A radiation correction method and system for remote sensing images

Abstract

The present invention discloses a radiation correction method and system for remote sensing images, which relates to the technical field of remote sensing image processing. The radiation correction method for remote sensing images includes: obtaining a dark channel image of a target remote sensing image, where the dark channel image is an image composed of the minimum channel pixel values of each pixel point in the target remote sensing image, and the minimum channel pixel value is the minimum value among the channel pixel values of the pixel point in each color channel; determining the atmospheric transmittance and reflectivity evaluation value of each image region in the target remote sensing image according to the dark channel image, and the reflectivity evaluation value is used to characterize the strength of the object reflectivity in the image region; determining the target dark pixels of the target remote sensing image according to the atmospheric transmittance and reflectivity evaluation value of each image region in the target remote sensing image; and performing radiation correction on the target remote sensing image based on the target dark pixels to obtain a corrected remote sensing image. The present invention can effectively perform radiation correction on the target remote sensing image and improve the effect of radiation correction.

Description

Technical Field

[0001] The present invention relates to the technical field of remote sensing image processing, and particularly to a method and system for radiometric correction of remote sensing images. Background Art

[0002] With the development of technology, the acquisition and application of remote sensing images have become more and more extensive, covering multiple fields such as environmental monitoring, urban planning, and agricultural management. When acquiring remote sensing images, it is usually necessary to perform radiometric correction on them, and an important link in radiometric correction is atmospheric correction.

[0003] Among various atmospheric correction methods, the dark pixel method is more suitable for atmospheric correction of remote sensing images in scenarios with complex atmospheric conditions or lack of meteorological information because it is simple and easy to use and does not require complex meteorological data. In the existing dark pixel method, when performing atmospheric correction on remote sensing images, generally the minimum value of the bands in the remote sensing image data or the value input by the user is selected as the dark pixel.

[0004] However, the accuracy of the dark pixels selected by the above method is poor, resulting in ineffective radiometric correction and poor radiometric correction effect. Summary of the Invention

[0005] Embodiments of the present invention provide a method and system for radiometric correction of remote sensing images, which can effectively perform radiometric correction on target remote sensing images and improve the radiometric correction effect.

[0006] In the first aspect of the embodiments of the present invention, a method for radiometric correction of remote sensing images is provided, including:

[0007] Obtaining a dark channel image of the target remote sensing image, where the dark channel image is an image composed of the minimum channel pixel values of each pixel point in the target remote sensing image, and the minimum channel pixel value is the minimum value among the channel pixel values of the pixel point in each color channel;

[0008] According to the dark channel image, determining the atmospheric transmittance and reflectance evaluation value of each image region in the target remote sensing image, where the reflectance evaluation value is used to characterize the strength of the object reflectance in the image region;

[0009] According to the atmospheric transmittance and reflectance evaluation value of each image region in the target remote sensing image, determining the target dark pixel of the target remote sensing image;

[0010] Based on the target dark pixel, performing radiometric correction on the target remote sensing image to obtain a corrected remote sensing image.

[0011] In the second aspect of the embodiments of the present invention, a system for radiometric correction of remote sensing images is provided, including:

[0012] An image acquisition module, configured to acquire a dark channel image of a target remote sensing image, where the dark channel image is an image composed of the minimum channel pixel values of each pixel point in the target remote sensing image, and the minimum channel pixel value is the minimum value among the channel pixel values of the pixel point in each color channel;

[0013] An image calculation module, configured to determine the atmospheric transmittance and reflectance evaluation value of each image region in the target remote sensing image according to the dark channel image, where the reflectance evaluation value is used to characterize the strength of the reflectance of the object in the image region;

[0014] A dark pixel determination module, configured to determine the target dark pixels of the target remote sensing image according to the atmospheric transmittance and reflectance evaluation value of each image region in the target remote sensing image;

[0015] A radiation correction module, configured to perform radiation correction on the target remote sensing image based on the target dark pixels to obtain a corrected remote sensing image.

[0016] In the radiation correction method for remote sensing images provided by the embodiments of the present invention, the atmospheric transmittance and reflectance evaluation value of each image region in the target remote sensing image are determined according to the dark channel image of the target remote sensing image. Then, the target dark pixels of the target remote sensing image are determined according to the atmospheric transmittance and reflectance evaluation value of each image region in the target remote sensing image. Finally, radiation correction is performed on the target remote sensing image based on the target dark pixels to obtain a corrected remote sensing image. In this way, the present invention comprehensively evaluates according to the reflectance of the object in the target remote sensing image and the transmittance of the atmosphere, so as to accurately determine the target dark pixels of the target remote sensing image. Furthermore, effective radiation correction can be performed on the target remote sensing image to improve the effect of radiation correction. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a schematic flowchart of the first radiation correction method for remote sensing images provided by an embodiment of the present invention;

[0019] Figure 2 It is a schematic flowchart of the second radiation correction method for remote sensing images provided by an embodiment of the present invention;

[0020] Figure 3 It is a schematic flowchart of the third radiation correction method for remote sensing images provided by an embodiment of the present invention;

[0021] Figure 4 Schematic flowchart of the fourth radiometric correction method for remote sensing images provided by an embodiment of the present invention;

[0022] Figure 5 Schematic structural diagram of a radiometric correction system for remote sensing images provided by an embodiment of the present invention. Detailed implementation manners

[0023] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, details the specific implementation manners, structures, features and effects of a radiometric correction method and system for remote sensing images proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0025] It should be noted that the acquisition, storage, use, processing, etc. of data in the technical solution of the present invention all comply with the relevant regulations of laws and regulations.

[0026] It should be noted that in the embodiments of the present invention, some industry-existing solutions such as certain software, components, models, etc. may be mentioned. They should be regarded as exemplary, and their purpose is only to illustrate the feasibility in the implementation of the technical solution of the present invention, but it does not mean that the applicant has already or necessarily used this solution.

[0027] In the existing atmospheric correction methods, the dark pixel method is more suitable for the atmospheric correction of remote sensing images in scenarios with relatively complex atmospheric conditions or lack of meteorological information because it is simple and easy to use and does not require complex meteorological data. In the existing dark pixel method, when performing atmospheric correction on remote sensing images, generally the minimum value of the bands in the remote sensing image data or the value input by the user is selected as the dark pixel. However, the accuracy of the selected dark pixel in this method is poor, resulting in ineffective radiometric correction and poor radiometric correction effect.

[0028] The object of the present invention is to provide a radiometric correction method and system for remote sensing images. In the radiometric correction method for remote sensing images provided by the embodiments of the present invention, according to the dark channel image of the target remote sensing image, the atmospheric transmittance and reflectivity evaluation values of each image region in the target remote sensing image are determined. Then, according to the atmospheric transmittance and reflectivity evaluation values of each image region in the target remote sensing image, the target dark pixels of the target remote sensing image are determined. Finally, based on the target dark pixels, radiometric correction is performed on the target remote sensing image to obtain a corrected remote sensing image. In this way, the present invention comprehensively evaluates according to the reflectivity of the objects in the target remote sensing image and the transmittance of the atmosphere, so as to accurately determine the target dark pixels of the target remote sensing image. Furthermore, effective radiometric correction can be performed on the target remote sensing image to improve the effect of radiometric correction.

[0029] The following introduces specific embodiments of a radiometric correction method and system for remote sensing images provided by the embodiments of the present invention.

[0030] Figure 1 A flowchart of a radiometric correction method for remote sensing images is provided. The radiometric correction method for remote sensing images can be applied to a server, and the radiometric correction method for remote sensing images may include the following S101 to S104.

[0031] S101, obtain the dark channel image of the target remote sensing image. The dark channel image is an image composed of the minimum channel pixel values of each pixel point in the target remote sensing image, and the minimum channel pixel value is the minimum value among the channel pixel values of the pixel point in each color channel.

[0032] In this embodiment, the target remote sensing image is usually a multi-band image, such as an image constructed from satellite data such as Landsat or Sentinel-2.

[0033] The dark channel image (DCI) is an image constructed based on a target hypothesis, that is, in most regions, at least one color channel has a very low value (close to 0) at each pixel position, which is usually caused by atmospheric scattering.

[0034] The channel pixel value is used to characterize the pixel value of the pixel point in the target remote sensing image corresponding to each color channel. Exemplarily, the channel pixel value of the pixel point may include the red channel pixel value 、the green channel pixel value and the blue channel pixel value .

[0035] As an example, the server first reads the target remote sensing image from the database. Then, for each pixel point in the target remote sensing image, the minimum value is found among the channel pixel values of all its corresponding color channels, and this minimum pixel value is used as the pixel value of the dark channel image at the corresponding pixel point position, thereby obtaining the dark channel image corresponding to the target remote sensing image.

[0036] S102. According to the dark channel image, determine the atmospheric transmittance and reflectivity evaluation value of each image region in the target remote sensing image. The reflectivity evaluation value is used to characterize the strength of the object reflectivity in the image region.

[0037] In this embodiment, the target remote sensing image includes multiple image regions. For example, after obtaining the target remote sensing image, the server can traverse the target remote sensing image based on a sliding window of a preset size, thereby dividing the target remote sensing image into multiple image regions.

[0038] During the process of the radiation of an object penetrating the atmosphere, it will be affected by three factors, specifically including atmospheric transmittance, atmospheric absorptance, and object reflectivity. Among them, the higher the atmospheric transmittance, the lower the atmospheric absorptance, and the higher the object reflectivity, the more accurate the radiation of the object collected.

[0039] Atmospheric transmittance and atmospheric absorptance are two related parameters, so only one of them needs to be determined. Among them, atmospheric transmittance is a parameter that describes the attenuation of radiation intensity after passing through the atmosphere, and its definition is the ratio of the radiation intensity after passing through the atmosphere to the radiation intensity before incidence; atmospheric absorptance is a parameter that describes the attenuation of radiation intensity when radiation passes through the atmosphere due to the absorption of atmospheric components, and its definition is the ratio of the absorption attenuation amount to the incident radiation after the radiation passes through a gas layer with only atmospheric absorption.

[0040] Object reflectivity is a parameter that describes the reflection performance of an object to radiation, and its definition is the ratio of the radiation flux reflected by the object surface to the incident radiation flux. The reflectivity evaluation value is used to characterize the strength of the object reflectivity in the image region. Specifically, the larger the reflectivity evaluation value, the larger the object reflectivity; the smaller the reflectivity evaluation value, the smaller the object reflectivity.

[0041] As an example, the server inputs the dark channel image into a pre-trained atmospheric transmittance recognition model to obtain the atmospheric transmittance of each image region in the target remote sensing image; then, calculate the average pixel value or pixel value standard deviation of each image region in the dark channel image, and compare the average pixel value or pixel value standard deviation between each image region in the dark channel image, thereby determining the reflectivity evaluation value of each image region in the target remote sensing image.

[0042] S103. Determine the target dark pixels of the target remote sensing image according to the atmospheric transmittance and reflectivity evaluation values of each image region in the target remote sensing image.

[0043] In this embodiment, the target dark pixels are used to represent the pixels in the target remote sensing image with very low atmospheric transmittance and object reflectivity in the visible light band. Among them, the target dark pixels can reflect the influence of the atmosphere on radiation, so they are often used as reference points for atmospheric correction.

[0044] As an example, the server can screen and determine the target dark pixels in the target remote sensing image based on the atmospheric transmittance and reflectivity evaluation values of each image region in the target remote sensing image.

[0045] Specifically, the server determines the image region corresponding to the target dark pixels as the image region with the lowest atmospheric transmittance and the lowest reflectivity evaluation value. Then, calculate the average value of the pixel values of each pixel point in this image region, and use this average value as the target dark pixel.

[0046] S104. Based on the target dark pixels, perform radiometric correction on the target remote sensing image to obtain a corrected remote sensing image.

[0047] In this embodiment, the server establishes a radiometric correction model based on the target dark pixels. Specifically, the radiometric correction model aims to eliminate the influence of the atmosphere on the target remote sensing image.

[0048] Then, apply the established radiometric correction model to the entire target remote sensing image. For each pixel point in the target remote sensing image, use the radiometric correction model to adjust the pixel value, so as to obtain a corrected remote sensing image.

[0049] In the radiometric correction method for remote sensing images provided in this embodiment, according to the dark channel image of the target remote sensing image, determine the atmospheric transmittance and reflectivity evaluation values of each image region in the target remote sensing image. Then, according to the atmospheric transmittance and reflectivity evaluation values of each image region in the target remote sensing image, determine the target dark pixels of the target remote sensing image. Finally, based on the target dark pixels, perform radiometric correction on the target remote sensing image to obtain a corrected remote sensing image. In this way, the present invention comprehensively evaluates according to the reflectivity of the object and the transmittance of the atmosphere in the target remote sensing image, so as to accurately determine the target dark pixels of the target remote sensing image. Furthermore, it can effectively perform radiometric correction on the target remote sensing image and improve the effect of radiometric correction.

[0050] As an alternative embodiment, as Figure 2 shown, S101 may specifically include the following S201 to S203.

[0051] S201. Obtain remote sensing image data;

[0052] S202. Convert the remote sensing image data to obtain the target remote sensing image;

[0053] S203. Determine the dark channel image of the target remote sensing image according to the pixel values of each pixel point in the target remote sensing image.

[0054] In this embodiment, the remote sensing image data is not presented as an image in the form of pixel points, but is presented in the form of data brightness values (Digital Number, DN) that are not clearly quantified and have no unified dimension. Therefore, it is necessary to first convert the remote sensing image data into a target remote sensing image.

[0055] As an example, the server first imports the remote sensing image data, and then performs radiometric calibration on the remote sensing image data, that is, converts the data brightness value of the remote sensing image data into a physically meaningful radiometric value (i.e., radiance). Then, according to the converted result, perform format conversion to convert it into a target remote sensing image in the GeoTIFF format containing three channels.

[0056] Then, for each pixel point in the target remote sensing image, find the minimum pixel value in all its corresponding color channels, and use this minimum pixel value as the pixel value of the dark channel image at the corresponding pixel point position, so as to obtain the dark channel image corresponding to the target remote sensing image. Further, a local minimum filter can be used to smooth the dark channel image to reduce noise and artifacts.

[0057] Through this embodiment, the obtained remote sensing image data is converted to obtain the target remote sensing image. Then, according to the pixel values of each pixel point in the target remote sensing image, the dark channel image of the target remote sensing image is determined. In this way, the dark channel image of the target remote sensing image can be accurately obtained, which helps to perform precise analysis on the dark channel image subsequently, enabling effective radiometric correction of the target remote sensing image and improving the effect of radiometric correction.

[0058] As an alternative embodiment, S203 may specifically include:

[0059] Obtain the channel pixel values of each pixel point in the target remote sensing image in each color channel;

[0060] For each pixel point in the target remote sensing image, respectively determine the minimum value among the channel pixel values as the dark channel pixel value of the pixel point;

[0061] Based on the dark channel pixel values of each pixel point, construct the dark channel image of the target remote sensing image.

[0062] In this embodiment, the dark channel pixel value is used to represent the pixel value of the dark channel image at the corresponding pixel point position.

[0063] As an example, the server separately obtains the channel pixel values of each pixel point in the red channel, green channel, and blue channel of the target remote sensing image. . Then, among the of each pixel point, the minimum value is found as the dark channel pixel value of the corresponding pixel point.

[0064] Finally, the dark channel pixel values of each obtained pixel point are projected and distributed according to the positions of each pixel point in the original target remote sensing image, so as to form the dark channel image corresponding to the target remote sensing image.

[0065] Through this embodiment, for each pixel point in the target remote sensing image, the minimum value among the channel pixel values is respectively determined as the dark channel pixel value of the pixel point; finally, based on the dark channel pixel values of each pixel point, the dark channel image of the target remote sensing image is constructed. In this way, the dark channel image of the target remote sensing image can be accurately obtained, which helps to perform precise analysis on the dark channel image subsequently, enabling effective radiometric correction of the target remote sensing image and improving the effect of radiometric correction.

[0066] As an alternative embodiment, as Figure 3 shown, S102 may specifically include the following S301 to S304.

[0067] S301, perform mean processing on the dark channel pixel values of each target pixel point in the dark channel image that meets the target screening condition to obtain the atmospheric light value, where the target screening condition is that the dark channel pixel value is greater than the preset pixel value threshold;

[0068] S302, divide the target remote sensing image into multiple image regions;

[0069] S303, based on the atmospheric light value, determine the atmospheric transmittance of each image region;

[0070] S304, according to the atmospheric transmittance of each image region and the dark channel pixel values of each pixel point in each image region, determine the reflectivity evaluation value of each image region.

[0071] In this embodiment, the preset pixel value threshold can be a pixel value threshold input by the user, or a pixel value threshold determined according to the pixel value distribution of each pixel point in the dark channel image. Exemplarily, the pixel values of each pixel point in the dark channel image can be sorted from large to small, and the pixel value at the 0.1% position is determined as the preset pixel value threshold, that is, the brightest top 0.1% of the pixel points in the dark channel image are screened out.

[0072] The atmospheric light value is used to characterize the light intensity of sunlight reaching the ground after passing through the atmosphere.

[0073] As an example, the server sorts the pixel values of each pixel point in the dark channel image from largest to smallest, and determines the preset pixel value threshold as the pixel value at the 0.1% position. Then, it traverses each pixel point in the dark channel image and compares its dark channel pixel value with the preset pixel value threshold; if it is greater than the preset pixel value threshold, it is considered that the pixel point meets the target screening condition and is included in the subsequent calculation. Then, the average value of the dark channel pixel values of all pixel points that meet the target screening condition is processed to obtain the atmospheric light value.

[0074] At the same time, the server determines a suitable partitioning method according to characteristics such as the size and resolution of the image. For example, the image can be divided into grid areas of a fixed size, or irregularly divided according to specific features in the image. Then, according to the determined partitioning method, the target remote sensing image is partitioned to obtain multiple image areas.

[0075] Then, the server uses an atmospheric scattering model (such as the Koschmieder model) to describe the light transmission process in the target remote sensing image. Among them, the atmospheric scattering model includes two key parameters: atmospheric transmittance and atmospheric light value. Therefore, based on the atmospheric scattering model and the known atmospheric light value, the server can obtain the atmospheric transmittance of each image area.

[0076] Finally, for each image area, the server extracts the dark channel pixel values of all pixel points in the image area and calculates the average value or standard deviation. Then, according to the atmospheric transmittance of each image area and the average pixel value or pixel value standard deviation of each image area, the reflectance evaluation value of each image area in the target remote sensing image is evaluated.

[0077] Through this embodiment, the average value of the dark channel pixel values of each target pixel point that meets the target screening condition in the dark channel image is processed to obtain the atmospheric light value; then, based on the atmospheric light value, the atmospheric transmittance of each image area is determined; finally, according to the atmospheric transmittance of each image area and the dark channel pixel values of each pixel point in each image area, the reflectance evaluation value of each image area is determined. In this way, this embodiment can accurately determine the atmospheric transmittance and reflectance evaluation value of each image area in the target remote sensing image, which helps to comprehensively evaluate according to the reflectance of the object in the target remote sensing image and the transmittance of the atmosphere, so as to accurately determine the target dark pixels of the target remote sensing image and improve the effect of radiometric correction.

[0078] As an alternative embodiment, S302 may specifically include:

[0079] Using The sliding window of divides the target remote sensing image to obtain the first information complexity mean value. The first information complexity mean value is The average information complexity of each first candidate region after partitioning with a sliding window, the initial value of N is 3;

[0080] Using Partition the target remote sensing image with a sliding window to obtain a second average information complexity, and the second average information complexity is the average information complexity of each first candidate region after partitioning with a sliding window;

[0081] When the absolute value of the difference between the first average information complexity and the second average information complexity is greater than a preset difference threshold, update N to N + 2 and update the first average information complexity to the second average information complexity, and return to loop and execute using a sliding window to partition the target remote sensing image to obtain a second average information complexity;

[0082] When the absolute value of the difference between the first average information complexity and the second average information complexity is not greater than the preset difference threshold, each first candidate region obtained by partitioning with a sliding window is determined as each image region of the target remote sensing image.

[0083] In this embodiment, the information complexity is used to characterize the intricacy of the information contained in the target remote sensing image, reflecting the visual complexity of the target remote sensing image.

[0084] This embodiment uses the information complexity of the image region as a basis for image region partitioning. Generally, the information complexity corresponding to the same scene is the same; at the same time, the air quality of the same scene is relatively close, so its atmospheric transmittance is basically the same.

[0085] As an example, the server first uses a sliding window to partition the target remote sensing image to obtain corresponding multiple first candidate regions. And calculate the information complexity of each first candidate region through the following formula 1:

[0086] , formula 1,

[0087] In formula 1, is used to characterize the information complexity of the mth first candidate region, is used to characterize the probability that a pixel point with a pixel value of g appears in the mth first candidate region, and log is used to represent the logarithmic operation.

[0088] Then, calculate the average value of the information complexity of each first candidate region to obtain the average information complexity in the case of partitioning with a sliding window.

[0089] Then use The target remote sensing image is divided by the sliding window to obtain a corresponding plurality of first candidate regions. Then, the information complexity of each first candidate region is calculated by the above formula 1. Next, the average value of the information complexity of each first candidate region is calculated to obtain the average value of the information complexity under the sliding window division of

[0090] Then, calculate the absolute value of the difference between the average value of the information complexity under the sliding window division of and the average value of the information complexity under the sliding window division of . If the absolute value of this difference is greater than the preset difference threshold, it indicates that the current division situation is unreasonable.

[0091] Then, continue to divide the target remote sensing image with the sliding window to obtain a corresponding plurality of first candidate regions. Then, the information complexity of each first candidate region is calculated by the above formula 1. Next, the average value of the information complexity of each first candidate region is calculated to obtain the average value of the information complexity under the sliding window division of

[0092] Meanwhile, calculate the absolute value of the difference between the average value of the information complexity under the sliding window division of and the average value of the information complexity under the sliding window division of . If the absolute value of this difference is not greater than the preset difference threshold, it indicates that the current division situation is reasonable. Therefore, each first candidate region obtained by the sliding window division of

[0093] is determined as each image region of the target remote sensing image.

[0094] As an optional embodiment, S304 may specifically include:

[0095] Calculate the average value of the dark channel pixel values of each pixel point in each image region respectively to obtain the average dark channel pixel value of each image region;

[0096] Calculate the cosine similarity between each image region to obtain the highest structural similarity region corresponding to each image region;

[0097] Based on each highest structural similarity region, determine the corresponding similarity compensation value of each image region respectively;

[0098] For each image region, the following operations are respectively performed: determining a reflectance evaluation value of the image region based on the atmospheric transmittance of the image region, the average value of the dark channel pixels of the image region, the average value of the dark channel pixels of the region with the highest structural similarity, and the similarity compensation value.

[0099] In this embodiment, the region with the highest structural similarity is used to represent other image regions that are most structurally similar to the image region; the similarity compensation value is used to adjust the reflectance evaluation value to improve the accuracy of the reflectance evaluation value.

[0100] As an example, the server first calculates the average value of the dark channel pixel values of each pixel point in each image region, so as to obtain the average value of the dark channel pixels of each image region.

[0101] Then, for each image region, the following operations are respectively performed: calculating the cosine similarity between the image region and other image regions, and selecting the image region corresponding to the maximum cosine similarity as the region with the highest structural similarity of the image region.

[0102] Then, for each image region, according to its corresponding region with the highest structural similarity, calculate the similarity compensation value of the image region through the following formula 2:

[0103] , formula 2,

[0104] In formula 2, is used to represent the similarity compensation value of the i-th image region, is used to represent the cosine similarity between the i-th image region and its corresponding region with the highest structural similarity, is used to represent the minimum value of the cosine similarities between the region with the highest structural similarity and each image region other than the i-th image region, is used to represent the maximum value of the cosine similarities between the region with the highest structural similarity and each image region other than the i-th image region.

[0105] Among them, since the similarity degrees between different image regions and the region with the highest structural similarity are different, the similarity degrees between all the image regions and the region with the highest structural similarity are made dimensionless, and the dimensionless unified value is used as the compensation value to adjust the reflectance evaluation value.

[0106] Finally, based on the atmospheric transmittance of the image region, the average value of the dark channel pixels of the image region, the average value of the dark channel pixels of the region with the highest structural similarity, and the similarity compensation value, determine the reflectance evaluation value of each image region through the following formula 3:

[0107] , formula 3,

[0108] In Equation 3, is used to represent the reflectance evaluation value of the i-th image region, is used to represent the dark channel pixel mean value of the i-th image region, is used to represent the dark channel pixel mean value of the highest structural similarity region of the i-th image region, is used to represent the similarity compensation value of the i-th image region, is used to represent the atmospheric transmittance of the i-th image region.

[0109] Among them, the lower the dark channel pixel mean value of the i-th image region, the lower the object reflectance of the i-th image region, that is, the smaller the reflectance evaluation value; the closer the dark channel pixel mean value of the i-th image region is to the dark channel pixel mean value of its corresponding highest structural similarity region, the higher the object reflectance of the i-th image region, that is, the larger the reflectance evaluation value; the larger the atmospheric transmittance of the i-th image region, the higher the object reflectance, that is, the larger the reflectance evaluation value.

[0110] Through this embodiment, for each image region, based on the atmospheric transmittance of the image region, the dark channel pixel mean value of the image region, the dark channel pixel mean value of the highest structural similarity region, and the similarity compensation value, the reflectance evaluation value of the image region is determined respectively. In this way, the reflectance evaluation values of each image region in the target remote sensing image can be accurately determined, which can help to accurately determine the target dark pixels of the target remote sensing image and improve the effect of radiometric correction.

[0111] As an alternative embodiment, as Figure 4 shown, S103 may specifically include the following S401 to S403.

[0112] S401, determine a plurality of candidate dark pixels according to the reflectance evaluation values of each image region;

[0113] S402, determine the screening parameter corresponding to each candidate dark pixel based on the atmospheric transmittance of the image region corresponding to each candidate dark pixel;

[0114] S403, screen each candidate dark pixel according to the screening parameter corresponding to each candidate dark pixel to obtain the target dark pixel of the target remote sensing image.

[0115] In this embodiment, the screening parameter is used to screen the candidate dark pixels to obtain the target dark pixel of the target remote sensing image.

[0116] As an example, the server compares the reflectance evaluation value of each image region with the corresponding evaluation value threshold, and determines the minimum value of the dark channel pixel values in each image region smaller than the evaluation value threshold as the candidate dark pixel respectively.

[0117] Then, for each candidate dark pixel, according to the atmospheric transmittance of the corresponding image region, the corresponding screening parameter is determined through the following formula (4):

[0118] , formula (4),

[0119] In formula (4), is used to represent the screening parameter corresponding to the b-th candidate dark pixel, is used to represent the dark channel pixel value corresponding to the b-th candidate dark pixel. is used to represent the atmospheric transmittance of the i-th image region, and I is used to represent the number of image regions of the target remote sensing image. is used to represent the atmospheric transmittance of the -th image region corresponding to the b-th candidate dark pixel, and e is used to represent the calculation of the natural exponential function.

[0120] Among them, the closer the atmospheric transmittance of the image region corresponding to the candidate dark pixel is to the atmospheric transmittance of the target remote sensing image, the better the atmospheric correction effect of the candidate dark pixel on the target remote sensing image, that is, the smaller the screening parameter corresponding to the candidate dark pixel.

[0121] Finally, the screening parameters corresponding to each candidate dark pixel are compared with a preset screening threshold, and each candidate dark pixel smaller than the preset screening threshold is determined as the target dark pixel of the target remote sensing image.

[0122] Through this embodiment, based on the atmospheric transmittance of the image region corresponding to each candidate dark pixel, the screening parameter corresponding to each candidate dark pixel is determined; thus, according to the screening parameter corresponding to each candidate dark pixel, each candidate dark pixel is screened to obtain the target dark pixel of the target remote sensing image. In this way, by accurately determining the target dark pixel of the target remote sensing image, it helps to perform effective radiometric correction on the target remote sensing image and improve the radiometric correction effect.

[0123] As an alternative embodiment, S401 may specifically include:[[]]

[0124] Sort the reflectance evaluation values of each image region in ascending order to obtain a reflectance evaluation value sequence;

[0125] Determine the image regions corresponding to the smallest Z reflectance evaluation values in the reflectance evaluation value sequence as the second candidate regions, where Z is a positive integer;

[0126] Determine the average dark channel pixel value of each second candidate region as the candidate dark pixel.

[0127] In this embodiment, the second candidate region is used to represent the image region where the candidate dark pixel is located.

[0128] As an example, the server sorts the reflectivity evaluation values of all image regions in ascending or descending order to obtain a reflectivity evaluation value sequence.

[0129] Then, the first 10% of the image regions with the smallest reflectivity evaluation values in the reflectivity evaluation value sequence are determined as the second candidate regions. Finally, for each second candidate region, the average value of the dark channel pixel values of each pixel point in the second candidate region is calculated to obtain the average dark channel pixel value. And the average dark channel pixel values of each second candidate region are determined as the candidate dark pixels.

[0130] Through this embodiment, the image regions corresponding to the first Z reflectivity evaluation values with the smallest reflectivity evaluation values are determined as the second candidate regions. Then, the average dark channel pixel values of each second candidate region are determined as the candidate dark pixels. In this way, this embodiment pre - determines multiple candidate dark pixels through the magnitude relationship of the reflectivity evaluation values. In this way, in the follow - up, only the target dark pixels of the target remote sensing image need to be determined from the candidate dark pixels, thus improving the determination efficiency of the target dark pixels.

[0131] As an alternative embodiment, S403 may specifically include:

[0132] The image region corresponding to the candidate dark pixel with the smallest screening parameter is determined as the dark pixel region;

[0133] The minimum value of the dark channel pixel values of each pixel point in the dark pixel region is determined as the target dark pixel of the target remote sensing image.

[0134] In this embodiment, the dark pixel region is used to represent the image region where the target dark pixel is located.

[0135] As an example, the server selects the image region corresponding to the candidate dark pixel with the smallest screening parameter as the dark pixel region. Then, the minimum value among the dark channel pixel values of each pixel point in the dark pixel region is determined as the target dark pixel of the target remote sensing image.

[0136] Through this embodiment, the candidate dark pixels are screened according to the screening parameters corresponding to each candidate dark pixel to obtain the target dark pixel of the target remote sensing image. In this way, the target dark pixel of the target remote sensing image can be accurately determined, thus helping to improve the effect of radiometric correction.

[0137] Based on the radiometric correction method for remote sensing images. Correspondingly, the present invention also provides a specific embodiment of a radiometric correction system for remote sensing images.

[0138] Figure 5The figure shows a schematic structural diagram of a radiation correction system for remote sensing images provided by an embodiment of the present invention. The radiation correction system 500 for remote sensing images may include an image acquisition module 510, an image calculation module 520, a dark pixel determination module 530, and a radiation correction module 540.

[0139] The image acquisition module 510 is configured to acquire a dark channel image of a target remote sensing image. The dark channel image is an image composed of the minimum channel pixel values of each pixel point in the target remote sensing image, and the minimum channel pixel value is the minimum value among the channel pixel values of the pixel point in each color channel.

[0140] The image calculation module 520 is configured to determine the atmospheric transmittance and reflectivity evaluation value of each image region in the target remote sensing image according to the dark channel image. The reflectivity evaluation value is used to characterize the strength of the object reflectivity in the image region.

[0141] The dark pixel determination module 530 is configured to determine the target dark pixels of the target remote sensing image according to the atmospheric transmittance and reflectivity evaluation value of each image region in the target remote sensing image.

[0142] The radiation correction module 540 is configured to perform radiation correction on the target remote sensing image based on the target dark pixels to obtain a corrected remote sensing image.

[0143] In the radiation correction system for remote sensing images provided by the embodiment of the present invention, the atmospheric transmittance and reflectivity evaluation value of each image region in the target remote sensing image is determined according to the dark channel image of the target remote sensing image. Then, the target dark pixels of the target remote sensing image are determined according to the atmospheric transmittance and reflectivity evaluation value of each image region in the target remote sensing image. Finally, radiation correction is performed on the target remote sensing image based on the target dark pixels to obtain a corrected remote sensing image. In this way, the present invention comprehensively evaluates according to the reflectivity of the object in the target remote sensing image and the transmittance of the atmosphere, so as to accurately determine the target dark pixels of the target remote sensing image. Furthermore, effective radiation correction can be performed on the target remote sensing image to improve the effect of radiation correction.

[0144] It should be clear that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, the detailed description of known methods is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order between steps after understanding the spirit of the present invention.

[0145] It should also be noted that the exemplary embodiments mentioned in the present invention describe some methods or systems based on a series of steps or devices. However, the present invention is not limited to the order of the above steps, that is, the steps can be executed in the order mentioned in the embodiments, or different from the order in the embodiments, or several steps can be executed simultaneously.

[0146] As described above, the foregoing is only a specific implementation manner of the present invention. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, modules, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein. It should be understood that the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should all be covered within the protection scope of the present invention.

Claims

1. A radiation correction method for remote sensing images, characterized in that: The method comprises: Acquire a dark channel image of the target remote sensing image, wherein the dark channel image is an image composed of the minimum channel pixel value of each pixel point in the target remote sensing image, and the minimum channel pixel value is the minimum value of the channel pixel values ​​of the pixel point in each color channel; Determine, according to the dark channel image, an atmospheric transmittance and reflectance evaluation value of each image area in the target remote sensing image, wherein the reflectance evaluation value is used to characterize the intensity of the reflectance of the object in the image area; Determining target dark pixels of the target remote sensing image according to evaluation values ​​of atmospheric transmittance and reflectance of each image area in the target remote sensing image; Based on the target dark pixels, performing radiation correction on the target remote sensing image to obtain a corrected remote sensing image; Determining a reflectivity evaluation value of each of the image regions according to the atmospheric transmittance of each of the image regions and the dark channel pixel value of each of the pixel points in each of the image regions; Determining the reflectivity evaluation value of each image area according to the atmospheric transmittance of each image area and the dark channel pixel value of each pixel point in each image area includes: Calculating the mean of the dark channel pixel values ​​of each pixel point in each image area to obtain the mean of the dark channel pixel values ​​of each image area; Calculating the cosine similarity between the image regions to obtain the highest structural similarity region corresponding to each image region; Based on each of the highest structural similarity regions, respectively determine a similarity compensation value of each corresponding image region; For each of the image regions, respectively, the following steps are performed: determining a reflectivity evaluation value of the image region based on the atmospheric transmittance of the image region, the dark channel pixel mean of the image region, the dark channel pixel mean of the highest structural similarity region, and the similarity compensation value.

2. The radiation correction method for remote sensing images according to claim 1, characterized in that: The step of acquiring a dark channel image of a target remote sensing image comprises: Acquisition of remote sensing image data; Converting the remote sensing image data to obtain the target remote sensing image; A dark channel image of the target remote sensing image is determined according to the pixel value of each pixel point in the target remote sensing image.

3. The radiation correction method for remote sensing images according to claim 2, characterized in that: Determining the dark channel image of the target remote sensing image according to the pixel value of each pixel point in the target remote sensing image includes: Obtaining the channel pixel value of each pixel point in each color channel in the target remote sensing image; For each pixel point in the target remote sensing image, the minimum value among the pixel values ​​of each channel is determined as the dark channel pixel value of the pixel point; Based on the dark channel pixel value of each of the pixel points, a dark channel image of the target remote sensing image is constructed.

4. The radiation correction method for remote sensing images according to claim 1, characterized in that: Determining the atmospheric transmittance of each image area in the target remote sensing image according to the dark channel image includes: Performing mean processing on the dark channel pixel values ​​of each target pixel point in the dark channel image that meets the target screening condition to obtain an atmospheric illumination value, wherein the target screening condition is that the dark channel pixel value is greater than a preset pixel value threshold; Dividing the target remote sensing image into a plurality of image regions; Based on the atmospheric illumination value, the atmospheric transmittance of each of the image regions is determined.

5. The radiation correction method for remote sensing images according to claim 4, characterized in that: The step of dividing the target remote sensing image into a plurality of image regions comprises: by The target remote sensing image is divided by a sliding window to obtain a first information complexity mean, wherein the first information complexity mean is The average value of the information complexity of each first candidate region after the sliding window is divided, and the initial value of N is 3; by The target remote sensing image is divided by a sliding window to obtain a second information complexity mean, wherein the second information complexity mean is The mean value of the information complexity of each first candidate region after the sliding window is divided; When the absolute value of the difference between the first information complexity mean and the second information complexity mean is greater than a preset difference threshold, N is updated to N+2 and the first information complexity mean is updated to the second information complexity mean, and the loop is returned to execute. The target remote sensing image is divided by a sliding window to obtain a second information complexity mean; If the absolute value of the difference between the first information complexity mean and the second information complexity mean is not greater than the preset difference threshold, The first candidate regions obtained by the sliding window division are determined as the image regions of the target remote sensing image.

6. The radiation correction method for remote sensing images according to claim 1, characterized in that: Determining the target dark pixel of the target remote sensing image according to the atmospheric transmittance and reflectance evaluation values ​​of each image area in the target remote sensing image comprises: Determining a plurality of candidate dark pixels according to the reflectivity evaluation values ​​of the image regions; Determining the screening parameters corresponding to each of the candidate dark pixels based on the atmospheric transmittance of the image area corresponding to each of the candidate dark pixels; The candidate dark pixels are screened according to the screening parameters corresponding to the candidate dark pixels to obtain the target dark pixels of the target remote sensing image.

7. The radiation correction method for remote sensing images according to claim 6, characterized in that: Determining a plurality of candidate dark pixels according to the reflectivity evaluation values ​​of the image regions comprises: Sorting the reflectivity evaluation values ​​of the image regions in order of magnitude to obtain a reflectivity evaluation value sequence; Determine the image area corresponding to the first Z smallest reflectivity evaluation values ​​in the reflectivity evaluation value sequence as the second candidate area, where Z is a positive integer; The mean value of the dark channel pixels of each of the second candidate areas is determined as the candidate dark pixel.

8. The radiation correction method for remote sensing images according to claim 6, characterized in that: The step of screening each of the candidate dark pixels according to the screening parameters corresponding to each of the candidate dark pixels to obtain the target dark pixel of the target remote sensing image comprises: Determine the image region corresponding to the candidate dark pixel with the smallest screening parameter as the dark pixel region; The minimum value of the dark channel pixel values ​​of the pixels in the dark pixel area is determined as the target dark pixel of the target remote sensing image.

9. A radiation correction system for remote sensing images, characterized in that: The system comprises: An image acquisition module is used to acquire a dark channel image of a target remote sensing image, wherein the dark channel image is an image composed of the minimum channel pixel value of each pixel point in the target remote sensing image, and the minimum channel pixel value is the minimum value of the channel pixel values ​​of the pixel point in each color channel; An image calculation module, used to determine the atmospheric transmittance and reflectance evaluation values ​​of each image area in the target remote sensing image according to the dark channel image, wherein the reflectance evaluation value is used to characterize the intensity of the reflectance of the object in the image area; A dark pixel determination module, used for determining the target dark pixel of the target remote sensing image according to the atmospheric transmittance and reflectance evaluation values ​​of each image area in the target remote sensing image; A radiation correction module, used for performing radiation correction on the target remote sensing image based on the target dark pixel to obtain a corrected remote sensing image; Determining a reflectivity evaluation value of each of the image regions according to the atmospheric transmittance of each of the image regions and the dark channel pixel value of each of the pixel points in each of the image regions; Determining the reflectivity evaluation value of each image area according to the atmospheric transmittance of each image area and the dark channel pixel value of each pixel point in each image area includes: Calculating the mean of the dark channel pixel values ​​of each pixel point in each image area to obtain the mean of the dark channel pixel values ​​of each image area; Calculating the cosine similarity between the image regions to obtain the highest structural similarity region corresponding to each image region; Based on each of the highest structural similarity regions, respectively determine a similarity compensation value of each corresponding image region; For each of the image regions, respectively, the following steps are performed: determining a reflectivity evaluation value of the image region based on the atmospheric transmittance of the image region, the dark channel pixel mean of the image region, the dark channel pixel mean of the highest structural similarity region, and the similarity compensation value.

Images