A camellia forest growth trend prediction method and system based on satellite remote sensing image recognition

Abstract

This application provides a method and system for predicting the growth trend of camellia oleifera forests based on satellite remote sensing image recognition. The method includes: acquiring remote sensing images; setting segmentation rules; performing semantic segmentation on the remote sensing images according to the segmentation rules to obtain several sub-regions; extracting features from several sub-regions; calculating the similarity between the sub-region features and preset feature values ​​to obtain feature similarity; determining whether the feature similarity is greater than or equal to a preset similarity threshold; if it is greater than or equal to, generating correction information and adjusting the segmentation rules according to the correction information; if it is less than, acquiring remote sensing data of several sub-regions; inputting the remote sensing data of several sub-regions into a preset growth prediction model to obtain growth prediction information; transmitting the growth prediction information to a terminal in a predetermined manner; and improving prediction accuracy by performing semantic segmentation on the remote sensing images and using the growth prediction model to predict the growth trend of camellia oleifera forests in real time.

Description

Technical Field

[0001] This application relates to the field of growth trend prediction, and more specifically, to a method and system for predicting the growth trend of Camellia oleifera forests based on satellite remote sensing image recognition. Background Technology

[0002] Camellia oleifera is a small tree or shrub belonging to the genus Camellia in the family Theaceae. Young branches are covered with coarse hairs; leaves are leathery, elliptical or obovate, with a blunt tip and a cuneate base; the midrib on the underside is covered with long hairs and has fine teeth; the petiole is covered with coarse hairs; flowers are terminal, leathery, broadly ovate, with white, obovate petals; stamens have nearly free filaments; the capsule is spherical. Remote sensing refers to non-contact, long-distance detection technology. It generally refers to the detection of the electromagnetic radiation and reflection characteristics of objects using sensors / remote sensors. Remote sensing uses instruments sensitive to electromagnetic waves, such as remote sensors, to detect target features at a distance and without contact with the target object. By analyzing remote sensing image information, the distribution information and growth trend of Camellia oleifera forests can be identified, thereby predicting the growth of Camellia oleifera forests. However, existing methods for predicting the growth of Camellia oleifera forests have poor accuracy and cannot accurately predict the growth of Camellia oleifera forests, resulting in significant prediction errors. Effective technical solutions are urgently needed to address these problems. Summary of the Invention

[0003] The purpose of this application is to provide a method and system for predicting the growth trend of Camellia oleifera forests based on satellite remote sensing image recognition. This method can improve the prediction accuracy by performing semantic segmentation on remote sensing images and using a growth prediction model to predict the growth trend of Camellia oleifera forests in real time.

[0004] This application also provides a method for predicting the growth trend of camellia oleifera forests based on satellite remote sensing image recognition, including:

[0005] Acquire remote sensing images, set segmentation rules, and perform semantic segmentation on the remote sensing images according to the segmentation rules to obtain several sub-regions;

[0006] Extract features from several sub-regions, calculate the similarity between the sub-region features and preset feature values, and obtain the feature similarity.

[0007] Determine whether the feature similarity is greater than or equal to a preset similarity threshold;

[0008] If it is greater than or equal to, then correction information is generated, and the segmentation rules are adjusted according to the correction information;

[0009] If it is less than, then obtain remote sensing data of several sub-regions, input the remote sensing data of several sub-regions into the preset growth prediction model, and obtain growth prediction information;

[0010] The growth prediction information is transmitted to the terminal in a predetermined manner.

[0011] Optionally, in the method for predicting the growth trend of camellia oleifera forests based on satellite remote sensing image recognition described in the embodiments of this application, the steps of acquiring remote sensing images, setting segmentation rules, and performing semantic segmentation of the remote sensing images according to the segmentation rules to obtain several sub-regions include:

[0012] Acquire remote sensing images, randomly divide the remote sensing images into several image domains, and calculate the pixel grayscale values ​​of the image domains;

[0013] The grayscale change rate is obtained by comparing the grayscale values ​​of two neighboring image domain pixels with the first grayscale mean.

[0014] Determine whether the grayscale change rate is greater than a preset grayscale change rate threshold;

[0015] If the area is greater than the specified value, the image domain will be segmented and the area adjusted accordingly.

[0016] If it is less than, the image domains are superimposed to calculate the superimposed area of ​​the image domains.

[0017] Optionally, in the method for predicting the growth trend of camellia oleifera forests based on satellite remote sensing image recognition described in the embodiments of this application, the step of adjusting the area of ​​the image domain by segmenting it if the value is greater than the specified value includes:

[0018] Obtain the size of the segmented area, calculate the difference between the segmented area and the preset area value, and obtain the segmented area difference.

[0019] Determine whether the difference in the segmented areas is greater than a preset area threshold;

[0020] If it is greater than, a second grayscale mean is generated, and the remote sensing image is segmented a second time based on the second grayscale mean.

[0021] If it is less than, then a grayscale mean adjustment amount is generated, and the first grayscale mean is adjusted by the same amount based on the grayscale mean adjustment amount.

[0022] Optionally, in the method for predicting the growth trend of camellia oleifera forests based on satellite remote sensing image recognition described in the embodiments of this application, the steps of acquiring remote sensing images, setting segmentation rules, and performing semantic segmentation of the remote sensing images according to the segmentation rules to obtain several sub-regions include:

[0023] The remote sensing image is divided into several small-area remote sensing images using a set segmentation algorithm.

[0024] Several small-area remote sensing images are merged sequentially according to a predetermined number to obtain merged remote sensing images;

[0025] Obtain parameter information for merged remote sensing images.

[0026] The merged remote sensing image parameter information is compared with the preset remote sensing image parameter information to obtain the merging deviation rate;

[0027] Determine whether the merged deviation rate is greater than or equal to a preset deviation rate threshold;

[0028] If the value is greater than or equal to the value, the merged remote sensing images will be split sequentially according to the reverse order of merging.

[0029] If the value is smaller, then continue merging the remote sensing images of the small area.

[0030] Optionally, in the method for predicting the growth trend of camellia oleifera forests based on satellite remote sensing image recognition described in the embodiments of this application, the step of merging several small-area remote sensing images sequentially according to a predetermined number to obtain merged remote sensing images includes:

[0031] Obtain the number of merged remote sensing images of a small area;

[0032] The difference between the merged quantity value and the preset threshold is calculated.

[0033] If the number is less than the preset threshold, the corresponding number of small area remote sensing images will be merged and fused.

[0034] If the difference is greater than the value, the remote sensing image of the small area will be split into smaller parts with the same number of differences.

[0035] Optionally, in the method for predicting the growth trend of camellia oleifera forests based on satellite remote sensing image recognition described in the embodiments of this application, the step of splitting the remote sensing images of a small area with a difference equal to the value if the difference is greater than the value includes:

[0036] Obtain the attribute information of a small area remote sensing image, calculate the similarity between the attribute information of the small area remote sensing image and the attribute information of the merged remote sensing image, and obtain the similarity score.

[0037] Determine whether the similarity is greater than a preset similarity threshold;

[0038] If the difference is greater than the value, the number of splits is generated based on the difference, and the small remote sensing images of the same number of regions that are finally merged are split sequentially based on the number of splits.

[0039] If the value is smaller than the target value, the corresponding small area of ​​remote sensing imagery will be split.

[0040] Secondly, embodiments of this application provide a system for predicting the growth trend of Camellia oleifera forests based on satellite remote sensing image recognition. This system includes a memory and a processor. The memory includes a program for predicting the growth trend of Camellia oleifera forests based on satellite remote sensing image recognition. When the program for predicting the growth trend of Camellia oleifera forests based on satellite remote sensing image recognition is executed by the processor, it performs the following steps:

[0041] Acquire remote sensing images, set segmentation rules, and perform semantic segmentation on the remote sensing images according to the segmentation rules to obtain several sub-regions;

[0042] Extract features from several sub-regions, calculate the similarity between the sub-region features and preset feature values, and obtain the feature similarity.

[0043] Determine whether the feature similarity is greater than or equal to a preset similarity threshold;

[0044] If it is greater than or equal to, then correction information is generated, and the segmentation rules are adjusted according to the correction information;

[0045] If it is less than, then obtain remote sensing data of several sub-regions, input the remote sensing data of several sub-regions into the preset growth prediction model, and obtain growth prediction information;

[0046] The growth prediction information is transmitted to the terminal in a predetermined manner.

[0047] Optionally, in the Camellia oleifera forest growth trend prediction system based on satellite remote sensing image recognition described in this application embodiment, the steps of acquiring remote sensing images, setting segmentation rules, and performing semantic segmentation of the remote sensing images according to the segmentation rules to obtain several sub-regions include:

[0048] Acquire remote sensing images, randomly divide the remote sensing images into several image domains, and calculate the pixel grayscale values ​​of the image domains;

[0049] The grayscale change rate is obtained by comparing the grayscale values ​​of two neighboring image domain pixels with the first grayscale mean.

[0050] Determine whether the grayscale change rate is greater than a preset grayscale change rate threshold;

[0051] If the area is greater than the specified value, the image domain will be segmented and the area adjusted accordingly.

[0052] If it is less than, the image domains are superimposed to calculate the superimposed area of ​​the image domains.

[0053] Optionally, in the Camellia oleifera forest growth trend prediction system based on satellite remote sensing image recognition described in this application embodiment, the step of adjusting the image domain segmentation area if the value is greater than the specified value includes:

[0054] Obtain the size of the segmented area, calculate the difference between the segmented area and the preset area value, and obtain the segmented area difference.

[0055] Determine whether the difference in the segmented areas is greater than a preset area threshold;

[0056] If it is greater than, a second grayscale mean is generated, and the remote sensing image is segmented a second time based on the second grayscale mean.

[0057] If it is less than, then a grayscale mean adjustment amount is generated, and the first grayscale mean is adjusted by the same amount based on the grayscale mean adjustment amount.

[0058] Optionally, in the Camellia oleifera forest growth trend prediction system based on satellite remote sensing image recognition described in this application embodiment, the steps of acquiring remote sensing images, setting segmentation rules, and performing semantic segmentation of the remote sensing images according to the segmentation rules to obtain several sub-regions include:

[0059] The remote sensing image is divided into several small-area remote sensing images using a set segmentation algorithm.

[0060] Several small-area remote sensing images are merged sequentially according to a predetermined number to obtain merged remote sensing images;

[0061] Obtain parameter information for merged remote sensing images.

[0062] The merged remote sensing image parameter information is compared with the preset remote sensing image parameter information to obtain the merging deviation rate;

[0063] Determine whether the merged deviation rate is greater than or equal to a preset deviation rate threshold;

[0064] If the value is greater than or equal to the value, the merged remote sensing images will be split sequentially according to the reverse order of merging.

[0065] If the value is smaller, then continue merging the remote sensing images of the small area.

[0066] As can be seen from the above, the method and system for predicting the growth trend of camellia oleifera forests based on satellite remote sensing image recognition provided in this application embodiment involves acquiring remote sensing images, setting segmentation rules, performing semantic segmentation of the remote sensing images according to the segmentation rules to obtain several sub-regions; extracting features of several sub-regions, calculating the similarity between the sub-region features and preset feature values ​​to obtain feature similarity; determining whether the feature similarity is greater than or equal to a preset similarity threshold; if it is greater than or equal to, generating correction information and adjusting the segmentation rules according to the correction information; if it is less than, acquiring remote sensing data of several sub-regions, inputting the remote sensing data of several sub-regions into a preset growth prediction model to obtain growth prediction information; transmitting the growth prediction information to the terminal in a predetermined manner; and improving prediction accuracy by performing semantic segmentation of remote sensing images and using the growth prediction model to predict the growth trend of camellia oleifera forests in real time.

[0067] Other features and advantages of this application will be set forth in the following description, and the advantages of this application may be inferred from the description or learned by practicing the embodiments of this application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0068] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0069] Figure 1 A flowchart of a method for predicting the growth trend of Camellia oleifera forests based on satellite remote sensing image recognition, provided in an embodiment of this application;

[0070] Figure 2 A flowchart illustrating the segmentation area adjustment process for a method for predicting the growth trend of Camellia oleifera forests based on satellite remote sensing image recognition, as provided in this application embodiment.

[0071] Figure 3 A flowchart of grayscale mean adjustment for a method for predicting the growth trend of Camellia oleifera forests based on satellite remote sensing image recognition provided in an embodiment of this application;

[0072] Figure 4 A flowchart of small-area remote sensing image merging for a method of predicting the growth trend of camellia oleifera forests based on satellite remote sensing image recognition provided in this application embodiment;

[0073] Figure 5 A flowchart illustrating the adjustment of planting status of Camellia oleifera forests using a satellite remote sensing image recognition-based growth trend prediction method provided in this application embodiment;

[0074] Figure 6 This is a schematic diagram of the structure of the Camellia oleifera forest growth trend prediction system based on satellite remote sensing image recognition provided in the embodiments of this application. Detailed Implementation

[0075] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0076] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0077] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a method for predicting the growth trend of Camellia oleifera forests based on satellite remote sensing image recognition, as described in some embodiments of this application. This method, used in a terminal device, includes the following steps:

[0078] S101, acquire remote sensing images, set segmentation rules, perform semantic segmentation on remote sensing images according to segmentation rules, and obtain several sub-regions;

[0079] S102, extract features from several sub-regions, calculate the similarity between the sub-region features and preset feature values, and obtain the feature similarity.

[0080] S103, determine whether the feature similarity is greater than or equal to the preset similarity threshold;

[0081] S104, if it is greater than or equal to, then generate correction information and adjust the segmentation rules according to the correction information;

[0082] S105, if it is less than, then acquire several sub-region remote sensing data, input the several sub-region remote sensing data into the preset growth prediction model, and obtain growth prediction information;

[0083] S106, the growth prediction information is transmitted to the terminal in a predetermined manner.

[0084] It should be noted that remote sensing is a new technology that combines inductive telemetry and resource management monitoring of the Earth's surface (such as resource management of trees, grasslands, soil, water, minerals, crops, fish, and wildlife) using telemetry instruments on platforms such as artificial Earth satellites and aircraft. Remote sensing refers to all non-contact, long-distance detection technologies. It utilizes modern vehicles and sensors to acquire the electromagnetic wave characteristics of target objects from a distance, and through the transmission, storage, satellite imagery, correction, and identification of this information, it ultimately achieves its function.

[0085] Please refer to Figure 2 , Figure 2This is a flowchart illustrating the segmentation area adjustment process of a method for predicting the growth trend of Camellia oleifera forests based on satellite remote sensing image recognition, as described in some embodiments of this application. According to embodiments of the present invention, remote sensing images are acquired, segmentation rules are set, and semantic segmentation is performed on the remote sensing images according to the segmentation rules to obtain several sub-regions, including:

[0086] S201, acquire remote sensing images, randomly divide the remote sensing images into several image domains, and calculate the grayscale values ​​of the pixels in the image domains;

[0087] S202, compare the gray values ​​of two neighboring image domain pixels with the first gray average value to obtain the gray change rate;

[0088] S203, determine whether the grayscale change rate is greater than the preset grayscale change rate threshold;

[0089] S204, if it is greater than, then the image domain will be segmented and the area adjusted;

[0090] If S205 is less than 5, then the image domain is superimposed to calculate the superimposed area of ​​the image domain.

[0091] It should be noted that region-based segmentation methods use defined rules to aggregate images into different sub-regions. Region-based segmentation methods can better filter noise, are suitable for images with significant intensity changes, and are less time-consuming.

[0092] Please refer to Figure 3 , Figure 3 This is a flowchart illustrating the grayscale mean adjustment process for a method of predicting the growth trend of Camellia oleifera forests based on satellite remote sensing image recognition, as described in some embodiments of this application. According to embodiments of the present invention, if the grayscale mean is greater than a certain value, the image domain is segmented and the area is adjusted, including:

[0093] S301, Obtain the size of the segmented area, calculate the difference between the segmented area and the preset area value, and obtain the segmented area difference;

[0094] S302, Determine whether the difference in the segmented area is greater than the preset area threshold;

[0095] If S303 is greater than 1, then a second grayscale mean is generated, and the remote sensing image is segmented a second time based on the second grayscale mean.

[0096] S304, if it is less than, then generate a grayscale mean adjustment amount, and adjust the first grayscale mean by the same amount according to the grayscale mean adjustment amount.

[0097] It should be noted that the threshold segmentation algorithm is based on the segmented area of ​​the image and inputs it into the prediction model to identify sub-regions. This method first determines one or more thresholds and compares the area of ​​the sub-region with the area threshold to obtain the specific category of the sub-region. It can handle situations that are prone to pixel value changes due to noise, illumination and other factors.

[0098] Please refer to Figure 4 , Figure 4 This is a flowchart of a small-area remote sensing image merging process for a method of predicting the growth trend of Camellia oleifera forests based on satellite remote sensing image recognition, as described in some embodiments of this application. According to embodiments of the present invention, remote sensing images are acquired, segmentation rules are set, and semantic segmentation is performed on the remote sensing images according to the segmentation rules to obtain several sub-regions, including:

[0099] S401, using a set segmentation algorithm, divides the remote sensing image into several small-area remote sensing images;

[0100] S402, merge several small area remote sensing images sequentially according to a predetermined number to obtain merged remote sensing images;

[0101] S403, Obtain parameter information for merged remote sensing images;

[0102] S404, compare the merged remote sensing image parameter information with the preset remote sensing image parameter information to obtain the merging deviation rate;

[0103] S405, determine whether the merged deviation rate is greater than or equal to the preset deviation rate threshold;

[0104] S406 If the value is greater than or equal to the value, the merged remote sensing images will be split sequentially according to the reverse order of merging; if the value is less than the value, the small area remote sensing images will continue to be merged.

[0105] It should be noted that by merging remote sensing images, the information reflected by the remote sensing images can be made closer to the actual growth status of the camellia oleifera forest, thereby enabling better prediction of the growth of the camellia oleifera forest and improving the prediction accuracy.

[0106] According to an embodiment of the present invention, several small-area remote sensing images are sequentially merged according to a predetermined number to obtain a merged remote sensing image, including:

[0107] Obtain the number of merged remote sensing images of a small area;

[0108] Calculate the difference between the merged quantity value and the preset threshold;

[0109] If the number is less than the preset threshold, the corresponding number of small area remote sensing images will be merged and fused.

[0110] If the difference is greater than the value, the remote sensing image of the small area will be split into smaller parts with the same number of differences.

[0111] It should be noted that by merging small-area remote sensing images, the number of merged small-area remote sensing images is kept equal to the predicted threshold. This ensures that the merged remote sensing images can accurately reflect the distribution and status information of the camellia oleifera forest, making the predicted growth results of the camellia oleifera forest closer to the actual values.

[0112] According to an embodiment of the present invention, if the difference is greater than the specified value, then the remote sensing images of small regions with a number equal to the difference are split, including:

[0113] Obtain the attribute information of a small area remote sensing image, calculate the similarity between the attribute information of the small area remote sensing image and the attribute information of the merged remote sensing image, and obtain the similarity score.

[0114] Determine whether the similarity is greater than a preset similarity threshold;

[0115] If the difference is greater than the value, the number of splits is generated based on the difference, and the small remote sensing images of the same number of regions that are finally merged are split sequentially based on the number of splits.

[0116] If the value is smaller than the target value, the corresponding small area of ​​remote sensing imagery will be split.

[0117] It should be noted that the splitting of remote sensing images can be understood as excessive fusion, and then the excess fused small areas of remote sensing images are split off again to ensure that the attribute deviation of the merged remote sensing images is small.

[0118] Please refer to Figure 5 , Figure 5 This is a flowchart illustrating the adjustment of planting status in a camellia oleifera forest, based on a satellite remote sensing image recognition method for predicting growth trends, as described in some embodiments of this application. According to embodiments of the present invention, remote sensing images are acquired, segmentation rules are set, and semantic segmentation is performed on the remote sensing images according to the segmentation rules to obtain several sub-regions, including:

[0119] S501, extract feature data from several sub-regions, input the feature data from several sub-regions into the preset camellia oleifera forest growth model, and obtain camellia oleifera forest growth information;

[0120] S502, Generate information on the production status and yield of camellia oleifera forests based on the growth information of the camellia oleifera forests;

[0121] S503 generates growth prediction information based on information on the condition and yield of tea leaves in camellia oleifera forests.

[0122] S504, compare the growth prediction information with the preset growth information to obtain the growth deviation rate;

[0123] S505, determine whether the growth deviation rate is greater than or equal to the preset growth deviation rate threshold;

[0124] S506: If it is greater than or equal to, then generate camellia oleifera forest adjustment information and adjust the planting status of the camellia oleifera forest according to the camellia oleifera forest adjustment information; if it is less than, then determine that the growth prediction of the camellia oleifera forest meets the requirements.

[0125] It should be noted that the planting of camellia oleifera forests is adjusted based on the growth prediction results to ensure the planting status of the forests, and to more accurately predict their growth, thereby improving the accuracy of the predictions.

[0126] According to an embodiment of the present invention, if the value is greater than or equal to the value, then camellia oleifera forest adjustment information is generated, and the planting status of the camellia oleifera forest is adjusted according to the camellia oleifera forest adjustment information, including:

[0127] The planting status of camellia oleifera forests includes the planting light conditions, planting temperature, types of fertilizers, fertilizer ratios between different fertilizers, planting time, and planting spacing.

[0128] It should be noted that different planting temperatures can also affect the growth rate and growth status of camellia oleifera forests. Soil temperature data and environmental temperature data under different geological or geomorphological conditions are obtained based on historical big data.

[0129] Based on soil temperature data and ambient temperature data, parameter variables are generated, the growth prediction model is optimized, and the growth relationship curve between soil temperature data and camellia oleifera forest is generated, which is denoted as the first growth curve, and the growth relationship curve between ambient temperature data and camellia oleifera forest is denoted as the second growth curve.

[0130] The temperature difference is obtained by determining the difference between soil temperature data and ambient temperature data.

[0131] The Euclidean distance between the first growth curve and the second growth curve is determined based on the temperature difference, and the distance information is obtained.

[0132] Determine whether the distance information is less than the preset distance value. If it is less, determine that the first growth curve and the second growth curve are close to the actual growth state of the camellia oleifera forest.

[0133] If the value is greater than or equal to the value, then the growth compensation information is used to adjust the type of fertilizer and the ratio between different fertilizers, thereby correcting the soil temperature.

[0134] It should be noted that when fertilizer is used on camellia oleifera forests, it can change the soil structure and the distribution of microorganisms in the soil, which can appropriately change the internal temperature of the soil. This can help adjust the soil temperature and make the first growth curve and the second growth curve closer to the actual growth state.

[0135] Planting altitude also affects soil temperature and ambient temperature, which can be dynamically adjusted by calculating the planting altitude of the camellia oleifera forest.

[0136] The growth trend of camellia oleifera forest is predicted in real time by a growth prediction model. If the growth trend does not meet the requirements, the actual growth status of the camellia oleifera forest can be adjusted by adjusting the planting spacing, that is, removing some camellia oleifera forests to expand the planting spacing, thereby optimizing and adjusting the planting of camellia oleifera forests in real time.

[0137] According to an embodiment of the present invention, if the value is less than a certain threshold, then several sub-regional remote sensing data are acquired, and the several sub-regional remote sensing data are input into a preset growth prediction model to obtain growth prediction information, including:

[0138] Training data is obtained through big data, and the growth prediction model is iteratively calculated using the training data to generate training results.

[0139] Determine whether the training results have converged;

[0140] If convergence occurs, stop training;

[0141] If convergence is not achieved, the growth prediction model will continue to be iteratively calculated until the training results converge.

[0142] It should be noted that the dataset is divided into three parts: training set, validation set, and test set. The model is iteratively trained on the training set, evaluated on the validation set using the corresponding evaluation metrics, and predicted on the test data on the test set.

[0143] The process of validating the growth prediction model using a validation set is as follows:

[0144] The parameters of the predictive model are evaluated using validation data, and evaluation information is obtained.

[0145] The evaluation information is compared with the preset evaluation information, and it is determined whether the verification data meets the requirements.

[0146] Determine whether the verification data meets the requirements. If it does, the growth prediction model is considered accurate, and an expert database is generated. Based on the expert database, a planting plan for the camellia oleifera forest is generated, which can provide reference data for the next planting season.

[0147] If the requirements are not met, compensation information is generated, and the parameters of the growth prediction model are adjusted based on the compensation information.

[0148] According to an embodiment of the present invention, training set data is obtained through big data, the growth prediction model is iteratively calculated using the training set data, and training results are generated, including:

[0149] Obtain the number of iterations, compare it with the preset number of iterations, and get the difference in the number of iterations;

[0150] Determine if the difference in iteration counts equals the preset number of iterations;

[0151] If equal, then determine whether the iteration result meets the training requirements;

[0152] If the condition is met, the number of iterations will be recorded and stored.

[0153] If they are not equal, calculate the difference in the number of iterations;

[0154] The training set data is adjusted and updated based on the difference in the number of iterations.

[0155] It should be noted that by continuously iterating the calculation of the model, the prediction accuracy of the model is improved, making the prediction results closer to the actual growth results of the camellia oleifera forest.

[0156] According to an embodiment of the present invention, it further includes:

[0157] Acquire remote sensing image information and extract texture features based on the remote sensing image information;

[0158] Based on texture features, the edge regions of camellia oleifera tree branches and trunks are identified, and parameter information of camellia oleifera tree branches and trunks is generated.

[0159] Set a sampling time interval, generate a time threshold, and re-identify branch parameter information;

[0160] By comparing branch parameter information at different time intervals, branch growth information can be obtained.

[0161] The growth trend of camellia oleifera forest is predicted based on the information on branch and trunk growth.

[0162] It should be noted that by extracting the dimensions of the branches and trunks of the camellia oleifera forest at different time periods or in different seasons, the growth status of the camellia oleifera forest can be judged. Based on the changes in dimensions, the growth trend of the camellia oleifera forest can be calculated, realizing multi-parameter judgment and prediction of the growth trend of the camellia oleifera forest. The changes in dimensions include the changes in the length of the branches and trunks or the changes caused by the radial growth of the branches and trunks.

[0163] Please refer to Figure 6 , Figure 6This is a schematic diagram of the structure of a Camellia oleifera forest growth trend prediction system based on satellite remote sensing image recognition, as described in some embodiments of this application. Secondly, embodiments of this application provide a Camellia oleifera forest growth trend prediction system 6 based on satellite remote sensing image recognition. This system includes: a memory 61 and a processor 62. The memory 61 includes a program for a Camellia oleifera forest growth trend prediction method based on satellite remote sensing image recognition. When the program for the Camellia oleifera forest growth trend prediction method based on satellite remote sensing image recognition is executed by the processor, it performs the following steps:

[0164] Acquire remote sensing images, set segmentation rules, and perform semantic segmentation on the remote sensing images according to the segmentation rules to obtain several sub-regions;

[0165] Extract features from several sub-regions, calculate the similarity between the sub-region features and preset feature values, and obtain the feature similarity.

[0166] Determine whether the feature similarity is greater than or equal to a preset similarity threshold;

[0167] If it is greater than or equal to, then correction information is generated, and the segmentation rules are adjusted according to the correction information;

[0168] If it is less than, then obtain remote sensing data of several sub-regions, input the remote sensing data of several sub-regions into the preset growth prediction model, and obtain growth prediction information;

[0169] The growth prediction information is transmitted to the terminal in a predetermined manner.

[0170] It should be noted that remote sensing is a new technology that combines inductive telemetry and resource management monitoring of the Earth's surface (such as resource management of trees, grasslands, soil, water, minerals, crops, fish, and wildlife) using telemetry instruments on platforms such as artificial Earth satellites and aircraft. Remote sensing refers to all non-contact, long-distance detection technologies. It utilizes modern vehicles and sensors to acquire the electromagnetic wave characteristics of target objects from a distance, and through the transmission, storage, satellite imagery, correction, and identification of this information, it ultimately achieves its function.

[0171] According to an embodiment of the present invention, the steps of acquiring remote sensing images, setting segmentation rules, and performing semantic segmentation on the remote sensing images according to the segmentation rules to obtain several sub-regions include:

[0172] Acquire remote sensing images, randomly divide the remote sensing images into several image domains, and calculate the pixel grayscale values ​​of the image domains;

[0173] The grayscale change rate is obtained by comparing the grayscale values ​​of two neighboring image domain pixels with the first grayscale mean.

[0174] Determine whether the grayscale change rate is greater than a preset grayscale change rate threshold;

[0175] If the area is greater than the specified value, the image domain will be segmented and the area adjusted accordingly.

[0176] If it is less than, the image domains are superimposed to calculate the superimposed area of ​​the image domains.

[0177] It should be noted that region-based segmentation methods use defined rules to aggregate images into different sub-regions. Region-based segmentation methods can better filter noise, are suitable for images with significant intensity changes, and are less time-consuming.

[0178] According to an embodiment of the present invention, if the area is greater than a certain threshold, then adjusting the segmented area of ​​the image domain includes:

[0179] Obtain the size of the segmented area, calculate the difference between the segmented area and the preset area value, and obtain the segmented area difference.

[0180] Determine whether the difference in the segmented areas is greater than a preset area threshold;

[0181] If it is greater than, a second grayscale mean is generated, and the remote sensing image is segmented a second time based on the second grayscale mean.

[0182] If it is less than, then a grayscale mean adjustment amount is generated, and the first grayscale mean is adjusted by the same amount based on the grayscale mean adjustment amount.

[0183] It should be noted that the threshold segmentation algorithm is based on the segmented area of ​​the image and inputs it into the prediction model to identify sub-regions. This method first determines one or more thresholds and compares the area of ​​the sub-region with the area threshold to obtain the specific category of the sub-region. It can handle situations that are prone to pixel value changes due to noise, illumination and other factors.

[0184] According to an embodiment of the present invention, the steps of acquiring remote sensing images, setting segmentation rules, and performing semantic segmentation on the remote sensing images according to the segmentation rules to obtain several sub-regions include:

[0185] The remote sensing image is divided into several small-area remote sensing images using a set segmentation algorithm.

[0186] Several small-area remote sensing images are merged sequentially according to a predetermined number to obtain merged remote sensing images;

[0187] Obtain parameter information for merged remote sensing images.

[0188] The merged remote sensing image parameter information is compared with the preset remote sensing image parameter information to obtain the merging deviation rate;

[0189] Determine whether the merged deviation rate is greater than or equal to a preset deviation rate threshold;

[0190] If the value is greater than or equal to the value, the merged remote sensing images will be split sequentially according to the reverse order of merging.

[0191] If the value is smaller, then continue merging the remote sensing images of the small area.

[0192] It should be noted that by merging remote sensing images, the information reflected by the remote sensing images can be made closer to the actual growth status of the camellia oleifera forest, thereby enabling better prediction of the growth of the camellia oleifera forest and improving the prediction accuracy.

[0193] According to an embodiment of the present invention, several small-area remote sensing images are sequentially merged according to a predetermined number to obtain a merged remote sensing image, including:

[0194] Obtain the number of merged remote sensing images of a small area;

[0195] Calculate the difference between the merged quantity value and the preset threshold;

[0196] If the number is less than the preset threshold, the corresponding number of small area remote sensing images will be merged and fused.

[0197] If the difference is greater than the value, the remote sensing image of the small area will be split into smaller parts with the same number of differences.

[0198] It should be noted that by merging small-area remote sensing images, the number of merged small-area remote sensing images is kept equal to the predicted threshold. This ensures that the merged remote sensing images can accurately reflect the distribution and status information of the camellia oleifera forest, making the predicted growth results of the camellia oleifera forest closer to the actual values.

[0199] According to an embodiment of the present invention, if the difference is greater than the specified value, then the remote sensing images of small regions with a number equal to the difference are split, including:

[0200] Obtain the attribute information of a small area remote sensing image, calculate the similarity between the attribute information of the small area remote sensing image and the attribute information of the merged remote sensing image, and obtain the similarity score.

[0201] Determine whether the similarity is greater than a preset similarity threshold;

[0202] If the difference is greater than the value, the number of splits is generated based on the difference, and the small remote sensing images of the same number of regions that are finally merged are split sequentially based on the number of splits.

[0203] If the value is smaller than the target value, the corresponding small area of ​​remote sensing imagery will be split.

[0204] It should be noted that the splitting of remote sensing images can be understood as excessive fusion, and then the excess fused small areas of remote sensing images are split off again to ensure that the attribute deviation of the merged remote sensing images is small.

[0205] According to an embodiment of the present invention, a remote sensing image is acquired, a segmentation rule is set, and semantic segmentation is performed on the remote sensing image according to the segmentation rule to obtain several sub-regions, including:

[0206] Extract feature data from several sub-regions, input the feature data from several sub-regions into a preset camellia oleifera forest growth model, and obtain camellia oleifera forest growth information;

[0207] Generate information on the production status and yield of camellia oleifera forests based on their growth information.

[0208] Growth prediction information is generated based on information on the condition and yield of tea plants in camellia oleifera forests.

[0209] The growth prediction information is compared with the preset growth information to obtain the growth deviation rate;

[0210] Determine whether the growth deviation rate is greater than or equal to the preset growth deviation rate threshold;

[0211] If the value is greater than or equal to the value, adjustment information for the camellia oleifera forest is generated, and the planting status of the camellia oleifera forest is adjusted according to the adjustment information; if the value is less than the value, it is determined that the growth prediction of the camellia oleifera forest meets the requirements.

[0212] It should be noted that the planting of camellia oleifera forests is adjusted based on the growth prediction results to ensure the planting status of the forests, and to more accurately predict their growth, thereby improving the accuracy of the predictions.

[0213] According to an embodiment of the present invention, if the value is greater than or equal to the value, then camellia oleifera forest adjustment information is generated, and the planting status of the camellia oleifera forest is adjusted according to the camellia oleifera forest adjustment information, including:

[0214] The planting status of camellia oleifera forests includes the planting light conditions, planting temperature, types of fertilizers, fertilizer ratios between different fertilizers, planting time, and planting spacing.

[0215] It should be noted that different planting temperatures can also affect the growth rate and growth status of camellia oleifera forests. Soil temperature data and environmental temperature data under different geological or geomorphological conditions are obtained based on historical big data.

[0216] Based on soil temperature data and ambient temperature data, parameter variables are generated, the growth prediction model is optimized, and the growth relationship curve between soil temperature data and camellia oleifera forest is generated, which is denoted as the first growth curve, and the growth relationship curve between ambient temperature data and camellia oleifera forest is denoted as the second growth curve.

[0217] The temperature difference is obtained by determining the difference between soil temperature data and ambient temperature data.

[0218] The Euclidean distance between the first growth curve and the second growth curve is determined based on the temperature difference, and the distance information is obtained.

[0219] Determine whether the distance information is less than the preset distance value. If it is less, determine that the first growth curve and the second growth curve are close to the actual growth state of the camellia oleifera forest.

[0220] If the value is greater than or equal to the value, then the growth compensation information is used to adjust the type of fertilizer and the ratio between different fertilizers, thereby correcting the soil temperature.

[0221] It should be noted that the growth of Camellia oleifera forests is characterized by multiple parameters such as growth height, trunk length, and radial dimensions of the trunk. Furthermore, it is understood that the growth characteristics of Camellia oleifera forests are not limited to these few. Those skilled in the art can make reasonable inferences and use experience to select or adjust the characteristics during the actual judgment process.

[0222] It should be noted that when fertilizer is used on camellia oleifera forests, it can change the soil structure and the distribution of microorganisms in the soil, which can appropriately change the internal temperature of the soil. This can help adjust the soil temperature and make the first growth curve and the second growth curve closer to the actual growth state.

[0223] Planting altitude also affects soil temperature and ambient temperature, which can be dynamically adjusted by calculating the planting altitude of the camellia oleifera forest.

[0224] The growth trend of camellia oleifera forest is predicted in real time by a growth prediction model. If the growth trend does not meet the requirements, the actual growth status of the camellia oleifera forest can be adjusted by adjusting the planting spacing, that is, removing some camellia oleifera forests to expand the planting spacing, thereby optimizing and adjusting the planting of camellia oleifera forests in real time.

[0225] According to an embodiment of the present invention, if the value is less than a certain threshold, then several sub-regional remote sensing data are acquired, and the several sub-regional remote sensing data are input into a preset growth prediction model to obtain growth prediction information, including:

[0226] Training data is obtained through big data, and the growth prediction model is iteratively calculated using the training data to generate training results.

[0227] Determine whether the training results have converged;

[0228] If convergence occurs, stop training;

[0229] If convergence is not achieved, the growth prediction model will continue to be iteratively calculated until the training results converge.

[0230] It should be noted that the dataset is divided into three parts: training set, validation set, and test set. The model is iteratively trained on the training set, evaluated on the validation set using the corresponding evaluation metrics, and predicted on the test data on the test set.

[0231] The process of validating the growth prediction model using a validation set is as follows:

[0232] Obtain validation set data, input the validation set data into the growth prediction model, perform iterative prediction, and obtain validation data;

[0233] The parameters of the predictive model are evaluated using validation data, and evaluation information is obtained.

[0234] The evaluation information is compared with the preset evaluation information, and it is determined whether the verification data meets the requirements.

[0235] Determine whether the verification data meets the requirements. If it does, the growth prediction model is considered accurate, and an expert database is generated. Based on the expert database, a planting plan for the camellia oleifera forest is generated, which can provide reference data for the next planting season.

[0236] If the requirements are not met, compensation information is generated, and the parameters of the growth prediction model are adjusted based on the compensation information.

[0237] According to an embodiment of the present invention, training set data is obtained through big data, the growth prediction model is iteratively calculated using the training set data, and training results are generated, including:

[0238] Obtain the number of iterations, compare it with the preset number of iterations, and get the difference in the number of iterations;

[0239] Determine if the difference in iteration counts equals the preset number of iterations;

[0240] If equal, then determine whether the iteration result meets the training requirements;

[0241] If the condition is met, the number of iterations will be recorded and stored.

[0242] If they are not equal, calculate the difference in the number of iterations;

[0243] The training set data is adjusted and updated based on the difference in the number of iterations.

[0244] It should be noted that by continuously iterating the calculation of the model, the prediction accuracy of the model is improved, making the prediction results closer to the actual growth results of the camellia oleifera forest.

[0245] According to an embodiment of the present invention, it further includes:

[0246] Acquire remote sensing image information and extract texture features based on the remote sensing image information;

[0247] Based on texture features, the edge regions of camellia oleifera tree branches and trunks are identified, and parameter information of camellia oleifera tree branches and trunks is generated.

[0248] Set a sampling time interval, generate a time threshold, and re-identify branch parameter information;

[0249] By comparing branch parameter information at different time intervals, branch growth information can be obtained.

[0250] The growth trend of camellia oleifera forest is predicted based on the information on branch and trunk growth.

[0251] It should be noted that by extracting the dimensions of the branches and trunks of the camellia oleifera forest at different time periods or in different seasons, the growth status of the camellia oleifera forest can be judged. Based on the changes in dimensions, the growth trend of the camellia oleifera forest can be calculated, realizing multi-parameter judgment and prediction of the growth trend of the camellia oleifera forest. The changes in dimensions include the changes in the length of the branches and trunks or the changes caused by the radial growth of the branches and trunks.

[0252] This invention discloses a method and system for predicting the growth trend of camellia oleifera forests based on satellite remote sensing image recognition. The method involves acquiring remote sensing images, setting segmentation rules, and performing semantic segmentation on the remote sensing images according to these rules to obtain several sub-regions. Features of these sub-regions are extracted, and the similarity between these features and preset feature values ​​is calculated to obtain feature similarity. It is then determined whether the feature similarity is greater than or equal to a preset similarity threshold. If it is greater than or equal to, correction information is generated, and the segmentation rules are adjusted based on this correction information. If it is less than, remote sensing data of these sub-regions is acquired and input into a preset growth prediction model to obtain growth prediction information. This growth prediction information is then transmitted to a terminal according to a predetermined method. By performing semantic segmentation on the remote sensing images and using the growth prediction model, the growth trend of the camellia oleifera forest can be predicted in real time, thereby improving prediction accuracy.

[0253] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0254] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0255] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0256] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0257] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

Claims

1. A method for predicting the growth trend of a tea-oil camellia forest based on satellite remote sensing image recognition, characterized in that, The method comprises the following steps: acquiring remote sensing images, setting segmentation rules, performing semantic segmentation on the remote sensing images according to the segmentation rules, and obtaining a plurality of sub-regions; extracting sub-region features, and performing similarity calculation on the sub-region features and preset feature values to obtain feature similarity; determining whether the feature similarity is greater than or equal to a preset similarity threshold value; if yes, generating correction information, and adjusting the segmentation rules according to the correction information; if no, acquiring remote sensing data of the plurality of sub-regions, inputting the remote sensing data of the plurality of sub-regions into a preset growth prediction model, and obtaining growth prediction information; transmitting the growth prediction information to a terminal in a predetermined manner; the method of acquiring remote sensing images, setting segmentation rules, and performing semantic segmentation on the remote sensing images according to the segmentation rules to obtain a plurality of sub-regions comprises the following steps: acquiring remote sensing images, randomly segmenting the remote sensing images into a plurality of image domains, and calculating image domain pixel grayscale values; comparing the pixel grayscale values of two image domains in a neighborhood with a first grayscale mean value to obtain a grayscale change rate; determining whether the grayscale change rate is greater than a preset grayscale change rate threshold value; if yes, adjusting the segmentation area of the image domain; if no, performing superposition calculation on the image domain to obtain an image domain superposition area; if yes, adjusting the segmentation area of the image domain, comprising the following steps: acquiring a segmentation area size, performing difference calculation on the segmentation area and a preset area value to obtain a segmentation area difference value; determining whether the segmentation area difference value is greater than a preset area threshold value; if yes, generating a second grayscale mean value, and performing secondary segmentation on the remote sensing images according to the second grayscale mean value; if no, generating a grayscale mean value adjustment amount, and adjusting the first grayscale mean value by the grayscale mean value adjustment amount.

2. The method according to claim 1, wherein the method is characterized by, the method of acquiring remote sensing images, setting segmentation rules, and performing semantic segmentation on the remote sensing images according to the segmentation rules to obtain a plurality of sub-regions comprises the following steps: segmenting the remote sensing images into a plurality of small-area remote sensing images through a set segmentation algorithm; merging the plurality of small-area remote sensing images in a predetermined number to obtain merged remote sensing images; acquiring merged remote sensing image parameter information, comparing the merged remote sensing image parameter information with preset remote sensing image parameter information to obtain a merging deviation rate; determining whether the merging deviation rate is greater than or equal to a preset deviation rate threshold value; if yes, sequentially splitting the merged remote sensing images in reverse order; if no, continue to merge the small-area remote sensing images.

3. The method according to claim 2, wherein the method comprises the following steps: 1) obtaining satellite remote sensing images of the tea-oil camellia forest; 2) determining the growth trend of the tea-oil camellia forest based on the satellite remote sensing images. the method of merging the plurality of small-area remote sensing images in a predetermined number to obtain merged remote sensing images comprises the following steps: acquiring a small-area remote sensing image merging number value; performing number difference calculation on the merging number value and a preset threshold value; if the merging number value is less than the preset threshold value, merging a corresponding number of small-area remote sensing images; if the merging number value is greater than the preset threshold value, splitting a number of small-area remote sensing images equal to the difference value.

4. The method according to claim 3, wherein the method comprises the following steps: 1) obtaining satellite remote sensing images of the tea-oil camellia forest; 2) determining the growth trend of the tea-oil camellia forest based on the satellite remote sensing images. the method of splitting a number of small-area remote sensing images equal to the difference value if the merging number value is greater than the preset threshold value comprises the following steps: Obtaining small area remote sensing image attribute information, calculating the similarity between the small area remote sensing image attribute information and the merged remote sensing image attribute information to obtain a similarity; Judging whether the similarity is greater than a preset similarity threshold value; If greater, generating a number of splits according to the difference, and sequentially splitting the last fused same number of small area remote sensing images according to the number of splits; If less, splitting the corresponding small area remote sensing image.

5. A system for predicting the growth trend of a tea-oil camellia forest based on satellite remote sensing image recognition, characterized in that, The system comprises a memory and a processor, the memory comprising a program of an oil tea forest growth trend prediction method based on satellite remote sensing image recognition, the program of the oil tea forest growth trend prediction method based on satellite remote sensing image recognition being implemented by the processor to realize the following steps: Obtaining a remote sensing image, setting a segmentation rule, performing semantic segmentation on the remote sensing image according to the segmentation rule to obtain a plurality of sub-regions; Extracting a plurality of sub-region features, calculating the similarity between the sub-region features and a preset feature value to obtain a feature similarity; Judging whether the feature similarity is greater than or equal to a preset similarity threshold value; If greater than or equal to, generating correction information and adjusting the segmentation rule according to the correction information; If less, obtaining a plurality of sub-region remote sensing data, inputting the plurality of sub-region remote sensing data into a preset growth prediction model to obtain growth prediction information; Transmitting the growth prediction information to a terminal in a predetermined manner; The obtaining of the remote sensing image, the setting of the segmentation rule, and the semantic segmentation of the remote sensing image according to the segmentation rule to obtain a plurality of sub-regions comprises: Obtaining a remote sensing image, randomly segmenting the remote sensing image into a plurality of image domains, and calculating the pixel gray value of the image domain; Comparing the pixel gray values of two image domains in the neighborhood with a first gray mean value to obtain a gray change rate; Judging whether the gray change rate is greater than a preset gray change rate threshold value; If greater, adjusting the segmentation area of the image domain; If less, performing superposition calculation on the image domain to obtain an image domain superposition area; If greater, adjusting the segmentation area of the image domain, comprising: Obtaining a segmentation area size, calculating the difference between the segmentation area and a preset area value to obtain a segmentation area difference; Judging whether the segmentation area difference is greater than a preset area threshold value; If greater, generating a second gray mean value and performing secondary segmentation on the remote sensing image according to the second gray mean value; If less, generating a gray mean value adjustment amount and adjusting the first gray mean value by the same amount according to the gray mean value adjustment amount. 6.The system for predicting growth trend of Camellia oleifera forest based on satellite remote sensing image recognition according to claim 5, characterized in that, The obtaining of the remote sensing image, the setting of the segmentation rule, and the semantic segmentation of the remote sensing image according to the segmentation rule to obtain a plurality of sub-regions comprises: Segmenting the remote sensing image into a plurality of small area remote sensing images through a set segmentation algorithm; Merging the plurality of small area remote sensing images in a predetermined number to obtain a merged remote sensing image; Obtaining merged remote sensing image parameter information, Comparing the merged remote sensing image parameter information with preset remote sensing image parameter information to obtain a merging deviation rate; Judging whether the merging deviation rate is greater than or equal to a preset deviation rate threshold value; If greater than or equal to, sequentially splitting the merged remote sensing image according to the inverse order principle of merging; If not, continue to merge the small area remote sensing image.

Images