing of multi-source heterogeneous remote sensing images. Samples created using this method can obtain the range of all lithologies within the same image by searching image samples. Therefore, during deep learning model training, they can serve as comprehensive samples for deep learning of multiple lithologies, ages, etc.

[0088] Example 2:

[0089] like Figure 2 As shown, this embodiment provides a satellite hyperspectral sample library construction device, the device including:

[0090] Data acquisition unit 1 is used to acquire at least two sets of satellite hyperspectral data and the corresponding basic data, including elevation data and geological data.

[0091] The segmented calculation unit 3 is used to segment the slope according to the elevation data corresponding to each satellite hyperspectral data to obtain at least two elevation sub-data, each of which contains a different slope.

[0092] The partitioning calculation unit 4 is used to perform weathering vegetation factor partitioning calculations based on each satellite hyperspectral data to obtain the partitioning results of the satellite hyperspectral data.

[0093] Labeling unit 5 is used to label each pixel in each satellite hyperspectral data according to geological data, elevation sub-data and zoning results, to obtain labeled satellite hyperspectral data. Each pixel in the labeled satellite hyperspectral data includes classification attributes in three dimensions: weathering vegetation, lithology type and slope.

[0094] Encoding unit 6 is used to perform sample cropping, intra-sample merging of samples with the same attribute, and sample encoding on the labeled satellite hyperspectral data, and to use the encoded satellite hyperspectral data as a sample in the satellite hyperspectral sample library.

[0095] In some specific embodiments, the partition calculation unit 4 includes:

[0096] Image acquisition unit 41 is used to acquire optical images corresponding to each satellite hyperspectral data from the same period.

[0097] Image calculation unit 42 is used to calculate the vegetation index of each pixel in the satellite hyperspectral image according to the NDVI or RENDVI calculation method.

[0098] Region division unit 43 is used to divide the satellite hyperspectral data into regions based on the vegetation index of each pixel in the satellite hyperspectral image, resulting in four partitions, including a non-vegetated area.

[0099] The deep interpretation unit 44 is used to interpret the non-vegetation area based on the optical image corresponding to the same period of satellite hyperspectral data, and to divide the non-vegetation area into the original rock outcrop area and the weathering area based on the interpretation results.

[0100] In some specific embodiments, the device further includes a screening unit 7, which includes:

[0101] Image processing unit 71 is used to first eliminate the envelope of all the original hyperspectral curves in the sample and then perform smoothing to obtain smooth hyperspectral curves after the envelope of all the corresponding curves in the sample are eliminated.

[0102] The absorption peak screening unit 72 is used to screen the absorption peak parameters corresponding to each smooth hyperspectral curve based on the smooth curve after envelope elimination. The absorption peak parameters include the number of absorption peaks and the wavelength of each absorption peak.

[0103] Evaluation value calculation unit 73 is used to calculate the difference evaluation value corresponding to each hyperspectral curve based on all the preprocessed original hyperspectral curves.

[0104] The marking unit 74 is used to mark pixels based on the difference evaluation value and absorption peak parameters corresponding to each hyperspectral curve.

[0105] The first calculation unit 75 is used to calculate the marking rate based on the marked pixels and the total number of pixels in the sample.

[0106] The labeling judgment unit 76 is used to determine the current sample screening result based on the labeling rate.

[0107] In some specific embodiments, the evaluation value calculation unit 73 includes:

[0108] Normalization unit 731 normalizes all the preprocessed original hyperspectral curves to obtain the normalized hyperspectral curve corresponding to each original hyperspectral curve.

[0109] The matrix construction unit 732 is used to construct an evaluation matrix corresponding to each original hyperspectral curve based on all normalized hyperspectral curves. Each row of the evaluation matrix contains a comparison value between a reference hyperspectral curve and one of the curves in all normalized hyperspectral curves. The reference hyperspectral curve is one of the curves in all normalized hyperspectral curves. The evaluation matrix is ​​an asymmetric matrix.

[0110] The second calculation unit 733 is used to calculate the average of all elements in the evaluation matrix corresponding to each original hyperspectral curve.

[0111] The third calculation unit 734 is used to calculate the difference value based on the evaluation matrix corresponding to each original hyperspectral curve, and uses the average and the difference value as the difference evaluation value.

[0112] In some specific embodiments, the matrix construction unit 732 includes:

[0113] The fourth calculation unit 7321 is used to calculate the adjustment factor matrix based on the reference hyperspectral curve and the comparison hyperspectral curve. Each element in the adjustment factor matrix is ​​the quotient of the values ​​of the reference hyperspectral curve and the comparison hyperspectral curve in the same channel.

[0114] The fifth calculation unit 7322 is used to calculate the evaluation value for each channel based on the evaluation value calculation formula, the reference hyperspectral curve, the comparison hyperspectral curve, and the adjustment factor matrix.

[0115] The sixth calculation unit 7323 is used to filter the evaluation values ​​under all channels to obtain the minimum evaluation value and the wavelength corresponding to the minimum evaluation value.

[0116] The seventh calculation unit 7324 is used to divide the minimum evaluation value by the total number of channels to obtain the comparison value.

[0117] In some specific embodiments, the device further includes a correction unit 2, used to sequentially perform sample cropping, intra-sample merging of samples with the same attribute, and sample encoding on the labeled satellite hyperspectral data, and to use the encoded satellite hyperspectral data as a sample in the satellite hyperspectral sample library. The correction unit 2 includes:

[0118] The first correction unit 212 is used to perform radiometric correction, atmospheric correction and geometric correction on all satellite hyperspectral data to obtain all satellite hyperspectral data after preliminary correction.

[0119] The second correction unit 222 is used to perform topographic radiation correction on all satellite hyperspectral data after preliminary correction to obtain all satellite hyperspectral data after secondary correction.

[0120] The third correction unit 232 is used to perform error value and negative value removal processing on all satellite hyperspectral data after secondary correction to obtain the corrected data.

[0121] It should be noted that the specific manner in which each module performs its operation in the apparatus described in the above embodiments has been described in detail in the embodiments of the method, and will not be elaborated here.

[0122] Example 3:

[0123] Corresponding to the above method embodiments, this embodiment also provides a satellite hyperspectral sample library construction device. The satellite hyperspectral sample library construction device described below and the satellite hyperspectral sample library construction method described above can be referred to each other.

[0124] Figure 3 This is a block diagram illustrating a satellite hyperspectral sample library construction device 800 according to an exemplary embodiment. Figure 3 As shown, the satellite hyperspectral sample library construction device 800 may include a processor 801 and a memory 802. The satellite hyperspectral sample library construction device 800 may also include one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0125] The processor 801 controls the overall operation of the satellite hyperspectral sample library construction device 800 to complete all or part of the steps in the aforementioned satellite hyperspectral sample library construction method. The memory 802 stores various types of data to support the operation of the satellite hyperspectral sample library construction device 800. This data may include, for example, instructions for any application or method operating on the satellite hyperspectral sample library construction device 800, as well as application-related data such as contact data, sent and received messages, images, audio, video, etc. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or transmitted via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. I / O interface 804 provides an interface between processor 801 and other interface modules, such as a keyboard, mouse, and buttons. These buttons can be virtual or physical. Communication component 805 is used for wired or wireless communication between the satellite hyperspectral sample library construction device 800 and other devices. Wireless communication includes Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination thereof. Therefore, the corresponding communication component 805 may include a Wi-Fi module, a Bluetooth module, and an NFC module.

[0126] In an exemplary embodiment, the satellite hyperspectral sample library construction device 800 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the satellite hyperspectral sample library construction method described above.

[0127] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, which, when executed by a processor, implement the steps of the satellite hyperspectral sample library construction method described above. For example, the computer-readable storage medium may be the memory 802 including program instructions described above, which may be executed by the processor 801 of the satellite hyperspectral sample library construction device 800 to complete the satellite hyperspectral sample library construction method described above.

[0128] Example 4:

[0129] Corresponding to the above method embodiments, this embodiment also provides a readable storage medium. The readable storage medium described below can be referred to in conjunction with the satellite hyperspectral sample library construction method described above.

[0130] A readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the satellite hyperspectral sample library construction method described in the above method embodiments.

[0131] Specifically, the readable storage medium can be a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, or any other readable storage medium capable of storing program code.

[0132] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0133] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for constructing a satellite hyperspectral sample library, characterized in that, include: Acquire at least two sets of satellite hyperspectral data and corresponding basic data, including elevation data and geological data; Based on the elevation data corresponding to each set of satellite hyperspectral data, slope segmentation is performed to obtain at least two sets of elevation sub-data, each set of elevation sub-data containing a different slope. Based on each set of satellite hyperspectral data, weathering vegetation factor zoning is calculated to obtain the zoning results of the satellite hyperspectral data; Based on the geological data, elevation sub-data and zoning results, each pixel in each set of satellite hyperspectral data is labeled to obtain labeled satellite hyperspectral data. Each pixel in the labeled satellite hyperspectral data includes classification attributes in three dimensions: weathering vegetation, lithology type and slope. The labeled satellite hyperspectral data are sequentially processed by sample cropping, merging of samples with the same attribute, and sample encoding. The satellite hyperspectral data with sample encoding is then used as a sample in the satellite hyperspectral sample library. This also includes screening samples from the satellite hyperspectral sample library, the sample screening including: All the original hyperspectral curves in the sample are first enveloped and then smoothed to obtain smooth hyperspectral curves after envelope removal. Based on the smooth curve after envelope elimination, the absorption peak parameters corresponding to each smooth hyperspectral curve are obtained by screening. The absorption peak parameters include the number of absorption peaks and the wavelength of each absorption peak. Based on all the preprocessed original hyperspectral curves, the difference evaluation value corresponding to each smoothed hyperspectral curve is calculated; Pixels are marked based on the difference evaluation value and absorption peak parameters corresponding to each smooth hyperspectral curve; The labeling rate is calculated based on the labeled pixels and the total number of pixels in the sample. The result of the current sample screening is obtained based on the labeling rate; The step of calculating the difference evaluation value corresponding to each hyperspectral curve based on all the preprocessed original hyperspectral curves includes: All the preprocessed original hyperspectral curves are normalized to obtain the normalized hyperspectral curve corresponding to each original hyperspectral curve. An evaluation matrix is ​​constructed for each original hyperspectral curve based on all normalized hyperspectral curves. Each row of the evaluation matrix contains a comparison value between a reference hyperspectral curve and one of the curves in all normalized hyperspectral curves. The reference hyperspectral curve is one of the curves in all normalized hyperspectral curves. The evaluation matrix is ​​an asymmetric matrix. Based on the average of all elements in the evaluation matrix corresponding to each original hyperspectral curve; The difference value is calculated based on the evaluation matrix corresponding to each original hyperspectral curve, and the average and the difference value are used as the difference evaluation value.

2. The method for constructing a satellite hyperspectral sample library according to claim 1, characterized in that... The step of calculating weathering and vegetation factor partitions based on each set of satellite hyperspectral data to obtain the partitioning results corresponding to the satellite hyperspectral data includes: Obtain optical images corresponding to each set of satellite hyperspectral data from the same period; Calculate the vegetation index for each pixel in the satellite hyperspectral image using the NDVI or RENDVI calculation method; The satellite hyperspectral data is divided into four regions based on the vegetation index of each pixel in the satellite hyperspectral image, including a region without vegetation. The vegetation-free area is interpreted based on the corresponding optical images from the same period of the satellite hyperspectral data, and the vegetation-free area is divided into the original rock outcrop area and the weathering area based on the interpretation results.

3. A satellite hyperspectral sample library construction device, characterized in that, include: The data acquisition unit is used to acquire at least two sets of satellite hyperspectral data and corresponding basic data, including elevation data and geological data. The segmented calculation unit is used to segment the slope according to the elevation data corresponding to each set of satellite hyperspectral data to obtain at least two sets of elevation sub-data, and each set of elevation sub-data contains a different slope. The partitioning calculation unit is used to perform weathering vegetation factor partitioning calculations based on each piece of satellite hyperspectral data to obtain the partitioning results of the satellite hyperspectral data. The annotation unit is used to annotate each pixel in each set of satellite hyperspectral data according to the geological data, elevation sub-data and partitioning results, so as to obtain the annotated satellite hyperspectral data. Each pixel in the annotated satellite hyperspectral data includes classification attributes in three dimensions: weathering vegetation, lithology type and slope. The encoding unit is used to sequentially perform sample cropping, intra-sample merging of samples with the same attribute, and sample encoding on the labeled satellite hyperspectral data, and to use the satellite hyperspectral data after sample encoding as a sample in the satellite hyperspectral sample library; The satellite hyperspectral sample library construction device further includes a screening unit for screening samples in the satellite hyperspectral sample library. The sample screening includes: All the original hyperspectral curves in the sample are first enveloped and then smoothed to obtain smooth hyperspectral curves after envelope removal. Based on the smooth curve after envelope elimination, the absorption peak parameters corresponding to each smooth hyperspectral curve are obtained by screening. The absorption peak parameters include the number of absorption peaks and the wavelength of each absorption peak. Based on all the preprocessed original hyperspectral curves, the difference evaluation value corresponding to each smoothed hyperspectral curve is calculated; Pixels are marked based on the difference evaluation value and absorption peak parameters corresponding to each smooth hyperspectral curve; The labeling rate is calculated based on the labeled pixels and the total number of pixels in the sample. The result of the current sample screening is obtained based on the labeling rate; The satellite hyperspectral sample library construction device further includes a calibration unit for calibrating all satellite hyperspectral data and updating the satellite hyperspectral data to the calibrated data. The calibration unit includes: The first correction unit is used to perform radiometric correction, atmospheric correction and geometric correction on all satellite hyperspectral data to obtain all satellite hyperspectral data after preliminary correction. The second correction unit is used to perform topographic radiation correction on all satellite hyperspectral data after preliminary correction, so as to obtain all satellite hyperspectral data after secondary correction. The third correction unit is used to process all satellite hyperspectral data after secondary correction by removing error values ​​and negative values, thus obtaining the corrected data.

4. The satellite hyperspectral sample library construction device according to claim 3, characterized in that, The partition calculation unit includes: An image acquisition unit is used to acquire optical images corresponding to each set of satellite hyperspectral data from the same period. The image calculation unit is used to calculate the vegetation index for each pixel in the satellite hyperspectral image according to the NDVI or RENDVI calculation method. The region division unit is used to divide the satellite hyperspectral data into regions based on the vegetation index of each pixel in the satellite hyperspectral image, resulting in four partitions, including a no-vegetation area. The deep interpretation unit is used to interpret the vegetation-free area based on the optical images corresponding to the same period of the satellite hyperspectral data, and to divide the vegetation-free area into the original rock outcrop area and the weathering area based on the interpretation results.

5. A satellite hyperspectral sample library construction device, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the satellite hyperspectral sample library construction method as described in any one of claims 1 to 2.

6. A readable storage medium, characterized in that, The readable storage medium stores a computer program that, when executed by a processor, implements the steps of the satellite hyperspectral sample library construction method as described in any one of claims 1 to 2.

Images