An image processing method and system for detailed marine land space planning

Through pre-training model and brightness reference table technology, the clarity of satellite remote sensing images is improved and segmented and combined, which solves the high cost and cloud impact of obtaining high-definition remote sensing images in detailed marine land space planning, achieving efficient and low-cost Image processing effect.

CN118628370BActive Publication Date: 2024-11-12SOUTH CHINA SEA PLANNING & ENVIRONMENT RES INST SOA
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411092613.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-09
Publication Date
2024-11-12
Estimated Expiration
2044-08-09

AI Technical Summary

Technical Problem

The prior art is costly to obtain high-definition remote sensing images in detailed planning of marine land space, and it is difficult to obtain large-scale high-definition images due to cloud layers.

Method used

The pre-trained model is used to improve the clarity of satellite remote sensing images, and the image color is standardized through the brightness reference table, and the high-definition fourth satellite remote sensing images are generated, reducing costs and reducing cloud impact.

Benefits of technology

It realizes the clarity and visibility of remote sensing images while reducing costs, and reduces the difficulty of acquiring high-definition remote sensing images.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118628370B_ABST
    Figure CN118628370B_ABST
Patent Text Reader

Abstract

The present invention belongs to the field of image processing technology, and discloses an image processing method and system applied to detailed planning of marine land space, the method comprising: obtaining N1 first satellite remote sensing images of a to-be-developed marine area, extracting any first satellite remote sensing image, inputting any first satellite remote sensing image into a pre-training model, generating a second satellite remote sensing image, traversing all first satellite remote sensing images, extracting any second satellite remote sensing image, standardizing the color of any second satellite remote sensing image based on a preset brightness reference table, generating a third satellite remote sensing image, traversing all second satellite remote sensing images, generating a fourth satellite remote sensing image of the to-be-developed marine area based on all third satellite remote sensing images, dividing the fourth satellite remote sensing image into marine units and land units, and sending the fourth satellite remote sensing image to a user terminal. The technical solution of the present invention can effectively improve the clarity of remote sensing images, and has low cost and good timeliness.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of image processing technology, and specifically relates to an image processing method and system applied to detailed marine land space planning. Background Art

[0002] When carrying out detailed marine land space planning for a sea area, it is necessary to extract information from high-definition remote sensing image data to obtain spatial distribution information on the current status of marine protection and utilization. However, there are several problems in obtaining remote sensing image data: the cost of high-definition remote sensing image data is relatively high; due to the influence of clouds, it is difficult to obtain high-definition images of a large area in a specific period.

[0003] It is an important research topic to obtain high-definition remote sensing images that can fully and accurately reflect the actual situation of marine land space at a relatively low cost. Similar prior art includes a Chinese patent application with publication number CN117934518A, which discloses a remote sensing image segmentation method and system for detailed marine land space planning, obtaining satellite remote sensing images of the undeveloped ocean and several local remote sensing images of the undeveloped ocean; obtaining all target feature points of each local remote sensing image of the undeveloped ocean according to the significance of each feature point of each local remote sensing image of the undeveloped ocean; obtaining the target matching scale of each local area according to the target degree of each local area at each scale; and making detailed planning of marine resources according to the clear satellite remote sensing images of the undeveloped ocean. There is also a Chinese patent application with publication number CN112393714A, which discloses an image correction method based on the fusion of drone aerial photography and satellite remote sensing. It obtains n photos taken by a drone in different directions in the power transmission and transformation tower area and aligns them, and further stitches them to generate a photo grid; calculates the elevation of each pixel in the photo grid to obtain the drone orthorectified image, and performs preprocessing; fuses the panchromatic orthophoto image and the orthophoto multispectral image to obtain a fused satellite remote sensing image M; constructs and updates a sample library; takes the fused satellite remote sensing image M as the reference image, and uses the position coordinates of the feature points and correction points corresponding to the satellite remote sensing image M in the sample library to correct the position coordinates of the feature points and correction points corresponding to the drone orthorectified image to obtain a corrected drone image.

[0004] The above two methods fuse high-definition local remote sensing images into large-scale satellite remote sensing images to improve the clarity of satellite remote sensing images, but the cost of collecting local remote sensing images by drones is still high. Therefore, it is an urgent problem to provide an image processing method and system for detailed marine land space planning to obtain high-definition remote sensing images while reducing costs. Summary of the invention

[0005] In response to the above-mentioned technical problems, the present invention provides an image processing method and system for detailed planning of marine land space.

[0006] In a first aspect, the present invention provides an image processing method for marine land space detailed planning, the method comprising the following steps:

[0007] Step 1, obtaining N1 first satellite remote sensing images of the ocean area to be developed, and obtaining imaging information of each first satellite remote sensing image;

[0008] Step 2: extract any first satellite remote sensing image, input any first satellite remote sensing image into a pre-training model, and generate a second satellite remote sensing image, wherein the clarity index value of the second satellite remote sensing image is greater than the second clarity index value of any first satellite remote sensing image;

[0009] Step 3: after traversing all the first satellite remote sensing images, extract any second satellite remote sensing image, standardize the color of any second satellite remote sensing image based on a preset brightness reference table, and generate a third satellite remote sensing image;

[0010] Step 4: After traversing all the second satellite remote sensing images, segment and combine all the third satellite remote sensing images to generate a fourth satellite remote sensing image of the ocean to be developed;

[0011] Step 5: Divide the fourth satellite remote sensing image into ocean units and land units, and send the result to the user terminal.

[0012] Specifically, in step 2, the training method of the pre-trained model is:

[0013] Step 211: Acquire M first satellite images, wherein any first satellite image has the same first clarity index value, and the first clarity index value is greater than the second clarity index value of any first satellite remote sensing image;

[0014] Step 212: extract any first satellite image, and perform degradation processing on any first satellite image according to the preset K third clarity index values ​​to generate K second satellite images, wherein one second satellite image corresponds to one third clarity index value;

[0015] Step 213: after traversing all the first satellite images, K training data sets are generated based on the clarity index values, wherein any training data set includes M training data pairs, and any training data pair is any first satellite image and a second satellite image whose clarity index value corresponding to any first satellite image is a fourth clarity index value, and the fourth clarity index value is a clarity index value corresponding to any training data set;

[0016] Step 214: extract any training data set, use any training data set to train a neural network model, and generate a pre-trained model corresponding to any training data set;

[0017] Step 215: Generate a clarity index range based on the clarity index value corresponding to any training data set, and store the clarity index range and the pre-trained model corresponding to any training data set in the first storage module in correspondence.

[0018] Specifically, in step 2, any first satellite remote sensing image is input into the pre-trained model to generate a second satellite remote sensing image, including:

[0019] Step 221, obtaining a second clarity index value of any first satellite remote sensing image;

[0020] Step 222: searching the first storage module based on the second clarity index value to obtain a first pre-trained model corresponding to the second clarity index value;

[0021] Step 223: Input any first satellite remote sensing image into the first pre-trained model to generate a second satellite remote sensing image.

[0022] Specifically, in step 3, the color of any second satellite remote sensing image is standardized based on a preset brightness reference table to generate a third satellite remote sensing image, including:

[0023] Step 311, extracting all ground objects in any second satellite remote sensing image, and classifying ground objects of the same type into the same ground object set;

[0024] Step 312: extract any set of ground objects and generate a first brightness distribution histogram of any color component of any set of ground objects;

[0025] Step 313: Calculate the first average brightness value of any color component in any set of ground objects, and generate the first pixel statistics of any set of ground objects based on the first average brightness values ​​of all color components;

[0026] Step 314: traverse the pixel statistics table based on the first pixel statistics, extract the second pixel statistics value with the greatest similarity to the first pixel statistics from the pixel statistics table, and determine whether the maximum similarity value is greater than or equal to a first preset value. If so, proceed to step 315; if not, proceed to step 316.

[0027] Step 315: obtaining a first identifier corresponding to the second pixel statistical value, searching a brightness reference table based on the first identifier, extracting a second brightness distribution histogram of any color component corresponding to the first identifier, and using the second brightness distribution histogram to calibrate the first brightness distribution histogram corresponding to any color component, so as to standardize the color of the ground objects in any ground object set;

[0028] Step 316: Generate a unique identifier for the first pixel statistical value, update the pixel statistical table based on the unique identifier and the first pixel statistical value, and update the brightness reference table based on the unique identifier and a first brightness distribution histogram of any color component;

[0029] Step 317: After traversing all sets of ground objects in any second satellite remote sensing image, output a third satellite remote sensing image.

[0030] Specifically, after step 3, the following steps are also included:

[0031] Step 321: select any third satellite remote sensing image, extract all ground objects in any third satellite remote sensing image, and classify the ground objects of the same type into the same adjusted ground object set;

[0032] Step 322: extract any adjustment object set, and generate a third brightness distribution histogram of any color component of any adjustment object set;

[0033] Step 323: Calculate the second average brightness value of any color component in any set of adjusted objects, and generate a third pixel statistic value of any set of adjusted objects based on the second average brightness values ​​of all color components;

[0034] Step 324: traverse the pixel statistics table based on the third pixel statistics value, extract the fourth pixel statistics value with the greatest similarity to the third pixel statistics value from the pixel statistics table, and further obtain the second identifier corresponding to the fourth pixel statistics value;

[0035] Step 325: Generate a fifth pixel statistical value based on the third pixel statistical value and the fourth pixel statistical value, and store the fifth pixel statistical value in a pixel statistical table corresponding to the second identifier;

[0036] Step 326, search the brightness reference table based on the second identifier, extract the fourth brightness distribution histogram of any color component corresponding to the second identifier, generate the fifth brightness distribution histogram corresponding to any color component based on the third brightness distribution histogram and the fourth brightness distribution histogram corresponding to any color component, and store the fifth brightness distribution histogram corresponding to any color component in the brightness reference table corresponding to the second identifier.

[0037] Specifically, step 4 includes:

[0038] Step 41: All third satellite remote sensing images are superimposed and placed based on the latitude and longitude information to generate a superimposed image, and then a circumscribed rectangle of the superimposed image is drawn, and the side length of the circumscribed rectangle is obtained;

[0039] Step 42: Generate segmentation parameters based on the first preset ratio and the side length, and divide the circumscribed rectangle into N2 rectangular areas based on the segmentation parameters;

[0040] Step 43, extract any rectangular area, combine and stitch the multiple layers of satellite remote sensing images in any rectangular area, and generate N3 regional stitched images, wherein the coverage area of ​​any regional stitched image overlaps with the coverage area of ​​the superimposed image in any rectangular area;

[0041] Step 44, extract any regional stitching image, determine whether the area covered by any regional stitching image is land or ocean, or contains both land and ocean. If it is land, define the ratio of the cloud area to the land area as the first cloud ratio of any regional stitching image. If it is ocean, define the ratio of the cloud area to the ocean area as the second cloud ratio of any regional stitching image. If it contains both land and ocean, extract the first land area and the first ocean area in any regional stitching image based on the pixel value, calculate the first ratio of the first cloud area to the first land area, and the second ratio of the second cloud area to the first ocean area. Calculate the third cloud ratio of any regional stitching image based on the first ratio and the second ratio. The calculation formula of the third cloud ratio RC3 is:

[0042] ,

[0043] in, and is the parameter weight coefficient, R1 is the first ratio, R2 is the second ratio, , the first cloud region is the cloud region above the first land region, and the second cloud region is the cloud region above the first ocean region;

[0044] Step 45: after traversing all the regional stitching images, the regional stitching image corresponding to the minimum cloud ratio is used as the regional image to be stitched corresponding to any rectangular area, wherein the cloud ratio is the first cloud ratio, the second cloud ratio or the third cloud ratio;

[0045] Step 46: After traversing all rectangular areas, extract all images of the area to be stitched, and stitch all images of the area to be stitched to generate a pending satellite remote sensing image;

[0046] Step 47, extracting the second land area and the second ocean area of ​​the pending satellite remote sensing image based on the pixel value, defining the ratio of the third cloud area to the second land area as the fourth cloud ratio of the pending satellite remote sensing image, and judging whether the fourth cloud ratio is less than or equal to the second preset value, if not, proceeding to step 48, and if so, proceeding to step 49, wherein the third cloud area is the cloud area above the second land area;

[0047] Step 48, reset the first preset ratio, and then return to step 42;

[0048] Step 49: output the pending satellite remote sensing image as the fourth satellite remote sensing image.

[0049] Specifically, in step 48, the first preset ratio is reset to reduce the first preset ratio according to the second preset ratio.

[0050] Specifically, before step 3, all second satellite remote sensing images are subjected to defogging, radiation correction and geometric correction.

[0051] In a second aspect, the present invention also provides an image processing system for marine land space detailed planning, the system comprising: an image acquisition module, a clarity adjustment module, a standardization processing module, an image synthesis module and an image division module;

[0052] An image acquisition module is used to acquire N1 first satellite remote sensing images of the ocean area to be developed, and to acquire imaging information of each first satellite remote sensing image;

[0053] A clarity adjustment module is used to extract any first satellite remote sensing image, input any first satellite remote sensing image into a pre-training model, and generate a second satellite remote sensing image, wherein the clarity index value of the second satellite remote sensing image is greater than the second clarity index value of any first satellite remote sensing image;

[0054] A standardization processing module is used to extract any second satellite remote sensing image after traversing all first satellite remote sensing images, and perform color standardization processing on any second satellite remote sensing image based on a preset brightness reference table to generate a third satellite remote sensing image;

[0055] An image synthesis module is used to segment and combine all the third satellite remote sensing images after traversing all the second satellite remote sensing images to generate a fourth satellite remote sensing image of the ocean to be developed;

[0056] The image division module is used to divide the fourth satellite remote sensing image into ocean units and land units, and send them to the user terminal.

[0057] In a third aspect, the present invention provides a computer storage medium storing program instructions, wherein when the program instructions are executed, the device where the computer storage medium is located is controlled to execute any one of the above-mentioned image processing methods applied to the detailed planning of marine land space.

[0058] The present invention discloses an image processing method and system for detailed marine land space planning, pre-training a pre-training model for improving image clarity, inputting a first satellite remote sensing image into the pre-training model to obtain a second satellite remote sensing image with improved clarity, then standardizing the colors of all second satellite remote sensing images based on a preset brightness reference table to unify the colors of each object in each second satellite remote sensing image, and finally obtaining a fourth satellite remote sensing image by segmenting and combining all third satellite remote sensing images to reduce the impact of clouds on satellite remote sensing images of marine areas to be developed, thereby improving the visibility of marine areas to be developed. According to the technical solution of the present invention, the cost of obtaining high-definition remote sensing images is further reduced, while the difficulty of obtaining high-definition images of a large range in a specific period is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0060] Figure 1 A flowchart of an image processing method for marine land space detailed planning according to the present invention;

[0061] Figure 2 A schematic diagram of a regional stitching image according to an embodiment of the present invention;

[0062] Figure 3 A modular schematic diagram of an image processing system for detailed marine land space planning according to the present invention. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical scheme and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. Obviously, the specific embodiments described herein are only used to explain the present invention and are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention.

[0064] It should be noted that if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0065] Figure 1 FIG. 1 is a flow chart of an embodiment of an image processing method for marine land space detailed planning provided by the present invention, and the flow chart specifically includes the following steps:

[0066] Step 1: Obtain N1 first satellite remote sensing images of the ocean area to be developed, and obtain imaging information of each first satellite remote sensing image.

[0067] Urban and rural construction is changing with each passing day. In order to ensure the accuracy of marine land space planning, the above N1 first satellite remote sensing images are satellite remote sensing images taken within the preset time. N1 is a positive integer greater than 1.

[0068] Exemplarily, the imaging information includes the shooting time of the first satellite remote sensing image, the shooting satellite number, the shooting angle, and the latitude and longitude coordinates of the first satellite remote sensing image.

[0069] Step 2: extract any first satellite remote sensing image, input any first satellite remote sensing image into a pre-trained model, and generate a second satellite remote sensing image, wherein the clarity index value of the second satellite remote sensing image is greater than the second clarity index value of any first satellite remote sensing image.

[0070] Clarity refers to the point density of pixels. The greater the point density of pixels, the finer the satellite remote sensing image and the higher the degree of realism.

[0071] Specifically, in step 2, the training method of the pre-trained model is:

[0072] Step 211: Acquire M first satellite images, wherein any first satellite image has the same first clarity index value, and the first clarity index value is greater than the second clarity index value of any first satellite remote sensing image.

[0073] Step 212: extract any first satellite image, and perform degradation processing on any first satellite image according to the preset K third clarity index values ​​to generate K second satellite images, wherein one second satellite image corresponds to one third clarity index value.

[0074] Step 213: After traversing all the first satellite images, K training data sets are generated based on the clarity index values, wherein any training data set includes M training data pairs, and any training data pair is any first satellite image, and a second satellite image whose clarity index value corresponding to any first satellite image is a fourth clarity index value, and the fourth clarity index value is a clarity index value corresponding to any training data set.

[0075] Step 214: extract any training data set, use any training data set to train a neural network model, and generate a pre-trained model corresponding to any training data set.

[0076] Step 215: Generate a clarity index range based on the clarity index value corresponding to any training data set, and store the clarity index range and the pre-trained model corresponding to any training data set in the first storage module in correspondence.

[0077] The values ​​of M and K are set according to the experience of those skilled in the art or according to actual application scenarios, and are not limited in this regard in the embodiments of the present application.

[0078] Exemplarily, the first clarity index value is R, the value of K is 5, the K third clarity index values ​​are 0.8R, 0.7R, 0.6R, 0.5R and 0.4R respectively, the value of M is 3, and the first satellite images are M1, M2 and M3 respectively. For the satellite image M1, five degradation processes are performed respectively to obtain five second satellite images M11, M12, M13, M14 and M15 with the third clarity index values ​​of 0.8R, 0.7R, 0.6R, 0.5R and 0.4R. Based on the same method, five second satellite images M21, M22, M23, M24 and M25 corresponding to the satellite image M2 and five second satellite images M31, M32, M33, M34 and M35 corresponding to the satellite image M3 are obtained. Then five training data sets are generated, namely ((M1, M11), (M2, M21), (M3, M31)), ((M1, M12), (M2, M22), (M3, M32)), ((M1, M13), (M2, M23), (M3, M33)), ((M1, M14) 4), (M2, M24), (M3, M34)) and ((M1, M15), (M2, M25), (M3, M35)), respectively use the above 5 training data sets to train the neural network model, obtain 5 pre-trained models, and then generate a clarity index range for each of the above pre-trained models based on the clarity index value of the training data set corresponding to each pre-trained model. For example, the clarity index range corresponding to the first pre-trained model is ≥0.75R, the clarity index range corresponding to the second pre-trained model is [0.65R, 0.75R), the clarity index range corresponding to the third pre-trained model is [0.55R, 0.65R), the clarity index range corresponding to the fourth pre-trained model is [0.45R, 0.55R), and the clarity index range corresponding to the fifth pre-trained model is <0.45R. A satellite image whose clarity index value is within any clarity index range is input into the pre-trained model corresponding to any clarity index range to obtain a satellite image whose clarity index value is R.

[0079] Since the clarity index values ​​of the satellite remote sensing images input into the pre-training model may be varied, if one pre-training model is used to train satellite remote sensing images with different clarity index values, the conversion accuracy will be low, and the clarity index value of the satellite remote sensing images finally generated will be low, failing to meet the requirements of high-definition images. According to the technical solution of the present invention, different pre-training models can be trained for different clarity index values, which can improve the accuracy and precision of the clarity index value conversion. At the same time, the pre-training model trained according to the technical solution of the present invention is highly practical and can be used not only in the present invention, but also when it is necessary to obtain a high-definition image with a clarity index value greater than that of the satellite remote sensing image, the satellite remote sensing image can be input into the pre-training model to obtain a high-definition image.

[0080] Specifically, in step 2, any first satellite remote sensing image is input into the pre-trained model to generate a second satellite remote sensing image, including:

[0081] Step 221: Obtain a second clarity index value of any first satellite remote sensing image.

[0082] Step 222: search the first storage module based on the second clarity index value to obtain a first pre-trained model corresponding to the second clarity index value.

[0083] Step 223: Input any first satellite remote sensing image into the first pre-trained model to generate a second satellite remote sensing image.

[0084] Exemplarily, if the second clarity index value of any first satellite remote sensing image is 0.7R, a second pre-trained model is obtained, and any first satellite remote sensing image is input into the second pre-trained model to obtain a second satellite remote sensing image with a clarity index value of R.

[0085] Step 3: After traversing all the first satellite remote sensing images, extract any second satellite remote sensing image, standardize the color of any second satellite remote sensing image based on a preset brightness reference table, and generate a third satellite remote sensing image.

[0086] Satellite remote sensing images are not taken at the same time. Due to the influence of weather and other factors, different satellite remote sensing images will have color differences. Therefore, standardization processing is performed to unify the color of each object in each satellite remote sensing image.

[0087] Specifically, in step 3, the color of any second satellite remote sensing image is standardized based on a preset brightness reference table to generate a third satellite remote sensing image, including:

[0088] Step 311: extract all ground objects in any second satellite remote sensing image, and classify ground objects of the same type into the same ground object set.

[0089] Step 312: extract any set of ground objects and generate a first brightness distribution histogram of any color component of any set of ground objects.

[0090] Step 313: Calculate the first average brightness value of any color component in any set of ground objects, and generate the first pixel statistics of any set of ground objects based on the first average brightness values ​​of all color components.

[0091] Step 314, traverse the pixel statistics table based on the first pixel statistics, extract the second pixel statistics value with the greatest similarity to the first pixel statistics from the pixel statistics table, and determine whether the maximum similarity value is greater than or equal to a first preset value. If so, proceed to step 315, if not, proceed to step 316.

[0092] Step 315, obtain the first identifier corresponding to the second pixel statistical value, search the brightness reference table based on the first identifier, extract the second brightness distribution histogram of any color component corresponding to the first identifier, and use the second brightness distribution histogram to calibrate the first brightness distribution histogram corresponding to any color component to standardize the color of the objects in any set of objects.

[0093] Step 316: Generate a unique identifier for the first pixel statistical value, update the pixel statistical table based on the unique identifier and the first pixel statistical value, and update the brightness reference table based on the unique identifier and the first brightness distribution histogram of any color component.

[0094] Step 317: After traversing all sets of ground objects in any second satellite remote sensing image, output a third satellite remote sensing image.

[0095] The first preset value is set according to the experience of those skilled in the art or according to an actual application scenario, and the embodiments of the present application are not limited to this.

[0096] All the objects in the second satellite remote sensing image are extracted based on the pixel values, and the object sets are divided. For example, the object types include roads, beaches, oceans, forests, grasslands, and the like.

[0097] A color image has three color components: red, green, and blue. Extracting any set of objects and drawing the first brightness distribution histogram of any color component of any set of objects means drawing the first red channel brightness distribution histogram, the first green channel brightness distribution histogram, and the first blue channel brightness distribution histogram of the set of objects.

[0098] Exemplarily, the first average brightness value of the red component is RE1, the first average brightness value of the green component is GR1, and the first average brightness value of the blue component is BL1, and the first pixel statistical value is (RE1, GR1, BL1).

[0099] Based on the first pixel statistical value, the pixel statistical table is traversed, and the second pixel statistical value with the greatest similarity to the first pixel statistical value is extracted from the pixel statistical table. If the maximum similarity value is less than the first preset value, it means that there is no feature of the same type as the feature corresponding to the first pixel statistical value in the pixel statistical table, then a unique identifier is generated for the first pixel statistical value, and the unique identifier and the first pixel statistical value are stored in the pixel statistical table correspondingly. At the same time, the unique identifier and the first brightness distribution histogram are stored in the brightness reference table correspondingly; if the maximum similarity value is greater than or equal to the first preset value, it means that the feature type corresponding to the first pixel statistical value is the same as the feature type corresponding to the second pixel statistical value, then the second red channel brightness distribution histogram corresponding to the first identifier is extracted to calibrate the first red channel brightness distribution histogram, the second green channel brightness distribution histogram corresponding to the first identifier is extracted to calibrate the first green channel brightness distribution histogram, and the second blue channel brightness distribution histogram corresponding to the first identifier is extracted to calibrate the first blue channel brightness distribution histogram. Based on this, the colors of the features in the feature set are standardized.

[0100] Specifically, after step 3, the following steps are also included:

[0101] Step 321: select any third satellite remote sensing image, extract all ground objects in any third satellite remote sensing image, and classify the ground objects of the same type into the same adjusted ground object set.

[0102] Step 322: extract any adjusted ground object set, and generate a third brightness distribution histogram of any color component of any adjusted ground object set.

[0103] Step 323: Calculate the second average brightness value of any color component in any adjusted ground object set, and generate a third pixel statistic value of any adjusted ground object set based on the second average brightness values ​​of all color components.

[0104] Step 324: traverse the pixel statistical table based on the third pixel statistical value, extract the fourth pixel statistical value with the greatest similarity to the third pixel statistical value from the pixel statistical table, and further obtain the second identifier corresponding to the fourth pixel statistical value.

[0105] Step 325: Generate a fifth pixel statistic value based on the third pixel statistic value and the fourth pixel statistic value, and store the fifth pixel statistic value in a pixel statistics table corresponding to the second identifier.

[0106] Step 326, search the brightness reference table based on the second identifier, extract the fourth brightness distribution histogram of any color component corresponding to the second identifier, generate the fifth brightness distribution histogram corresponding to any color component based on the third brightness distribution histogram and the fourth brightness distribution histogram corresponding to any color component, and store the fifth brightness distribution histogram corresponding to any color component in the brightness reference table corresponding to the second identifier.

[0107] Preferably, an average value of the third pixel statistic value and the fourth pixel statistic value is taken as the fifth pixel statistic value.

[0108] Preferably, the average value of the number of pixels of the third brightness distribution histogram and the fourth brightness distribution histogram of any color component at any brightness level is used as the new number of pixels of any of the above-mentioned brightness levels, and after traversing all brightness levels, the above-mentioned fifth brightness distribution histogram is generated based on all new pixel points; or, the sum of the number of pixels of the third brightness distribution histogram and the fourth brightness distribution histogram of any color component at any brightness level is used as the new number of pixels of any of the above-mentioned brightness levels, and after traversing all brightness levels, the above-mentioned fifth brightness distribution histogram is generated based on all new pixel points.

[0109] According to the technical solution of the present invention, after the colors of all the second satellite remote sensing images are standardized, the pixel statistics table and the brightness reference table are updated based on the generated third satellite remote sensing image, and new pixel statistics table and brightness reference table can be automatically generated based on the latest ground texture information and / or the latest ground objects.

[0110] Specifically, before step 3, all second satellite remote sensing images are subjected to defogging, radiation correction and geometric correction.

[0111] Step 4: After traversing all the second satellite remote sensing images, segment and combine all the third satellite remote sensing images to generate a fourth satellite remote sensing image of the ocean to be developed.

[0112] Due to the large area of ​​the undeveloped ocean area and / or the influence of clouds, multiple third satellite remote sensing images are segmented and combined to generate a fourth satellite remote sensing image that covers the entire undeveloped ocean area and has little cloud influence, laying the foundation for detailed marine national space planning.

[0113] Specifically, step 4 includes:

[0114] Step 41: All third satellite remote sensing images are superimposed and placed based on the latitude and longitude information to generate a superimposed image, and then a circumscribed rectangle of the superimposed image is drawn, and the side length of the circumscribed rectangle is obtained.

[0115] Step 42: Generate segmentation parameters based on the first preset ratio and the side length, and divide the circumscribed rectangle into N2 rectangular areas based on the segmentation parameters.

[0116] Step 43: extract any rectangular area, combine and stitch the multiple layers of satellite remote sensing images in any rectangular area, and generate N3 regional stitched images, wherein the coverage area of ​​any regional stitched image overlaps with the coverage area of ​​the superimposed image in any rectangular area.

[0117] Step 44, extract any regional stitching image, determine whether the area covered by any regional stitching image is land or ocean, or contains both land and ocean. If it is land, define the ratio of the cloud area to the land area as the first cloud ratio of any regional stitching image. If it is ocean, define the ratio of the cloud area to the ocean area as the second cloud ratio of any regional stitching image. If it contains both land and ocean, extract the first land area and the first ocean area in any regional stitching image based on the pixel value, calculate the first ratio of the first cloud area to the first land area, and the second ratio of the second cloud area to the first ocean area. Calculate the third cloud ratio of any regional stitching image based on the first ratio and the second ratio. The calculation formula of the third cloud ratio RC3 is:

[0118] ,

[0119] in, and is the parameter weight coefficient, R1 is the first ratio, R2 is the second ratio, , the first cloud region is the cloud region above the first land region, and the second cloud region is the cloud region above the first ocean region.

[0120] Step 45: After traversing all the regional stitching images, the regional stitching image corresponding to the minimum cloud ratio is used as the regional image to be stitched corresponding to any rectangular area, wherein the cloud ratio is the first cloud ratio, the second cloud ratio or the third cloud ratio.

[0121] Step 46: After traversing all rectangular areas, extract all images of the area to be stitched, and stitch all images of the area to be stitched to generate a pending satellite remote sensing image.

[0122] Step 47, extract the second land area and the second ocean area of ​​the pending satellite remote sensing image based on the pixel value, define the ratio of the third cloud area to the second land area as the fourth cloud ratio of the pending satellite remote sensing image, and determine whether the fourth cloud ratio is less than or equal to the second preset value. If not, proceed to step 48; if so, proceed to step 49, wherein the third cloud area is the cloud area above the second land area.

[0123] Step 48 , reset the first preset ratio, and then return to step 42 .

[0124] Step 49: output the pending satellite remote sensing image as the fourth satellite remote sensing image.

[0125] The second preset value and the first preset ratio are set according to the experience of those skilled in the art or according to actual application scenarios, and the embodiments of the present application are not limited thereto.

[0126] Preferably, all third satellite remote sensing images are rectangular images of equal size, and all third satellite remote sensing images are superimposed and placed based on latitude and longitude information to generate a superimposed image. Exemplarily, the circumscribed rectangle of the superimposed image is a 20×30 rectangle, and the first preset ratio is 1 / 5 (the first preset ratio of the long side and the wide side may be the same or different), and the circumscribed rectangle is divided into 20 rectangular areas.

[0127] by Figure 2 Take as an example to explain the generation process of the regional mosaic image in the technical solution of the present invention. For ease of understanding, it is displayed in a three-dimensional pattern. In fact, each third satellite remote sensing image is superimposed. Figure 2 As shown, there are three third satellite remote sensing images P1, P2 and P3, B1 is the selected rectangular area, B2 is the area where the rectangular area B1 is projected on the satellite remote sensing image P2, A1 is the area where the rectangular area B1 is projected on the satellite remote sensing image P1, A2 is the area where the rectangular area B1 is projected on the satellite remote sensing image P3, A3 is the partial area where the rectangular area B1 is projected on the satellite remote sensing image P2 (the total area of ​​A1, A2 and A3 is equal to the area of ​​the rectangular area B1), and in addition, the area where A1 is projected on the satellite remote sensing image P2 is C1 (not shown), and the area where A2 is projected on the satellite remote sensing image P2 is C2 (not shown). The three layers of satellite remote sensing images in the rectangular area B1 are combined and spliced ​​to generate four regional spliced ​​images, namely (A1+A2+A3), (A1+A3+C2), (C1+A2+A3), (A1+A2+A3), and B2.

[0128] Since clouds have a greater impact on land areas and a smaller impact on ocean areas, when the area covered by the stitched image of any region includes both land and ocean, a larger parameter weight coefficient is set for the first ratio of the land area, and a smaller parameter weight coefficient is set for the second ratio of the ocean area. Preferably, only the cloud ratio of the land area may be considered, and the cloud ratio of the ocean area may not be considered, that is, the cloud ratio of the land area is used as the cloud ratio of the stitched image of any region.

[0129] At the same time, since clouds have a greater impact on land areas and a smaller impact on ocean areas, only the cloud amount above the land area is considered when calculating the fourth cloud ratio. In addition to this method, the cloud amount above the land area and the ocean area can also be considered at the same time, and the fourth cloud ratio of the pending satellite remote sensing image can be calculated according to the calculation formula of the third cloud ratio.

[0130] The image of the area to be stitched with the smallest cloud ratio is extracted to generate a pending satellite remote sensing image, where the pending satellite remote sensing image is a satellite remote sensing image with the smallest cloud ratio that can be combined based on the regional stitching images segmented at the first preset ratio. According to the technical solution of the present invention, a satellite remote sensing image with the smallest cloud ratio can be quickly generated.

[0131] Specifically, in step 48, the first preset ratio is reset to reduce the first preset ratio according to the second preset ratio.

[0132] The second preset ratio is set according to the experience of those skilled in the art or according to the actual application scenario, and the embodiment of the present application does not limit this. For example, the second preset ratio is 1 / 9.

[0133] The smaller the rectangular area is, the greater the probability of generating a satellite remote sensing image containing a small amount of clouds. Therefore, when the proportion of clouds in the land area in the pending satellite remote sensing image is greater than the second preset value, the size of the rectangular area is reduced.

[0134] The smaller the generated rectangular area is, the greater the amount of calculation is, and the cloud ratio in the land area will tend to be stable as the rectangular area becomes smaller. Therefore, after the number of updates is met or the area of ​​the rectangular area reaches the preset threshold, the first preset ratio is stopped from being updated, and the pending satellite remote sensing image corresponding to the smallest cloud ratio in the land area is output as the fourth satellite remote sensing image.

[0135] Step 5: Divide the fourth satellite remote sensing image into ocean units and land units, and send the result to the user terminal.

[0136] Figure 3 FIG. 2 is a schematic diagram of a structure of an embodiment of an image processing system for marine land space detailed planning provided by the present invention. Figure 3 As shown, the system includes: an image acquisition module 101, a clarity adjustment module 102, a standardization processing module 103, an image synthesis module 104 and an image division module 105.

[0137] The image acquisition module 101 is used to acquire N1 first satellite remote sensing images of the ocean area to be developed, and acquire imaging information of each first satellite remote sensing image.

[0138] The clarity adjustment module 102 is used to extract any first satellite remote sensing image, input any first satellite remote sensing image into a pre-trained model, and generate a second satellite remote sensing image, wherein the clarity index value of the second satellite remote sensing image is greater than the second clarity index value of any first satellite remote sensing image.

[0139] The standardization processing module 103 is used to extract any second satellite remote sensing image after traversing all first satellite remote sensing images, and perform color standardization processing on any second satellite remote sensing image based on a preset brightness reference table to generate a third satellite remote sensing image.

[0140] The image synthesis module 104 is used to segment and combine all the third satellite remote sensing images after traversing all the second satellite remote sensing images to generate a fourth satellite remote sensing image of the ocean to be developed.

[0141] The image division module 105 is used to divide the fourth satellite remote sensing image into ocean units and land units, and send the units to the user terminal.

[0142] According to another aspect of an embodiment of the present invention, a computer storage medium is provided, which stores program instructions, wherein when the program instructions are executed, the device where the computer storage medium is located is controlled to execute any one of the above-mentioned image processing methods applied to detailed planning of marine land space.

[0143] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0144] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the above-mentioned program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0145] The above-mentioned embodiments only express the preferred implementation mode of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. An image processing method for detailed marine land space planning, It is characterized in that The steps include: Step 1, obtaining N1 first satellite remote sensing images of the ocean area to be developed, and obtaining imaging information of each first satellite remote sensing image; Step 2: extract any first satellite remote sensing image, input any of the first satellite remote sensing images into a pre-trained model, and generate a second satellite remote sensing image, wherein the clarity index value of the second satellite remote sensing image is greater than the second clarity index value of any of the first satellite remote sensing images; Step 3: after traversing all the first satellite remote sensing images, extract any second satellite remote sensing image, and perform color standardization on any second satellite remote sensing image based on a preset brightness reference table to generate a third satellite remote sensing image; Step 4: After traversing all the second satellite remote sensing images, segment and combine all the third satellite remote sensing images to generate a fourth satellite remote sensing image of the ocean to be developed; Step 5: dividing the fourth satellite remote sensing image into ocean units and land units, and sending the images to a user terminal; The step 4 comprises: Step 41: superimpose all the third satellite remote sensing images based on the latitude and longitude information to generate a superimposed image, then draw a circumscribed rectangle of the superimposed image, and obtain the side length of the circumscribed rectangle; Step 42: generating segmentation parameters based on the first preset ratio and the side length, and dividing the circumscribed rectangle into N2 rectangular areas based on the segmentation parameters; Step 43, extract any rectangular area, combine and stitch the multi-layer satellite remote sensing images in any of the rectangular areas, and generate N3 regional stitched images, wherein the coverage area of ​​any regional stitched image overlaps with the coverage area of ​​the superimposed image in any of the rectangular areas; Step 44: extract any regional stitching image, determine whether the area covered by any of the regional stitching images is land or ocean, or includes both land and ocean. If it is land, define the ratio of the cloud area to the land area as the first cloud ratio of any of the regional stitching images. If it is ocean, define the ratio of the cloud area to the ocean area as the second cloud ratio of any of the regional stitching images. If it includes both land and ocean, extract the first land area and the first ocean area in any of the regional stitching images based on the pixel values, calculate a first ratio of the first cloud area to the first land area, and a second ratio of the second cloud area to the first ocean area. Calculate a third cloud ratio of any of the regional stitching images based on the first ratio and the second ratio. The third cloud ratio RC 3 The calculation formula is: ; in, and is the parameter weight coefficient, R1 is the first ratio, R2 is the second ratio, , the first cloud region is the cloud region above the first land region, and the second cloud region is the cloud region above the first ocean region; Step 45: after traversing all the regional stitching images, the regional stitching image corresponding to the minimum cloud proportion is used as the regional image to be stitched corresponding to any of the rectangular areas, wherein the cloud proportion is the first cloud proportion, the second cloud proportion or the third cloud proportion; Step 46: after traversing all rectangular areas, extract all images of the area to be stitched, and stitch all the images of the area to be stitched to generate a pending satellite remote sensing image; Step 47, extracting the second land area and the second ocean area of ​​the pending satellite remote sensing image based on the pixel value, defining the ratio of the third cloud area to the second land area as the fourth cloud ratio of the pending satellite remote sensing image, and judging whether the fourth cloud ratio is less than or equal to a second preset value, if not, proceeding to step 48, and if so, proceeding to step 49, wherein the third cloud area is the cloud area above the second land area; Step 48, resetting the first preset ratio, and then returning to step 42; Step 49: output the pending satellite remote sensing image as the fourth satellite remote sensing image.

2. The method according to claim 1, It is characterized in that In step 2, the training method of the pre-trained model is: Step 211: Acquire M first satellite images, wherein any first satellite image has the same first clarity index value, and the first clarity index value is greater than the second clarity index value of any first satellite remote sensing image; Step 212: extract any first satellite image, and perform degradation processing on any first satellite image according to the preset K third clarity index values ​​to generate K second satellite images, wherein one second satellite image corresponds to one third clarity index value; Step 213: after traversing all the first satellite images, K training data sets are generated based on the clarity index values, wherein any training data set includes M training data pairs, and any training data pair is any of the first satellite images and a second satellite image whose clarity index value corresponding to any of the first satellite images is a fourth clarity index value, and the fourth clarity index value is a clarity index value corresponding to any of the training data sets; Step 214: extract any of the training data sets, use any of the training data sets to train a neural network model, and generate a pre-trained model corresponding to any of the training data sets; Step 215: Generate a clarity index range based on the clarity index value corresponding to any of the training data sets, and store the clarity index range in a first storage module in correspondence with the pre-trained model corresponding to any of the training data sets.

3. The method according to claim 2, It is characterized in that In step 2, inputting any of the first satellite remote sensing images into a pre-trained model to generate a second satellite remote sensing image includes: Step 221, obtaining a second clarity index value of any of the first satellite remote sensing images; Step 222: searching the first storage module based on the second clarity index value to obtain a first pre-trained model corresponding to the second clarity index value; Step 223: input any of the first satellite remote sensing images into the first pre-trained model to generate the second satellite remote sensing image.

4. The method according to claim 1, It is characterized in that In step 3, the step of performing color standardization processing on any of the second satellite remote sensing images based on a preset brightness reference table to generate a third satellite remote sensing image includes: Step 311, extracting all the ground objects in any of the second satellite remote sensing images, and classifying the ground objects of the same type into the same ground object set; Step 312: extract any set of ground objects and generate a first brightness distribution histogram of any color component of any of the ground object sets; Step 313: Calculate the first average brightness value of any color component in any of the ground object sets, and generate a first pixel statistic value of any of the ground object sets based on the first average brightness values ​​of all color components; Step 314: traverse the pixel statistics table based on the first pixel statistics value, extract the second pixel statistics value with the greatest similarity to the first pixel statistics value from the pixel statistics table, and determine whether the maximum similarity value is greater than or equal to a first preset value. If so, proceed to step 315; if not, proceed to step 316. Step 315: obtaining a first identifier corresponding to the second pixel statistical value, searching the brightness reference table based on the first identifier, extracting a second brightness distribution histogram of any color component corresponding to the first identifier, and using the second brightness distribution histogram to calibrate the first brightness distribution histogram corresponding to any color component, so as to standardize the color of any ground object in the ground object set; Step 316: Generate a unique identifier for the first pixel statistical value, update the pixel statistical table based on the unique identifier and the first pixel statistical value, and update the brightness reference table based on the unique identifier and the first brightness distribution histogram of any of the color components; Step 317: After traversing all sets of ground objects in any of the second satellite remote sensing images, output the third satellite remote sensing image.

5. The method according to claim 4, It is characterized in that After step 3, the following steps are also included: Step 321: select any third satellite remote sensing image, extract all ground objects in any of the third satellite remote sensing images, and classify the ground objects of the same type into the same adjusted ground object set; Step 322: extract any one of the adjustment ground object sets, and generate a third brightness distribution histogram of any one of the color components of any one of the adjustment ground object sets; Step 323: Calculate the second average brightness value of any color component in any of the adjusted ground object sets, and generate a third pixel statistic value of any of the adjusted ground object sets based on the second average brightness values ​​of all color components; Step 324: traverse the pixel statistical table based on the third pixel statistical value, extract a fourth pixel statistical value having the greatest similarity to the third pixel statistical value from the pixel statistical table, and further obtain a second identifier corresponding to the fourth pixel statistical value; Step 325: Generate a fifth pixel statistical value based on the third pixel statistical value and the fourth pixel statistical value, and store the fifth pixel statistical value in the pixel statistical table in correspondence with the second identifier; Step 326, search the brightness reference table based on the second identifier, extract the fourth brightness distribution histogram of any color component corresponding to the second identifier, generate a fifth brightness distribution histogram corresponding to any color component based on the third brightness distribution histogram and the fourth brightness distribution histogram corresponding to any color component, and store the fifth brightness distribution histogram corresponding to any color component in the brightness reference table corresponding to the second identifier.

6. The method according to claim 1, It is characterized in that In step 48, resetting the first preset ratio is to reduce the first preset ratio according to a second preset ratio.

7. The method according to claim 1, It is characterized in that Before step 3, all the second satellite remote sensing images are subjected to defogging, radiation correction and geometric correction.

8. An image processing system for marine land space detailed planning, used to implement the method according to any one of claims 1 to 7, It is characterized in that include: Image acquisition module, clarity adjustment module, standardization processing module, image synthesis module and image division module; The image acquisition module is used to acquire N1 first satellite remote sensing images of the ocean area to be developed, and to acquire imaging information of each first satellite remote sensing image; The clarity adjustment module is used to extract any first satellite remote sensing image, input any of the first satellite remote sensing images into a pre-training model, and generate a second satellite remote sensing image, wherein the clarity index value of the second satellite remote sensing image is greater than the second clarity index value of any of the first satellite remote sensing images; The standardization processing module is used to extract any second satellite remote sensing image after traversing all the first satellite remote sensing images, and perform color standardization processing on any second satellite remote sensing image based on a preset brightness reference table to generate a third satellite remote sensing image; The image synthesis module is used to segment and combine all the third satellite remote sensing images after traversing all the second satellite remote sensing images to generate a fourth satellite remote sensing image of the ocean to be developed; The image division module is used to divide the fourth satellite remote sensing image into ocean units and land units, and send them to the user terminal; The segmenting and combining of all the third satellite remote sensing images to generate the fourth satellite remote sensing image of the ocean to be developed comprises: Step 41: superimpose all the third satellite remote sensing images based on the latitude and longitude information to generate a superimposed image, then draw a circumscribed rectangle of the superimposed image, and obtain the side length of the circumscribed rectangle; Step 42: generating segmentation parameters based on the first preset ratio and the side length, and dividing the circumscribed rectangle into N2 rectangular areas based on the segmentation parameters; Step 43, extract any rectangular area, combine and stitch the multi-layer satellite remote sensing images in any of the rectangular areas, and generate N3 regional stitched images, wherein the coverage area of ​​any regional stitched image overlaps with the coverage area of ​​the superimposed image in any of the rectangular areas; Step 44: extract any regional stitching image, determine whether the area covered by any of the regional stitching images is land or ocean, or includes both land and ocean. If it is land, define the ratio of the cloud area to the land area as the first cloud ratio of any of the regional stitching images. If it is ocean, define the ratio of the cloud area to the ocean area as the second cloud ratio of any of the regional stitching images. If it includes both land and ocean, extract the first land area and the first ocean area in any of the regional stitching images based on the pixel values, calculate a first ratio of the first cloud area to the first land area, and a second ratio of the second cloud area to the first ocean area. Calculate a third cloud ratio of any of the regional stitching images based on the first ratio and the second ratio. The third cloud ratio RC 3 The calculation formula is: ; in, and is the parameter weight coefficient, R1 is the first ratio, R2 is the second ratio, , the first cloud region is the cloud region above the first land region, and the second cloud region is the cloud region above the first ocean region; Step 45: after traversing all the regional stitching images, the regional stitching image corresponding to the minimum cloud proportion is used as the regional image to be stitched corresponding to any of the rectangular areas, wherein the cloud proportion is the first cloud proportion, the second cloud proportion or the third cloud proportion; Step 46: after traversing all rectangular areas, extract all images of the area to be stitched, and stitch all the images of the area to be stitched to generate a pending satellite remote sensing image; Step 47, extracting the second land area and the second ocean area of ​​the pending satellite remote sensing image based on the pixel value, defining the ratio of the third cloud area to the second land area as the fourth cloud ratio of the pending satellite remote sensing image, and judging whether the fourth cloud ratio is less than or equal to a second preset value, if not, proceeding to step 48, and if so, proceeding to step 49, wherein the third cloud area is the cloud area above the second land area; Step 48, resetting the first preset ratio, and then returning to step 42; Step 49: output the pending satellite remote sensing image as the fourth satellite remote sensing image.

9. A computer storage medium, It is characterized in that The computer storage medium stores program instructions, wherein when the program instructions are executed, the device where the computer storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

Citation Information

Patent Citations

  • Image correction method based on unmanned aerial vehicle aerial photography and satellite remote sensing fusion

    CN112393714A

  • Remote sensing image segmentation method and system for marine territorial space detailed planning

    CN117934518A

  • Remote sensing image processing method and system based on adversarial neural network model

    CN111160128A

  • Roof information automatic identification method and device based on high-definition satellite image

    CN115937708A

  • Image processing method and system, storage medium and electronic equipment

    CN116703777A