A method, device, equipment and medium for detecting cloud and cloud shadows of remote sensing satellite

Abstract

The present application provides a cloud and cloud shadow detection method, device, equipment and medium for remote sensing satellites, wherein the method comprises: preprocessing the image to be detected to obtain a preprocessed first intermediate image; segmenting the first intermediate image in a target segmentation manner to obtain multiple second intermediate images; for each second intermediate image, determining the information amount ratio between the second intermediate image and the standard image according to the standard image corresponding to the second intermediate image and the information amount value of the second intermediate image; determining whether there are clouds and / or cloud shadows in the image to be detected according to the information amount ratio between each second intermediate image and the standard image. The effect of accurately detecting clouds and cloud shadows of remote sensing satellites is achieved.

Description

Technical Field

[0001] The present application relates to the field of remote sensing satellite image technology, and in particular to a cloud and cloud shadow detection method, device, equipment and medium for a remote sensing satellite. Background Art

[0002] Satellite remote sensing data is ground data collected by sensors on remote sensing satellites at high altitudes. These data contain rich geographical, environmental and resource information, and are of great significance for scientific research, environmental protection, resource investigation, disaster monitoring and other aspects.

[0003] Due to the influence of gases on the earth's surface, remote sensing data often contain interference factors such as clouds, fog, and cloud shadows, which block or weaken ground information to varying degrees and affect the availability of remote sensing data. In order to quickly find high-quality data that meets the needs in massive remote sensing data, it is necessary to automatically identify cloud and fog areas and cloud shadow areas through automatic detection methods and comprehensively evaluate data quality.

[0004] Currently, there are many methods for detecting clouds and cloud shadows from remote sensing satellites, which are usually implemented by using information such as spectrum, texture and spatial structure of remote sensing data itself, through threshold method, pattern recognition, machine learning, deep learning and other technologies. However, current methods for detecting clouds and cloud shadows from remote sensing satellites often only consider the data itself, and fail to make full use of the massive historical remote sensing data resources. Therefore, the available information is relatively limited and is easily affected by factors such as illumination, surface and atmosphere. There are certain limitations, and there are generally problems such as sensitivity to environmental conditions, need for human participation, and lack of algorithm versatility. Summary of the invention

[0005] In view of this, the purpose of the present application is to provide a cloud and cloud shadow detection method, device, equipment and medium for remote sensing satellites, which can determine the quality of the image to be detected based on the comparison of information values ​​between the standard image and the segmented remote sensing satellite image, and solve the problems that the available information in the prior art is relatively limited, and it is easily affected by factors such as lighting, surface, and atmosphere. There are certain limitations, and there are generally problems such as sensitivity to environmental conditions, need for human participation, and insufficient algorithm versatility, so as to achieve the effect of accurately detecting clouds and cloud shadows of remote sensing satellites.

[0006] In a first aspect, an embodiment of the present application provides a cloud and cloud shadow detection method for a remote sensing satellite, characterized in that the method includes: preprocessing the image to be detected to obtain a preprocessed first intermediate image; segmenting the first intermediate image in a target segmentation manner to obtain multiple second intermediate images; for each second intermediate image, determining an information ratio between the second intermediate image and the standard image according to a standard image corresponding to the second intermediate image and an information value of the second intermediate image; determining whether there are clouds and / or cloud shadows in the image to be detected according to the information ratio between each second intermediate image and the standard image.

[0007] Optionally, the first intermediate image is segmented in a target segmentation manner through the following steps: the first intermediate image is segmented according to the geographical range to obtain multiple third intermediate images, each third intermediate image has a geographical code; for each third intermediate image, a comparison image corresponding to the third intermediate image is determined according to the geographical code and resolution of the third intermediate image, and the comparison image is a standard image in a pre-established standard image library; according to the information volume of the third intermediate image and the information volume of the comparison image, it is judged whether the third intermediate image needs to be further segmented; if the third intermediate image does not need to be further segmented, the third intermediate image is determined to be the second intermediate image; if the third intermediate image needs to be further segmented, the third intermediate image is continued to be segmented until the difference between the information volume of the comparison image corresponding to the segmented image and the information volume of the segmented image is lower than the target difference or the segmented image is of the minimum size.

[0008] Optionally, the standard image is stored in a pre-established standard image library, wherein the method further comprises: establishing the standard image library according to pre-set rules.

[0009] Optionally, a standard image library is established through the following steps: acquiring multiple historical images of remote sensing satellites, wherein the historical images include multiple historical images at multiple deficiencies; screening out multiple target historical images from the multiple historical images, wherein the multiple target historical images are clear historical images; preprocessing the multiple target historical images at multiple deficiencies to obtain multiple target historical images at multiple deficiencies after preprocessing; framing the multiple target historical images based on preselected map framing standards of multiple different sizes to obtain multiple standard images of multiple different framing sizes; and establishing a standard image library based on the multiple standard images of the multiple different framing sizes.

[0010] Optionally, the method also includes: determining a ratio of clouds and cloud shadows in the image to be detected to the image to be detected based on an information amount ratio between a plurality of second intermediate images and the standard image; determining an overall quality index of the image to be detected based on the ratio of clouds and cloud shadows in the image to be detected to the image to be detected; and determining the quality of the image to be detected based on the overall quality index of the image to be detected.

[0011] Optionally, the information content ratio between the plurality of second intermediate images and the standard image is calculated by the following formula:

[0012]

[0013] Wherein, Q represents the information ratio between the second intermediate image and the standard image, A q represents the area of ​​the i-th second intermediate image, q i represents the information content of the i-th second intermediate image, A s represents the area of ​​the standard image corresponding to the i-th second intermediate image, s i represents the information amount of the standard image corresponding to the i-th second intermediate image, and n represents the number of second intermediate images.

[0014] In a second aspect, an embodiment of the present application further provides a cloud and cloud shadow detection device for a remote sensing satellite, the device comprising:

[0015] The image preprocessing module to be detected is used to preprocess the image to be detected to obtain a first intermediate image after preprocessing;

[0016] A first intermediate image segmentation module is used to segment the first intermediate image in a target segmentation manner to obtain a plurality of second intermediate images;

[0017] an information content ratio determination module, configured to determine, for each second intermediate image, an information content ratio between the second intermediate image and the standard image according to a standard image corresponding to the second intermediate image and an information content value of the second intermediate image;

[0018] The cloud and cloud shadow determination module is used to determine whether clouds and / or cloud shadows exist in the image to be detected based on the information amount ratio between each second intermediate image and the standard image.

[0019] Optionally, the standard image is stored in a pre-established standard image library, wherein the device further comprises: a standard image library establishment module, which is used to establish the standard image library according to pre-established rules.

[0020] In a third aspect, an embodiment of the present application also provides an electronic device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, the steps of the cloud and cloud shadow detection method for the remote sensing satellite as described above are performed.

[0021] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, executes the steps of cloud and cloud shadow detection by a remote sensing satellite as described above.

[0022] The cloud and cloud shadow detection method, device, equipment and medium for remote sensing satellites provided in the embodiments of the present application can determine the quality of the image to be detected based on the comparison of information values ​​between the standard image and the segmented remote sensing satellite image, so as to solve the problems existing in the prior art that the available information is relatively limited, and it is easily affected by factors such as lighting, surface, and atmosphere. There are certain limitations, and there are generally problems such as sensitivity to environmental conditions, need for human participation, and insufficient algorithm versatility, so as to achieve the effect of accurately detecting clouds and cloud shadows for remote sensing satellites.

[0023] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0025] Figure 1 A flow chart of a cloud and cloud shadow detection method for a remote sensing satellite provided in an embodiment of the present application;

[0026] Figure 2 A flowchart of another remote sensing satellite cloud and cloud shadow detection method provided in an embodiment of the present application;

[0027] Figure 3 A schematic diagram of the structure of a cloud and cloud shadow detection device for a remote sensing satellite provided in an embodiment of the present application;

[0028] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application usually described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work belongs to the scope of protection of the present application.

[0030] First, the application scenarios to which this application is applicable are introduced. This application can be applied in the field of remote sensing satellite imaging technology.

[0031] Research has found that satellite remote sensing data is ground data collected by sensors on remote sensing satellites at high altitudes. These data contain rich geographical, environmental and resource information, and are of great significance for scientific research, environmental protection, resource investigation, disaster monitoring and other aspects.

[0032] Due to the influence of gases on the earth's surface, remote sensing data often contain interference factors such as clouds, fog, and cloud shadows, which block or weaken ground information to varying degrees and affect the availability of remote sensing data. In order to quickly find high-quality data that meets the needs in massive remote sensing data, it is necessary to automatically identify cloud and fog areas and cloud shadow areas through automatic detection methods and comprehensively evaluate data quality.

[0033] Currently, there are many methods for detecting clouds and cloud shadows from remote sensing satellites, which are usually implemented by using information such as spectrum, texture and spatial structure of remote sensing data itself, through threshold method, pattern recognition, machine learning, deep learning and other technologies. However, current methods for detecting clouds and cloud shadows from remote sensing satellites often only consider the data itself, and fail to make full use of the massive historical remote sensing data resources. Therefore, the available information is relatively limited and is easily affected by factors such as illumination, surface and atmosphere. There are certain limitations, and there are generally problems such as sensitivity to environmental conditions, need for human participation, and lack of algorithm versatility.

[0034] Here, this application gives several existing cloud and cloud shadow detection algorithms for remote sensing satellites, (1) Threshold method: Identify clouds and cloud shadows by comparing the pixel values ​​in the remote sensing image with the preset threshold. The advantage of this algorithm is that the calculation is simple and efficient, but the selection of the threshold has great uncertainty, is greatly affected by the image quality and lighting conditions, and the detection accuracy is not high. (2) Texture recognition: Use the texture features of clouds to distinguish from the texture features of other ground backgrounds. However, the extraction and recognition of texture features are difficult, especially for natural surface types with insignificant textures. (3) Edge detection: Use edge detection technology (such as the Canny operator) to find the boundaries of clouds, and then classify them according to the size of the connected domain. However, this method is difficult to identify clouds with blurred edges and fog without clear edges. (4) Machine learning method: There is a support vector machine (SVM) method, which learns to distinguish between cloud and non-cloud areas through training samples. This type of method requires a large amount of labeled data, and the model has limited generalization ability, and is difficult to migrate to data of different types and sources. There is also a fuzzy mean clustering method, which uses fuzzy set theory to cluster remote sensing data to distinguish between cloud and non-cloud areas. This type of method does not require manual feature selection and has a high degree of automation, but poor stability. (5) Convolutional Neural Network (CNN): Use deep learning for feature extraction and classification. It has high recognition accuracy and strong adaptability, but the algorithm is complex and has poor interpretability. The training process requires a lot of computing resources and has a high overall cost. (6) Spectral feature combination: Use multi-spectral or full-spectral data and combine the differences between bands to identify clouds and cloud shadows. However, this method is more suitable for remote sensing data with rich bands. It is not effective for data with fewer bands and is prone to misjudgment and omission. (7) Comprehensive method: Integrate the results of multiple algorithms, such as combining texture, edge, spectrum, etc. to improve detection accuracy. However, this also leads to increased algorithm complexity and excessive resource consumption.

[0035] Based on this, an embodiment of the present application provides a cloud and cloud shadow detection method for a remote sensing satellite, which can determine the quality of the image to be detected based on a comparison of information values ​​between a standard image and a segmented remote sensing satellite image, thereby solving the problems in the prior art that the available information is relatively limited, is easily affected by factors such as lighting, surface, and atmosphere, has certain limitations, and is generally sensitive to environmental conditions, requires human participation, and has insufficient algorithm versatility, thereby achieving the effect of accurately detecting clouds and cloud shadows of remote sensing satellites.

[0036] See also Figure 1 , Figure 1 This is a flow chart of a cloud and cloud shadow detection method for a remote sensing satellite provided in an embodiment of the present application. Figure 1 As shown in, the cloud and cloud shadow detection method of the remote sensing satellite provided in the embodiment of the present application includes:

[0037] S101 , preprocessing the image to be detected to obtain a first intermediate image after preprocessing.

[0038] Here, the methods of preprocessing the image to be detected include: radiation correction, geometric correction, etc., so as to make it reflect the actual state of the surface.

[0039] S102: Segment the first intermediate image in a target segmentation manner to obtain a plurality of second intermediate images.

[0040] Specifically, the first intermediate image is segmented in a target segmentation manner through the following steps: the first intermediate image is segmented according to a geographic range to obtain a plurality of third intermediate images, each of which is provided with a geographic code; for each third intermediate image, a comparison image corresponding to the third intermediate image is determined according to the geographic code and resolution of the third intermediate image, the comparison image being a standard image in a pre-established standard image library; according to the information volume of the third intermediate image and the information volume of the comparison image, it is determined whether the third intermediate image needs to be further segmented; if the third intermediate image does not need to be further segmented, the third intermediate image is determined to be the second intermediate image; if the third intermediate image needs to be further segmented, the third intermediate image is continued to be segmented until the difference between the information volume of the comparison image corresponding to the segmented image and the information volume of the segmented image is lower than the target difference or the segmented image is of the minimum size.

[0041] Here, the segmentation of the first intermediate image is consistent with the segmentation method of the standard image library.

[0042] Furthermore, in order to improve algorithm efficiency, the present application adopts recursive level-by-level segmentation based on information volume calculation, that is, the segmentation from high level to low level is not done all at once, but whether to continue segmentation is determined level by level according to the amount of information.

[0043] Exemplarily, each sub-image formed by segmentation is uniformly encoded according to the standard, and high-quality images of corresponding levels, codes, and resolutions are selected from the standard image library to form matching image pairs with consistent geographical ranges and similar resolutions. At this time, it is necessary to calculate the amount of information of the image pairs separately, and then compare them. If the amount of information of the image pairs is similar, the segmentation will not continue. If there is a difference of more than 20% in the amount of information of the image pairs, the segmentation will continue to the next level, and the next level will be matched and calculated one by one. In this step-by-step manner, a group of difference image pairs with different levels and codes are finally locked. The areas covered by these image pairs are the areas where there are problems with image quality, and there is a significant lack of information in the image. After the initial extraction of the information missing area, the cloud fog, that is, the cloud shadow area, can be further identified.

[0044] It should be noted that the standard images are stored in a pre-established standard image library.

[0045] Wherein, the method further comprises: establishing a standard image library according to preset rules.

[0046] Specifically, a standard image library is established through the following steps: obtaining a plurality of historical images of a remote sensing satellite, wherein the historical images include a plurality of historical images at a plurality of definition levels; screening out a plurality of target historical images from the plurality of historical images, wherein the plurality of target historical images are clear historical images; preprocessing the plurality of target historical images at a plurality of definition levels to obtain a plurality of target historical images at a plurality of definition levels after preprocessing; framing the plurality of target historical images based on a plurality of preselected map framing standards of different sizes to obtain a plurality of standard images of a plurality of different framing sizes; and establishing a standard image library based on the plurality of standard images of the plurality of different framing sizes.

[0047] Exemplarily, before establishing a standard image library, it is necessary to select and update the materials of the standard image library, and it is necessary to select the visible light bands (red light, green light, blue light) and near-infrared bands of high-quality remote sensing images taken in history. Remote sensing images must be clear, free of clouds, fog and cloud shadows, and the shooting time should be as recent as possible, covering the entire area. The standard library uses high, medium and low resolution data, such as 2 meters, 10 meters, and 50 meters, or 1 meter, 10 meters, 100 meters, or other resolutions. In short, the resolution span of the standard library should basically cover the resolution range of common remote sensing data. New high-quality data can be supplemented at an appropriate frequency to ensure the timeliness of the standard image library. Since this patent adopts the core method of information measurement, it is less affected by seasons and light, and general surface changes are not easy to affect the detection and evaluation effects. Therefore, for most areas, it can be updated every 3-5 years. For areas where the ground state changes dramatically, the update frequency can be appropriately increased.

[0048] For example, the data in the standard image library needs to be corrected by radiation and geometry to reflect the actual state of the surface. The geographic grid is divided according to longitude and latitude, and the fully covered image is divided according to the geographic grid so that the image has a unified geographic coding. Here, the image segmentation and coding rules of the basic scale topographic map of my country can be referred to for image segmentation and coding.

[0049] According to the National Basic Scale Topographic Map Framing and Numbering (GB / T 13989-2012), the framing of 1:1000000 topographic maps adopts the international 1:1000000 map framing standard. The range of each 1:1000000 topographic map is 6 longitude difference and 4 latitude difference. The longitude difference is 12° and the latitude difference is 4° between latitudes 60° and 76°, and the longitude difference is 24° and the latitude difference is 4° between latitudes 76° and 88°.

[0050] The subsequent segmentation is based on the 1:1000000 topographic map, and the range is divided according to the specified longitude and latitude differences. The range of the 1:500000 topographic map is 3° longitude difference and 2° latitude difference; the range of the 1:250000 topographic map is 1°30' longitude difference and 1° latitude difference; the range of the 1:100000 topographic map is 30' longitude difference and 20' latitude difference; the range of the 1:50000 topographic map is 15' longitude difference and 10' latitude difference; the range of the 1:25000 topographic map is 7'30" longitude difference and 5' latitude difference; the range of the 1:10000 topographic map is 3'45" longitude difference and 2'30" latitude difference; the range of the 1:50000 topographic map is 15' longitude difference and 10' latitude difference; the range of the 1:25000 topographic map is 7'30" longitude difference and 5' latitude difference; the range of the 1:10000 topographic map is 3'45" longitude difference and 2'30" latitude difference; The range of 1:2000 topographic map is 37.5" longitude and 25" latitude; the range of 1:1000 topographic map is 18.75" longitude and 12.5" latitude; the range of 1:500 topographic map is 9.375" longitude and 6.25" latitude. As the coverage of clouds and cloud shadows is generally large, the image does not need to be segmented too finely. Segmentation according to 1:2000 topographic map can meet most of the cloud and cloud shadow detection needs.

[0051] The final standard image library should contain N standard images corresponding to 1:1000000 topographic maps, 4*N standard images corresponding to 1:500000 topographic maps, 16*N standard images corresponding to 1:250000 topographic maps, 144*N standard images corresponding to 1:100000 topographic maps, 576*N standard images corresponding to 1:50000 topographic maps, 2304*N standard images corresponding to 1:25000 topographic maps, 9216*N standard images corresponding to 1:10000 topographic maps, 36864*N standard images corresponding to 1:5000 topographic maps, 331776*N standard images corresponding to 1:2000 topographic maps, 1327104*N standard images corresponding to 1:1000 topographic maps, and 5308416*N standard images corresponding to 1:500 topographic maps.

[0052] Among them, each standard image should contain three types of high-quality images: high resolution, medium resolution, and low resolution.

[0053] S103. For each second intermediate image, determine an information quantity ratio between the second intermediate image and the standard image according to a standard image corresponding to the second intermediate image and an information quantity value of the second intermediate image.

[0054] Here, the information amount of the second intermediate image, the third intermediate image, or the standard image may be calculated in a variety of ways. This application provides several optional calculation methods, including:

[0055] (1) Image entropy. Image entropy reflects the average amount of information in an image and indicates the aggregation characteristics of the image grayscale distribution. Image entropy is expressed as the bit average of the image grayscale set.

[0056] (2) Number and density of feature points: Feature descriptors (such as SIFT, SURF, etc.) can be used to extract feature points in an image, and then the amount of information is calculated based on the number and distribution of feature points. The more feature points there are and the more evenly they are distributed, the greater the amount of information in the image. This is done by searching for key points (feature points) in different scale spaces and calculating the direction of the key points.

[0057] For example, SIFT (Scale-invariant feature transform) is a computer vision algorithm. It is used to detect and describe local features in images. It finds extreme points in the spatial scale and extracts their position, scale, and rotation invariants. The description and detection of local image features can help identify objects. SIFT features are based on some local appearance interest points on the object and are independent of the size and rotation of the image. The tolerance for light, noise, and slight changes in perspective is also quite high. The detection rate of partial object occlusion using SIFT feature description is also quite high, and even only more than 3 SIFT object features are needed to calculate the position and orientation. The essence of the SIFT algorithm is to find key points (feature points) in different scale spaces and calculate the direction of the key points. The key points found by SIFT are some very prominent points that will not change due to factors such as lighting, affine transformation and noise, such as corner points, edge points, bright spots in dark areas and dark spots in bright areas.

[0058] SURF (Speeded Up Robust Features) is an efficient variant of SIFT and also extracts scale-invariant features. The algorithm steps are roughly the same as the SIFT algorithm. The determinant value of the Hessian matrix is ​​used for feature point detection and the integral image is used to accelerate the operation. The descriptor of SURF is based on the 2D discrete wavelet transform response and effectively utilizes the integral image.

[0059] (3) Structural complexity calculation. Structural complexity can be calculated using a variety of methods. A common method is to use image texture feature methods to calculate structural complexity. This method can extract texture information from an image and evaluate the structural complexity of an image by analyzing the regularity, complexity, and frequency of the texture. In addition, methods such as wavelet transform can be used to decompose an image into different frequency sub-bands, and the structural complexity of an image can be evaluated by analyzing the energy distribution and frequency characteristics of different sub-bands. In addition, methods based on image edge information can also be used to calculate structural complexity, and the structural features of an image can be evaluated by analyzing the complexity of the image edge.

[0060] Complexity calculation based on texture. First, we need to extract the texture features of the image. We can use the LBP feature method or the grayscale co-occurrence matrix method. The LBP feature method divides the detection window into small areas (cells). For a pixel in each cell, the 8 points in its annular neighborhood are compared clockwise or counterclockwise. If the central pixel value is larger than the neighboring point, the neighboring point is assigned a value of 1, otherwise it is assigned a value of 0. In this way, each point will obtain an 8-bit binary number (usually converted to a decimal number). Then calculate the histogram of each cell, that is, the frequency of occurrence of each number (assuming it is a decimal number) (that is, a binary sequence of statistics on whether each pixel point is larger than the point in the neighborhood), and then normalize the histogram. Finally, connect the statistical histograms of each cell to obtain the LBP texture features of the entire image. The gray-level co-occurrence matrix defines a direction (orientation) and a step (step) in pixels for an image. The gray-level co-occurrence matrix T (N×N) is defined as the frequency at which pixels with gray levels i and j appear at the same point and at the point along the defined direction span step. N is the number of gray-level divisions. Then, it is necessary to analyze the local changes and spatial distribution of the texture and calculate the complexity of the texture. Based on the results of texture feature extraction, the entropy value of the texture image is calculated. The higher the entropy value, the higher the complexity of the texture, the higher the complexity of the image, and the greater the amount of information.

[0061] Complexity calculation based on wavelet transform. First, perform wavelet transform on the image. You can choose to use different wavelet basis functions (such as Haar, Daubechies, Symlets, etc.) to obtain approximate components (rough versions of approximate images) and detail components (describe the details of the image, such as edges and textures). Wavelet transform can decompose the image into multiple frequency subbands, each with different spatial frequency and directional characteristics. By analyzing the frequency characteristics of each subband, the structural complexity of the image can be understood. Find the subband containing a lot of edge information in the subband of the standard image, compare it with the corresponding subband of the image to be detected, and compare the intensity of the edge information. If the intensity is similar, it means that the complexity of the image pair is basically the same. This method can eliminate the influence of abnormal noise on complexity calculation and has higher reliability.

[0062] Complexity calculation based on edge detection. First, perform edge detection on the image. You can use edge detection algorithms such as Sobel and Canny. These algorithms can detect edges in an image and generate an edge image. Extract the number and intensity of edges from the edge image, or calculate the ratio of the number of pixels in the edge image to the number of pixels in the entire image to represent the edge density in the image and reflect the complexity and information content of the image.

[0063] S104: Determine whether clouds and / or cloud shadows exist in the image to be detected according to the information amount ratio between each second intermediate image and the standard image.

[0064] Here, if the information ratio is close to 1, it is determined that the information ratio between each second intermediate image and the standard image is basically the same, indicating that the area is basically cloudless and shadowless. If the information ratio is abnormally low, it means that there are cloud, fog and / or cloud shadow coverage problems in the area.

[0065] Optionally, the method also includes: determining a ratio of clouds and cloud shadows in the image to be detected to the image to be detected based on an information amount ratio between a plurality of second intermediate images and the standard image; determining an overall quality index of the image to be detected based on the ratio of clouds and cloud shadows in the image to be detected to the image to be detected; and determining the quality of the image to be detected based on the overall quality index of the image to be detected.

[0066] The information content ratio between the plurality of second intermediate images and the standard image can be calculated by the following formula:

[0067]

[0068] Wherein, Q represents the information ratio between the second intermediate image and the standard image, A q represents the area of ​​the i-th second intermediate image, q irepresents the information content of the i-th second intermediate image, A s represents the area of ​​the standard image corresponding to the i-th second intermediate image, s i represents the information amount of the standard image corresponding to the i-th second intermediate image, and n represents the number of second intermediate images.

[0069] For example, if the amount of information of the image to be detected is less than 30% of the information of the standard image, it is judged to have thick clouds; if the amount of information of the image to be detected is greater than 30% of the information of the standard image and less than 60% of the information of the standard image, it is judged to have thin clouds or thick cloud shadows; if the amount of information of the image to be detected is greater than 60% of the information of the standard image and less than 90% of the information of the standard image, it is judged to have fog or thin cloud shadows; if the amount of information of the image to be detected is greater than 90% of the entropy value of the standard image, it is judged to be a high-quality image. Combined with the image pair encoding, based on the above standards, four detection results data of thick cloud area, thin cloud area, fog area and high-quality area can be generated respectively.

[0070] In this way, the quality of the image taken by the remote sensing satellite is determined by the ratio between the amount of information of the image to be detected and the amount of information of the standard image.

[0071] For example, see Figure 2 , Figure 2 A flowchart of another remote sensing satellite cloud and cloud shadow detection method provided in an embodiment of the present application. Another remote sensing satellite cloud and cloud shadow detection method provided in an embodiment of the present application includes:

[0072] S201 , segment a first intermediate image into M levels.

[0073] Here, M levels are the number of times the first intermediate image is divided, and dividing the first intermediate image once is 1 level.

[0074] S202. Determine a standard image corresponding to the first intermediate image.

[0075] S203: Calculate the information amount of the first intermediate image and the standard image.

[0076] S204: Determine whether the information amount of the first intermediate image is similar to the information amount of the standard image

[0077] If the information content of the first intermediate image is similar to that of the standard image, then step S205 is executed to analyze the clouds and cloud shadows.

[0078] S206: Determine the quality of the image to be detected.

[0079] If the information amount of the first intermediate image is not similar to the information amount of the standard image, step S207 is executed and the level M is increased by 1.

[0080] Here, the segmentation of the first intermediate image, the calculation method of the information amount and the calculation method of the quality of the image to be detected are the same as those in Figure 1 The descriptions of the segmentation of the first intermediate image, the method for calculating the amount of information, and the method for calculating the quality of the image to be detected are the same or similar and will not be repeated here.

[0081] The cloud and cloud shadow detection method for remote sensing satellites provided in the embodiments of the present application can determine the quality of the image to be detected based on the comparison of information values ​​between the standard image and the segmented remote sensing satellite image, thereby solving the problems in the prior art that the available information is relatively limited and is easily affected by factors such as lighting, surface, and atmosphere. It has certain limitations and is generally sensitive to environmental conditions, requires human participation, and has insufficient algorithm versatility, thereby achieving the effect of accurately detecting clouds and cloud shadows for remote sensing satellites.

[0082] See also Figure 3 , Figure 3 This is a schematic diagram of the structure of a cloud and cloud shadow detection device for a remote sensing satellite provided in an embodiment of the present application. Figure 3 As shown in , the cloud and cloud shadow detection device 300 of the remote sensing satellite includes:

[0083] The image preprocessing module 301 is used to preprocess the image to be detected to obtain a first intermediate image after preprocessing;

[0084] A first intermediate image segmentation module 302 is used to segment the first intermediate image in a target segmentation manner to obtain a plurality of second intermediate images;

[0085] An information content ratio determination module 303 is used to determine, for each second intermediate image, an information content ratio between the second intermediate image and the standard image according to the standard image corresponding to the second intermediate image and the information content value of the second intermediate image;

[0086] The cloud and cloud shadow determination module 304 is used to determine whether clouds and / or cloud shadows exist in the image to be detected according to the information amount ratio between each second intermediate image and the standard image.

[0087] Optionally, the standard image is stored in a pre-established standard image library, wherein the device further comprises: a standard image library establishment module, which is used to establish the standard image library according to pre-established rules.

[0088] The cloud and cloud shadow detection device for a remote sensing satellite provided in the embodiment of the present application can determine the quality of the image to be detected based on the comparison of information values ​​between a standard image and a segmented remote sensing satellite image, thereby solving the problems in the prior art that the available information is relatively limited, and the device is easily affected by factors such as illumination, surface, and atmosphere, and has certain limitations. In addition, the device is generally sensitive to environmental conditions, requires human participation, and has insufficient algorithm versatility, thereby achieving the effect of accurately detecting clouds and cloud shadows for remote sensing satellites.

[0089] See also Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 4 As shown in , the electronic device 400 includes a processor 410 , a memory 420 and a bus 440 .

[0090] The memory 420 stores machine-readable instructions executable by the processor 410. When the electronic device 400 is running, the processor 410 communicates with the memory 420 via the bus 440. When the machine-readable instructions are executed by the processor 410, the above-mentioned Figure 1 The specific implementation of the steps of the method for detecting clouds and cloud shadows of a remote sensing satellite in the method embodiment shown can be found in the method embodiment, and will not be described in detail here.

[0091] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1 The specific implementation of the steps of the method for detecting clouds and cloud shadows of a remote sensing satellite in the method embodiment shown can be found in the method embodiment, and will not be described in detail here.

[0092] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0093] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0094] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0095] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0096] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application can essentially or in other words, the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0097] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A cloud and cloud shadow detection method for a remote sensing satellite, characterized in that: The method comprises: Preprocessing the image to be detected to obtain a first intermediate image after preprocessing; Segmenting the first intermediate image in a target segmentation manner to obtain a plurality of second intermediate images; For each second intermediate image, determining an information content ratio between the second intermediate image and the standard image according to a standard image corresponding to the second intermediate image and an information content value of the second intermediate image; determining whether there are clouds and / or cloud shadows in the image to be detected according to an information amount ratio between each second intermediate image and the standard image; Wherein, the method further comprises: Determining a ratio of clouds and cloud shadows in the image to be detected to the image to be detected according to an information amount ratio between the plurality of second intermediate images and the standard image; Determining an overall quality index of the image to be detected according to a ratio of clouds and cloud shadows in the image to be detected to the image to be detected; Determining the quality of the image to be detected according to the overall quality index of the image to be detected; The information content ratio between the plurality of second intermediate images and the standard image is calculated by the following formula: Wherein, Q represents the information ratio between the second intermediate image and the standard image, A q represents the area of ​​the i-th second intermediate image, q i represents the information content of the i-th second intermediate image, A s represents the area of ​​the standard image corresponding to the i-th second intermediate image, s i represents the information amount of the standard image corresponding to the i-th second intermediate image, and n represents the number of second intermediate images.

2. The method according to claim 1, characterized in that The first intermediate image is segmented in a target segmentation manner by the following steps: Segmenting the first intermediate image according to the geographical range to obtain a plurality of third intermediate images, each of which has a geographical code; For each third intermediate image, determining a comparison image corresponding to the third intermediate image according to the geocoding and resolution of the third intermediate image, wherein the comparison image is a standard image in a pre-established standard image library; determining whether to continue segmenting the third intermediate image according to the information amount of the third intermediate image and the information amount of the comparison image; If it is not necessary to further segment the third intermediate image, determining the third intermediate image as the second intermediate image; If the third intermediate image needs to be further segmented, the third intermediate image is further segmented until the difference between the amount of information of the contrast image corresponding to the segmented image and the amount of information of the segmented image is lower than the target difference or the segmented image reaches the minimum size.

3. The method according to claim 1, characterized in that Standard images are stored in a pre-established standard image library. Wherein, the method further comprises: Establish a standard image library based on pre-set rules.

4. The method according to claim 1, characterized in that: Create a standard image library by following these steps: Acquire a plurality of historical images of a remote sensing satellite, wherein the historical images include a plurality of historical images at a plurality of resolutions; Screening out a plurality of target historical images from a plurality of historical images, wherein the plurality of target historical images are cleaned historical images; Preprocessing a plurality of target historical images at various resolutions to obtain preprocessed target historical images at various resolutions; Based on a plurality of pre-selected map framing standards of different sizes, a plurality of target historical images are framed to obtain a plurality of standard images of a plurality of different framing sizes; A standard image library is established based on multiple standard images of different frame sizes.

5. A cloud and cloud shadow detection device for a remote sensing satellite, characterized in that: The device comprises: The image preprocessing module to be detected is used to preprocess the image to be detected to obtain a first intermediate image after preprocessing; A first intermediate image segmentation module is used to segment the first intermediate image in a target segmentation manner to obtain a plurality of second intermediate images; an information content ratio determination module, configured to determine, for each second intermediate image, an information content ratio between the second intermediate image and the standard image according to a standard image corresponding to the second intermediate image and an information content value of the second intermediate image; a cloud and cloud shadow determination module, configured to determine whether there are clouds and / or cloud shadows in the image to be detected according to the information amount ratio between each second intermediate image and the standard image; The information amount ratio determination module is further used to determine the ratio of clouds and cloud shadows in the image to be detected to the image to be detected according to the information amount ratio between the plurality of second intermediate images and the standard image; determine the overall quality index of the image to be detected according to the ratio of clouds and cloud shadows in the image to be detected to the image to be detected; and determine the quality of the image to be detected according to the overall quality index of the image to be detected; The information content ratio between the plurality of second intermediate images and the standard image is calculated by the following formula: Wherein, Q represents the information ratio between the second intermediate image and the standard image, A q represents the area of ​​the i-th second intermediate image, q i represents the information content of the i-th second intermediate image, A s represents the area of ​​the standard image corresponding to the i-th second intermediate image, s i represents the information amount of the standard image corresponding to the i-th second intermediate image, and n represents the number of second intermediate images.

6. The device according to claim 5, characterized in that Standard images are stored in a pre-established standard image library. Wherein, the device further comprises: The standard image library establishment module is used to establish a standard image library according to pre-set rules.

7. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory via the bus, and the processor executes the machine-readable instructions to perform the steps of the cloud and cloud shadow detection method for a remote sensing satellite as claimed in any one of claims 1 to 4.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the cloud and cloud shadow detection method for a remote sensing satellite as claimed in any one of claims 1 to 4 are executed.

Images