A radiation ground truth region of interest extraction method and system

Abstract

The application relates to the field of remote sensing data processing, and provides a radiation true value region of interest extraction method and system, which comprises the following steps: S1, acquiring a reference data and an overlapping area of a to-be-inspected image, blocking the to-be-inspected image, and removing image blocks that do not meet the conditions; S2, calculating a variation coefficient of pixel brightness values in a radiation true value region of interest, and removing the radiation true value region of interest with a variation coefficient greater than a threshold value; S3, performing layered processing on the pixel brightness values in the radiation true value region of interest; S4, performing difference calculation on the average value of the pixel brightness values in the radiation true value region of interest and the reference pixel brightness values; and S5, extracting radiation true value regions of interest with different reflection brightness levels. The method combines the variation coefficient and uses the layered sampling method in the extraction of the planar radiation true value region of interest, and the extracted radiation true value region of interest is mainly distributed in an area with flat terrain, uniform ground object types and high radiation characteristic consistency, so that the planar true value is more accurate and more representative.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of remote sensing data processing, in particular to a radiation true value region of interest extraction method and system. BACKGROUND

[0002] Remote sensing data acquisition is a complex process, which is affected by many factors and links such as atmospheric radiation transmission characteristics, remote sensor operating environment, remote sensor working state, and observed target state. The remote sensing data received by the sensor may have errors, and must be verified for authenticity.

[0003] The commonly used authenticity verification methods are divided into two types: direct verification method based on ground measured data and indirect verification method based on reference image. Among them, the direct verification method based on ground measured data verifies the radiation accuracy authenticity by obtaining ground point true value data; the indirect verification method based on reference image can reduce the complexity of scale conversion in the process of directly converting points to surfaces to a certain extent, and has the advantages of not needing to carry out ground experiments, not being limited by the geographical space position and sample area quantity of the research area.

[0004] However, in the indirect verification method based on reference image, how to scientifically and reasonably extract the surface true value data to improve the accuracy of radiation authenticity verification is a big problem faced by the industry at present.

[0005] Therefore, it is necessary to provide an improved technical solution to the above-mentioned deficiencies of the prior art. SUMMARY

[0006] The purpose of the present application is to provide a radiation true value region of interest extraction method and system to solve or alleviate the problems existing in the prior art.

[0007] In order to achieve the above-mentioned purpose, the present application provides the following technical solution:

[0008] The present application provides a radiation true value region of interest extraction method, comprising:

[0009] Step S1, extracting an initial radiation true value region of interest according to a reference image and a to-be-verified image; the reference image and the to-be-verified image both contain multiple bands;

[0010] Step S2, calculating the coefficient of variation of the brightness value of each pixel in the initial radiation true value region of interest according to the band;

[0011] According to the coefficient of variation of the brightness value of the pixel in the initial radiation true value region of interest, the initial radiation true value region of interest is extracted, and the average value of the brightness value of the pixel in the extracted radiation true value region of interest and the corresponding central longitude and latitude are calculated;

[0012] Step S3, the maximum pixel brightness value and the minimum pixel brightness value of each band are obtained by counting the extracted radiation true value region of interest according to the band, and each band is processed based on the maximum pixel brightness value and the minimum pixel brightness value to obtain the reference pixel brightness value corresponding to each layer in each band.

[0013] Step S4, the absolute value of the difference between the average value of the pixel brightness value in the extracted radiation true value region of interest and the reference pixel brightness value of each layer of the corresponding band is calculated according to the band.

[0014] Step S5, the radiation true value region of interest with the minimum difference absolute value in each band is taken as the final radiation true value region of interest of the band.

[0015] Preferably, in step S1, the initial radiation true value region of interest is extracted according to the reference image and the image to be verified, specifically:

[0016] Step S101, the overlapping area of the reference image and the image to be verified is obtained.

[0017] Step S102, the image to be verified is divided into blocks according to a preset window size to obtain original blocks, and the latitude and longitude range of each original block is determined.

[0018] Step S103, whether each original block overlaps with the reference image is judged according to the latitude and longitude range of the original block and the overlapping area of the reference image and the image to be verified.

[0019] If the original block completely overlaps with the reference image, the original block is taken as the initial radiation true value region of interest.

[0020] Preferably, in step S2, the initial radiation true value region of interest is extracted according to the coefficient of variation of the pixel brightness value in the initial radiation true value region of interest, and the average value of the pixel brightness value in the extracted radiation true value region of interest and the corresponding central latitude and longitude are calculated, specifically:

[0021] Whether the coefficient of variation of the pixel brightness value in the initial radiation true value region of interest is less than the maximum threshold value of the coefficient of variation is judged, if yes, the average value of the pixel brightness value in the radiation true value region of interest and the corresponding central latitude and longitude are calculated, if not, the radiation true value region of interest is rejected.

[0022] Preferably, in step S2, the calculation formula of the coefficient of variation of the pixel brightness value in each initial radiation true value region of interest is as follows:

[0023]

[0024] In the formula, CV iσ is the coefficient of variation of the region of interest for the i-th initial true radiation value, i = 1, 2, 3, ..., n; i is the standard deviation of the pixel brightness values ​​within the region of interest for the i-th initial true radiation value; is the average pixel brightness value within the region of interest for the i-th initial true radiation value.

[0025] Preferably, in step S3, the extracted radiation true value region of interest is statistically analyzed according to bands to obtain the maximum and minimum pixel brightness values ​​for each band. Based on the maximum and minimum pixel brightness values, each band is layered to obtain the reference pixel brightness values ​​corresponding to each layer in each band. Specifically:

[0026] The extracted radiation true values ​​of the region of interest are statistically analyzed according to the spectral bands to obtain the maximum and minimum pixel brightness values ​​for each spectral band.

[0027] The pixel brightness values ​​of each band are sorted according to the preset sorting rules;

[0028] Based on the maximum and minimum pixel brightness values ​​of each band, and the sorting results of the pixel brightness values ​​of each band, the pixel brightness values ​​of the band are divided into multiple layers, and the reference pixel brightness values ​​corresponding to each layer in each band are obtained. The reference pixel brightness value corresponding to the j-th layer is denoted as:

[0029] Preferably, in step S4, the formula for calculating the absolute value of the difference between the average value of the pixel brightness value within the region of interest of the extracted true radiation and the brightness value of the reference pixel in each layer of the corresponding band is as follows:

[0030]

[0031] In the formula: The average pixel brightness value of the i-th radiometric region of interest. With pixel brightness value The absolute value of the difference between them, i = 1, 2, 3, ..., n; Let be the reference pixel brightness value corresponding to the j-th layer, where j = 1, 2, 3, ...

[0032] Preferably, step S5 further includes:

[0033] If multiple regions of interest (ROIs) with the same minimum absolute difference in their radiation true values ​​are found, the first ROI with the same minimum value is taken as the final ROI for that band.

[0034] This application provides a radiation truth region of interest extraction system, including:

[0035] The first extraction unit is configured to extract an initial radiation true value region of interest according to a reference image and a to-be-inspected image; the reference image and the to-be-inspected image both contain multiple wave bands;

[0036] The second extraction unit is configured to calculate a coefficient of variation of pixel brightness values in each of the initial radiation true value regions of interest according to wave bands;

[0037] According to the coefficient of variation of the pixel brightness values in the initial radiation true value region of interest, the initial radiation true value region of interest is extracted, and the average value of the pixel brightness values in the extracted radiation true value region of interest and the corresponding central longitude and latitude are calculated;

[0038] The hierarchical processing unit is configured to count the extracted radiation true value region of interest according to wave bands to obtain the maximum pixel brightness value and the minimum pixel brightness value of each wave band, and perform hierarchical processing on each wave band based on the maximum pixel brightness value and the minimum pixel brightness value to obtain the reference pixel brightness value corresponding to each hierarchical layer in each wave band;

[0039] The calculation unit is configured to calculate the absolute value of the difference between the average value of the pixel brightness values in the extracted radiation true value region of interest and the reference pixel brightness value of each hierarchical layer of the corresponding wave band according to wave bands;

[0040] The determination unit is configured to take the radiation true value region of interest with the minimum difference absolute value in each wave band as the final radiation true value region of interest of the wave band.

[0041] Advantageous effects:

[0042] The radiation true value region of interest extraction method and system provided in the application combine the coefficient of variation and use the hierarchical sampling method in the extraction of the planar radiation true value region of interest, avoid the uncertainty problem caused by human factors in the traditional manual visual recognition of the region of interest, solve the problem of large radiation difference and insufficient representation of the region of interest in the traditional region of interest extraction method, and the extracted radiation true value region of interest is mainly distributed in the region with flat terrain, uniform ground object type and high consistency of radiation characteristics, so that the planar true value is more accurate and more representative. BRIEF DESCRIPTION OF DRAWINGS

[0043] The drawings accompanying the specification of the application form part of the application and are used to provide a further understanding of the application. The illustrative embodiments of the application and their descriptions serve to explain the application without unduly limiting it. Among them:

[0044] FIG. 1 The technical logic diagram of the radiation true value region of interest extraction method provided according to some embodiments of the application;

[0045] FIG. 2A flowchart of a radiation true value region of interest extraction method according to some embodiments of the present application is provided;

[0046] FIG. 3 A flowchart of an initial radiation true value region of interest extraction method according to some embodiments of the present application is provided;

[0047] FIG. 4 A structural diagram of a radiation true value region of interest extraction system according to some embodiments of the present application is provided. DETAILED DESCRIPTION

[0048] As described in the background, the current commonly used remote sensing data authenticity verification method is divided into two kinds: direct verification method based on ground measured data and indirect verification method based on reference data. Compared with the radiation precision authenticity verification method based on ground point true value data, the verification method based on reference image data uses area true value data as an intermediate scale to reduce the complexity of the scale conversion from point to area to some extent, and has the advantages of not needing to carry out ground experiments, not being limited by the geographical space position and sample area quantity of the research area, etc.

[0049] However, how to scientifically and reasonably extract area true value data between reference data and data to be verified to make the verification result more objective and representative is a problem faced by the relative radiation precision authenticity verification method based on reference data.

[0050] At present, the area true value is mainly extracted by using the traditional manual visual recognition method, and the specific process is as follows: first, through visual interpretation, the region of interest (ROI) is selected in the "uniform" and "flat" area of the image; then, the region of interest that does not meet the condition is removed; and finally, the remaining region of interest is the area true value to be extracted. The region of interest extracted by this method has problems such as large pixel radiation difference, insufficient representativeness, and uncertainty caused by human subjective factors.

[0051] Therefore, the present application provides a radiation true value region of interest extraction method, which combines the coefficient of variation and uses the stratified sampling method for extracting the area radiation true value region of interest. The extracted radiation true value region of interest is mainly distributed in the area with flat terrain, uniform ground object type, and high consistency of radiation characteristics, and the area true value is more accurate and more representative.

[0052] The present application will be described in detail below with reference to the accompanying drawings and embodiments.

[0053] Exemplary method

[0054] The present application provides a radiation true value region of interest extraction method, as shown in FIG. 1 to FIG. 3 The method comprises the following steps:

[0055] Step S1, according to the reference image and the image to be tested, an initial radiation true value region of interest is extracted; the reference image and the image to be tested both contain multiple bands.

[0056] In the embodiments of the present application, the reference image data with high radiation precision is used as the reference data representing the relative "true value" of the ground target, so as to realize the radiation true value test of the image to be tested. The reference image can be a high-precision satellite image or aerial image.

[0057] It should be noted that the region of interest is a region to be processed determined from the image to be processed in the form of a box, a circle, an ellipse, or an irregular polygon. In the process of true value test of the image to be tested based on the test method of the reference image data, the region of interest can be a ground same-named region meeting the uniformity requirement and having a certain size extracted from the reference image and the image to be tested for carrying out the true value test.

[0058] In the embodiments of the present application, the reference image and the image to be tested both contain multiple bands, and the bands between the reference image and the image to be tested have a corresponding relationship. For example, the image to be tested contains three bands of red, green and blue, and the reference image also contains three bands of red, green and blue. Since the bands have a corresponding relationship, the region of interest extraction process in the subsequent steps is carried out band by band.

[0059] In some embodiments, in step S1, according to the reference image and the image to be tested, an initial radiation true value region of interest is extracted, specifically:

[0060] Step S101, an overlapping area of the reference image and the image to be tested is obtained. The overlapping area refers to a geographical space range covered by the reference image and the image to be tested together, and determining the overlapping area can lay a foundation for extracting the same-named point (i.e. the radiation true value region of interest) representing the same position on the ground. It should be understood that the overlapping area of the reference image and the image to be tested can be obtained by using corresponding software, such as GIS software.

[0061] Step S102, the image to be tested is divided into blocks according to a preset window size, to obtain original blocks, and the latitude and longitude range of each original block is determined. Here, the window size for dividing the image to be tested into blocks can be set according to actual needs, for example, the image to be tested is divided into 30*30 pixel blocks, or the image to be tested is divided into 50*50 pixel blocks.

[0062] Step S103, whether each original block overlaps with the reference image is judged according to the latitude and longitude range of the original block and the overlapping area of the reference image and the image to be tested. If the original block completely overlaps with the reference image, the original block is taken as an initial radiation true value region of interest.

[0063] Specifically, according to the latitude and longitude range of the original patch, whether each original patch falls entirely within the overlapping region of the reference image and the image to be verified can be determined based on the geographical range thereof. That is, for each original patch, if the geographical range thereof falls entirely within the overlapping region of the reference image and the image to be verified, the original patch is retained, otherwise, the original patch is excluded, thereby obtaining a set of original patches whose entire geographical range falls within the overlapping region, and taking the set as the initial radiation ground truth region of interest to ensure that each initial radiation ground truth region of interest is located within the overlapping region.

[0064] In step S2, the coefficient variation CV of the pixel brightness values in each initial radiation ground truth region of interest is calculated; according to the coefficient variation of the pixel brightness values in the initial radiation ground truth region of interest, the initial radiation ground truth region of interest is extracted, and the average value of the pixel brightness values in the extracted radiation ground truth region of interest and the corresponding central latitude and longitude are calculated.

[0065] The pixel brightness value is also referred to as a DN value, which records the gray value of a ground object and has no unit and is an integer. The DN value is related to the radiation resolution of a sensor, the reflectivity of a ground object, the atmospheric transmittance and scattering rate, etc.

[0066] The coefficient variation, also referred to as a dispersion coefficient, is a relative statistical quantity for measuring the variation degree of data, and is used to compare the variation degrees of two or more sample data with different average values. The coefficient variation is dimensionless, and can eliminate the influence of different units and (or) average values on the comparison of the variation degrees of two or more data. Generally, the smaller the coefficient variation, the smaller the variation degree of data; on the contrary, the larger the coefficient variation, the larger the variation degree of data.

[0067] In the embodiments of the present application, the initial radiation ground truth region of interest is screened according to the coefficient variation of the pixel brightness values in the initial radiation ground truth region of interest, so that the DN values of each radiation ground truth region of interest satisfy a preset condition, thereby obtaining a relatively uniform and flat ROI.

[0068] In the embodiments of the present application, in step S2, the calculation formula of the coefficient variation of the pixel brightness values in each initial radiation ground truth region of interest is as follows:

[0069]

[0070] In the formula, CVi is the coefficient variation of the i-th initial radiation ground truth region of interest, i = 1, 2, 3,..., n; σi is the standard deviation of the pixel brightness values in the i-th initial radiation ground truth region of interest; and μi is the average value of the pixel brightness values in the i-th initial radiation ground truth region of interest. i

[0071] 1, 2, 3,..., n; σi is the standard deviation of the pixel brightness values in the i-th initial radiation ground truth region of interest; and μi is the average value of the pixel brightness values in the i-th initial radiation ground truth region of interest. i In the formula, CVi is the coefficient variation of the i-th initial radiation ground truth region of interest, i = 1, 2, 3,..., n; σi is the standard deviation of the pixel brightness values in the i-th initial radiation ground truth region of interest; and μi is the average value of the pixel brightness values in the i-th initial radiation ground truth region of interest.​

[0072] is the average value of the DN values of the i-th initial radiation true value region of interest.

[0073] Further, in step S2, the initial radiation true value region of interest is extracted according to the coefficient of variation of the DN values in the initial radiation true value region of interest, and the average value of the DN values in the extracted radiation true value region of interest and the corresponding central longitude and latitude are calculated. Specifically, it is determined whether the coefficient of variation of the DN values in the initial radiation true value region of interest is less than the maximum threshold of the coefficient of variation. If yes, the average value of the DN values in the radiation true value region of interest and the corresponding central longitude and latitude are calculated. If no, the radiation true value region of interest is excluded.

[0074] The maximum threshold of the coefficient of variation can be set according to the heterogeneity of the ground surface and the spatial resolution of the image to be tested. For example, the maximum threshold of the coefficient of variation can be set to 2%. That is, if the coefficient of variation of the DN values in a certain radiation true value region of interest is less than 2%, the average value of the DN values in the radiation true value region of interest and the corresponding central longitude and latitude are recorded and calculated. If the coefficient of variation of the DN values in the radiation true value region of interest is greater than 2%, the radiation true value region of interest is excluded, so as to ensure that the extracted radiation true value region of interest is a region with uniform underlying surface radiation characteristics.

[0075] In step S3, the extracted radiation true value region of interest is counted according to the wave band to obtain the maximum pixel brightness value and the minimum pixel brightness value of each wave band, and each wave band is processed by layering based on the maximum pixel brightness value and the minimum pixel brightness value to obtain the reference pixel brightness value corresponding to each layer in each wave band.

[0076] In some embodiments, in step S3, the extracted radiation true value region of interest is counted according to the wave band to obtain the maximum pixel brightness value and the minimum pixel brightness value of each wave band, and each wave band is processed by layering based on the maximum pixel brightness value and the minimum pixel brightness value to obtain the reference pixel brightness value corresponding to each layer in each wave band. Specifically, the extracted radiation true value region of interest is counted according to the wave band to obtain the maximum pixel brightness value and the minimum pixel brightness value of each wave band. The pixel brightness values of each wave band are sorted according to a predetermined sorting rule. Based on the maximum pixel brightness value and the minimum pixel brightness value of each wave band and the sorting result of the pixel brightness values of each wave band, the pixel brightness values of the wave band are divided into multiple layers to obtain the reference pixel brightness value corresponding to each layer in each wave band, wherein the reference pixel brightness value corresponding to the j-th layer is denoted as

[0077] Exemplarily, the hierarchical processing of the DN value of the extracted radiation true value region of interest, in other words, the extracted radiation true value region of interest is counted according to the DN value of each band, and the maximum and minimum DN values of each band are obtained; then the maximum and minimum DN values are sorted in ascending order or descending order, and then a plurality of breakpoints are selected according to the sorting result, for example, 10 breakpoints are selected to divide the sorting result into 10 equal parts, and 11 reference values of DN values from small to large are obtained, denoted as In this way, it is ensured that the extracted radiation true value region of interest from the image can cover various ground object types with radiation characteristics from low reflection to high reflection.

[0078] The sorting rule of the pixel brightness value and the number of hierarchical layers and the selection method of the breakpoints can be set according to actual needs, for example, the hierarchical layers can also be divided into 20 equal parts or 30 equal parts, and the embodiments of the present application do not limit this.

[0079] Step S4, calculating the absolute value of the difference between the average value of the pixel brightness value in the extracted radiation true value region of interest and each hierarchical reference pixel brightness value of the corresponding band.

[0080] In some embodiments, in step S4, the calculation formula of the absolute value of the difference between the average value of the pixel brightness value in the extracted radiation true value region of interest and each hierarchical reference pixel brightness value of the corresponding band is as follows:

[0081]

[0082] In the formula: is the average value of the pixel brightness value of the i th radiation true value region of interest is the absolute value of the difference between the average value of the pixel brightness value and the pixel brightness value , i = 1, 2, 3, … n; is the reference pixel brightness value corresponding to the j th hierarchical layer, j = 1, 2, 3, ….

[0083] In the embodiments of the present application, the absolute value of the difference between the average value of the DN value of the radiation true value region of interest and the DN value of the reference image is calculated for each band, thereby obtaining the measurement of each radiation true value region of interest to the reference DN value of each hierarchical layer in each band, which lays the foundation for selecting the final radiation true value region of interest of the band.

[0084] Step S5, taking the radiation true value region of interest with the minimum absolute value of the difference in each band as the final radiation true value region of interest of the band.

[0085] In the embodiments of the present application, the radiation true value region of interest with the minimum absolute value in the multiple layers of each wave band is taken as the final radiation true value region of interest, so that the accuracy of the radiation true value region of interest is improved, and the accuracy of the authenticity inspection is further improved.

[0086] In some embodiments, the step S5 further includes: if the minimum absolute value of the difference values calculated by the multiple radiation true value regions of interest is equal, taking the first radiation true value region of interest with the minimum absolute value as the final radiation true value region of interest of the wave band. In this way, one radiation true value region of interest with the minimum absolute value can be obtained near the reference DN value of each layer of each wave band, and multiple radiation true value regions of interest are extracted, wherein the number of the extracted radiation true value regions of interest is the same as the number of the layers.

[0087] The radiation true value region of interest extraction method provided by the embodiments of the present application combines the coefficient of variation and uses the layer sampling method in the extraction of the planar radiation true value region of interest, avoids the uncertainty problem caused by human factors in the traditional manual visual identification of the region of interest, solves the problem of large radiation difference and insufficient representation of the pixels in the region of interest in the traditional region of interest extraction method, and extracts the radiation true value region of interest mainly distributed in the areas with flat terrain, uniform land cover types and high consistency of radiation characteristics. The planar true value is more accurate and more representative.

[0088] Exemplary system

[0089] The embodiments of the present application also provide a radiation true value region of interest extraction system, as shown in FIG. 4 The system includes a first extraction unit 401, a second extraction unit 402, a layer processing unit 403, a calculation unit 404 and a determination unit 405. Wherein:

[0090] The first extraction unit 401 is configured to extract an initial radiation true value region of interest according to a reference image and a to-be-inspected image; the reference image and the to-be-inspected image both contain multiple wave bands;

[0091] The second extraction unit 402 is configured to calculate the coefficient of variation of the brightness values of the pixels in each initial radiation true value region of interest according to the wave band;

[0092] According to the coefficient of variation of the brightness values of the pixels in the initial radiation true value region of interest, the initial radiation true value region of interest is extracted, and the average value of the brightness values of the pixels in the extracted radiation true value region of interest and the corresponding central longitude and latitude are calculated;

[0093] The hierarchical processing unit 403 is configured to count the extracted radiation true value region of interest according to the wave band, obtain the maximum pixel brightness value and the minimum pixel brightness value of each wave band, and perform hierarchical processing on each wave band based on the maximum pixel brightness value and the minimum pixel brightness value, to obtain the reference pixel brightness value corresponding to each hierarchical layer in each wave band.

[0094] The calculation unit 404 is configured to calculate the absolute value of the difference between the average value of the pixel brightness value in the extracted radiation true value region of interest and the reference pixel brightness value of each hierarchical layer corresponding to the wave band.

[0095] The determination unit 405 is configured to take the radiation true value region of interest with the minimum difference absolute value in each wave band as the final radiation true value region of interest of the wave band.

[0096] The radiation true value region of interest extraction system provided by the embodiments of the present application can realize the steps and processes of the radiation true value region of interest extraction method of any of the above embodiments and achieve the same technical effects, and thus will not be described here in detail.

[0097] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for extracting regions of interest from radiation truth values, characterized in that, include: Step S1: Extract the region of interest for the initial radiometric truth based on the reference image and the image to be tested; Both the reference image and the image to be tested contain multiple bands; Step S2: Calculate the coefficient of variation of the pixel brightness value within the region of interest for each of the initial true radiation values ​​according to the waveband; Based on the coefficient of variation of the pixel brightness values ​​within the initial true radiation region of interest, the initial true radiation region of interest is extracted, and the average value of the pixel brightness values ​​and the corresponding center latitude and longitude within the extracted true radiation region of interest are calculated. Step S3: Statistically analyze the extracted radiation true value region of interest according to the band to obtain the maximum and minimum pixel brightness values ​​of each band. Then, based on the maximum and minimum pixel brightness values, perform layer processing on each band to obtain the reference pixel brightness values ​​corresponding to each layer in each band. Step S4: Calculate the absolute value of the difference between the average value of the pixel brightness value in the region of interest extracted by the radiometric true value and the brightness value of the reference pixel in each layer of the corresponding band. Step S5: Select the region of interest with the smallest absolute difference in radiation true value in each band as the final region of interest for radiation true value in that band.

2. The method for extracting the region of interest from the true radiation value according to claim 1, characterized in that, In step S1, the step of extracting the region of interest for the initial radiometric truth based on the reference image and the image to be tested specifically involves: Step S101: Obtain the overlapping area between the reference image and the image to be tested; Step S102: Divide the image to be inspected into blocks according to the preset window size to obtain the original blocks, and determine the latitude and longitude range of each original block; Step S103: Based on the latitude and longitude range of the original blocks and the overlapping area between the reference image and the image to be tested, determine whether each original block overlaps with the reference image. If the original block completely overlaps with the reference image, then the original block is taken as the initial radiometric ground truth region of interest.

3. The method for extracting the region of interest from the true radiation value according to claim 1, characterized in that, In step S2, the region of interest (ROI) of the initial radiometric truth region of interest is extracted based on the coefficient of variation of the pixel brightness values ​​within the ROI, and the average value of the pixel brightness values ​​and the corresponding center latitude and longitude within the extracted ROI are calculated, specifically as follows: Determine whether the coefficient of variation of the pixel brightness value within the region of interest of the initial radiometric ground truth is less than the maximum threshold of the coefficient of variation; if so, calculate the average value of the pixel brightness value within the region of interest of the radiometric ground truth and the corresponding center latitude and longitude; if not, remove the region of interest of the radiometric ground truth.

4. The method for extracting the region of interest from the true radiation value according to claim 1, characterized in that, In step S2, the formula for calculating the coefficient of variation of the pixel brightness value within the region of interest of each initial radiation true value is as follows: Where: CV i σ is the coefficient of variation of the region of interest for the i-th initial true radiation value, i = 1, 2, 3, ..., n; i is the standard deviation of the pixel brightness values ​​within the region of interest for the i-th initial true radiation value; is the average pixel brightness value within the region of interest for the i-th initial true radiation value.

5. The method for extracting the region of interest from the true radiation value according to claim 1, characterized in that, In step S3, the extracted radiation true value regions of interest are statistically analyzed according to bands to obtain the maximum and minimum pixel brightness values ​​for each band. Based on the maximum and minimum pixel brightness values, each band is layered to obtain the reference pixel brightness values ​​corresponding to each layer in each band. Specifically: The extracted radiation true values ​​of the region of interest are statistically analyzed according to the spectral bands to obtain the maximum and minimum pixel brightness values ​​for each spectral band. The pixel brightness values ​​of each band are sorted according to the preset sorting rules; Based on the maximum and minimum pixel brightness values ​​of each band, and the sorting results of the pixel brightness values ​​of each band, the pixel brightness values ​​of the band are divided into multiple layers, and the reference pixel brightness values ​​corresponding to each layer in each band are obtained. The reference pixel brightness value corresponding to the j-th layer is denoted as:

6. The method for extracting the region of interest from the true radiation value according to claim 1, characterized in that, In step S4, the formula for calculating the absolute value of the difference between the average value of the pixel brightness value within the region of interest of the extracted true radiation and the brightness value of the reference pixel in each layer of the corresponding band is as follows: In the formula: The average pixel brightness value of the i-th radiometric region of interest. With pixel brightness value The absolute value of the difference between them, i = 1, 2, 3, ..., n; Let be the reference pixel brightness value corresponding to the j-th layer, where j = 1, 2, 3, ...

7. The method for extracting the region of interest from the true radiation value according to claim 1, characterized in that, Step S5 also includes: If multiple regions of interest (ROIs) with the same minimum absolute difference in their radiation true values ​​are found, the first ROI with the same minimum value is taken as the final ROI for that band.

8. A radiation truth region of interest extraction system, characterized in that, include: The first extraction unit is configured to extract the region of interest of initial radiometric truth based on the reference image and the image to be examined; Both the reference image and the image to be tested contain multiple bands; The second extraction unit is configured to calculate the coefficient of variation of the pixel brightness value within the region of interest of each initial radiation true value according to the band. Based on the coefficient of variation of the pixel brightness values ​​within the initial true radiation region of interest, the initial true radiation region of interest is extracted, and the average value of the pixel brightness values ​​and the corresponding center latitude and longitude within the extracted true radiation region of interest are calculated. The layered processing unit is configured to perform statistical analysis on the extracted radiation true value region of interest according to the band, obtain the maximum and minimum pixel brightness values ​​of each band, and perform layered processing on each band based on the maximum and minimum pixel brightness values ​​to obtain the reference pixel brightness values ​​corresponding to each layer in each band. The calculation unit is configured to calculate the absolute value of the difference between the average value of the pixel brightness value in the region of interest extracted by the true radiation value by band and the brightness value of the reference pixel in each layer of the corresponding band. The element is determined and configured to take the region of interest with the smallest absolute value of the difference in radiation truth value in each band as the final region of interest for radiation truth value in that band.

Images