A land type identification management method and system based on unmanned aerial vehicle assistance

Abstract

The application discloses a kind of based on unmanned aerial vehicle auxiliary land type identification management method and system, it is related to land type identification technical field, including through unmanned aerial vehicle collection land remote sensing image, and processing and feature extraction, strengthen the land type features in image;Land type identification model is constructed, land type identification standard is set, input is according to the land type feature image after strengthening training and analysis, obtain land type identification result;The identification result is rechecked and corrected, according to the land type after correction is authenticated and managed.The application obtains high-resolution ground surface image using unmanned aerial vehicle, can clearly show the detailed features of ground surface, improve identification accuracy, can obtain image at any time and is not affected by factors such as cloud cover, night lighting etc.Improve identification precision, ensure the accuracy of identification result, help land management and planning, have important application value for disaster emergency, environmental protection and other tasks.

Description

Technical Field

[0001] This invention relates to the field of land type identification technology, and in particular to a land type identification and management method and system based on unmanned aerial vehicle (UAV) assistance. Background Technology

[0002] In recent years, China has elevated optimizing its land space development pattern and strengthening ecological civilization construction to a national strategy. To achieve the rational application of different land types, land classification is necessary. Land types are typically categorized based on factors such as soil, rock, and topography, according to their natural attributes. Land type classification has become a crucial research topic in the field of land use, and remote sensing technology has become one of the important technical means for land type identification.

[0003] Remote sensing technology, a product of the information age, can remotely acquire electromagnetic wave information radiated and reflected by target objects, and ultimately output it as images or videos. Using UAV remote sensing imagery as the initial research object and identifying land types based on the information contained therein has become the main principle and method for land type identification. Summary of the Invention

[0004] This invention addresses the problems existing in current land type identification and management methods. While existing methods can automatically extract features, their performance depends on a large amount of training data. For some rare or complex land types, it may be difficult to obtain sufficient training data, which affects the model's performance. Existing methods typically only perform single-level classification of images.

[0005] Therefore, the problem to be solved by the present invention is to provide a method and system for land type identification and management based on unmanned aerial vehicle (UAV) assistance.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, embodiments of the present invention provide a land type identification and management method based on unmanned aerial vehicle (UAV) assistance, which includes: collecting land remote sensing images through UAVs, processing and extracting features to enhance land type features in the images; constructing a land type identification model, setting land type identification standards, inputting images with enhanced land type features for training and analysis, and obtaining land type identification results; performing secondary verification and correction on the land type identification results, and certifying and managing the land types based on the corrected land types.

[0008] As a preferred embodiment of the UAV-assisted land type identification and management method of the present invention, the land remote sensing image includes unchanged patches and changed patches. The unchanged patches include suspected newly added construction patches, suspected farmland outflow patches, suspected newly added farmland unchanged patches, construction and facility agricultural land change patches, agricultural land change patches, and unused land change patches. The changed patches include updated patches other than the monitored patches.

[0009] As a preferred embodiment of the UAV-assisted land type identification and management method of the present invention, the processing includes color correction, noise reduction and image enhancement.

[0010] The color correction calculation formula is as follows:

[0011] I'(x,y)=D*I(x,y)

[0012]

[0013] In the formula, D is a diagonal matrix, and diag is a diagonalization function. The average value is the white reference. These represent the average pixel values ​​in red, green, and blue, respectively.

[0014] The noise reduction process includes,

[0015] I'(x,y)=∑w(x,y,i,j)*I(i,j)

[0016] w(x,y,i,j)=exp||v(x,y)-v(i,j)|| 2 / h

[0017] In the formula, (i,j) represents the entire image, w is the weight used to represent the similarity between pixels (x,y) and (i,j); v is a small window around pixel (x,y), ||v(x,y)-v(i,j)|| 2 is the Euclidean distance between windows, and h is a parameter that controls the filtering strength;

[0018] The image enhancement includes,

[0019]

[0020] Variables z and Z represent the brightness values ​​before and after image enhancement processing, respectively. i+1 and l i To enhance the stretch range before and after, M i+1 and M i This corresponds to the brightness value range before and after enhancement, where N is the number of segments.

[0021] As a preferred embodiment of the UAV-assisted land type identification and management method of the present invention, the feature extraction related calculation formula is as follows:

[0022] Z[i,j,k]=∑∑∑X[a,b,c]*W[ia,jb,kc]+b[k]

[0023] Z′[i,j,k]=max(X[i*s:i*s+f,j*s:j*s+f,k])

[0024] Z = W i *A p +b

[0025] In the formula, Z[i,j,k] is the convolutional image; X is the input image; W is the convolution kernel; b is the bias; i and j are the spatial positions; and k is the depth position; Z′[i,j,k] is the max-pooled image; s is the stride; f is the pooling window size; Z is the fully connected image; W i It's the weight, A p 'b' is the activation value of the previous layer, and 'b' is the bias.

[0026] As a preferred embodiment of the UAV-assisted land type identification and management method of the present invention, the land type identification model specifically includes: creating a dataset from processed and feature-enhanced land remote sensing images; training on the constructed dataset; removing similar data from the trained dataset; segmenting the remote sensing images of the land to be tested; labeling each land region; extracting the quantitative features of the corresponding land; and performing similarity calculations between the land to be tested and the standard land type. The relevant calculation formulas are as follows:

[0027]

[0028] In the formula, S represents the similarity, and t i To extract land type features from remote sensing images, t j The standard land type features are set; the accuracy of land type identification is calculated.

[0029] As a preferred embodiment of the UAV-assisted land type identification and management method of the present invention, the land type identification standard includes a primary category and a secondary category. After identification by the identification model, if the similarity between the identified type and the land standard type reaches 0.85 or higher, it is initially certified as a primary category. Further identification and classification are performed according to the primary category identification standard to divide it into secondary categories. A preliminary review and correction is performed on the divided secondary categories. If the review result is the same as the preliminary identification result, the identification result is identified as category 1 and recorded. If the review result is different from the preliminary identification result, the primary category identification is re-performed. If the similarity reaches 0.95 or higher, secondary category identification is performed. A preliminary second review and correction is performed on the divided secondary categories. If the second review result is the same as the preliminary identification result, the identification result is identified as category 2 and recorded. If the second review result is different from the preliminary identification result, the identification result is identified as category 3 and recorded, and a second review and correction is performed. If the similarity between the identified type and the land standard type is below 0.85, it cannot be determined as a primary category and is directly transferred to manual review and correction for a second review and correction.

[0030] As a preferred embodiment of the UAV-assisted land type identification and management method of the present invention, the secondary verification and correction includes: manual verification and judgment, on-site inspection of land type identification results, land type authentication through investigation and inquiry, and secondary verification and correction based on the secondary category classification. When the identification type result of the secondary verification and correction is the same as the result of the first verification, the identification result is identified as category 4 and the authentication result is recorded. When the identification type result of the secondary verification and correction is different from the result of the first verification, the identification type result of the secondary verification and correction is taken as the final result, identified as category 5 and recorded.

[0031] Secondly, embodiments of the present invention provide a land type identification and management system based on unmanned aerial vehicle (UAV) assistance, comprising: an acquisition module for acquiring land remote sensing images via UAV and inputting the collected images into an image processing module via a communication network; an image processing module for processing and feature extraction of the collected images, processing the images into images with enhanced features, and sending them to the identification model in the identification module for identification; an identification module for identifying the images processed and feature-extracted by the image processing module, calculating similarity, performing preliminary identification and land type determination, and obtaining a preliminary category determination result; and a verification and correction module for verifying and correcting the preliminary category determination result, checking for errors, and if the correction result is the same as the preliminary land category determination result, then the land type identification result is considered correct; if the correction result is different from the preliminary land category determination result, then the verification and correction result shall prevail.

[0032] Thirdly, embodiments of the present invention provide a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement any of the steps of the above-described UAV-assisted land type identification and management method.

[0033] Fourthly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any of the steps of the above-described UAV-assisted land type identification and management method.

[0034] The beneficial effects of this invention are that it utilizes drones to acquire high-resolution surface images, clearly displaying detailed surface features and thus improving identification accuracy. Furthermore, drones can acquire images at any time, unaffected by cloud cover, nighttime lighting, or other factors. This allows for more accurate identification of different land types. After determining the primary category, a model specifically trained for that category can be used to determine the secondary category, further improving identification accuracy. Verification further ensures the accuracy of the identification results. This improves the accuracy and efficiency of land type identification, contributing to land management and planning. Simultaneously, the use of drones allows for the acquisition of real-time surface information, which has significant application value for disaster emergency response, environmental protection, and other tasks. Attached Figure Description

[0035] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0036] Figure 1 This is a flowchart of a land type identification and management method based on drones. Detailed Implementation

[0037] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0038] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0039] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0040] Example 1

[0041] Reference Figure 1 This is the first embodiment of the present invention, which provides a land type identification and management method based on unmanned aerial vehicle (UAV) assistance, including:

[0042] S1: Collect land remote sensing images using drones, and process and extract features to enhance land type characteristics in the images.

[0043] Specifically, land remote sensing images include unchanged and changed land features. Unchanged land features include suspected newly added construction features, suspected farmland outflow features, suspected newly added farmland features without changes, features showing changes in construction and facility agricultural land use, farmland change features, and unused land change features. Changed land features include updated features other than those monitored. Processing includes color correction, noise reduction, and image enhancement. The color correction calculation formula is as follows:

[0044] I'(x,y)=D*I(x,y)

[0045] D=diag([mean(Iw) / mean(I(x,y) R ),mean(Iw) / mean(I(x,y) G ),mean(Iw) / mean(I(x,y) B In the formula, D is a diagonal matrix, diag is a diagonalization function, and I... w The white area serves as a reference, and "mean" is the function used to calculate the mean.

[0046] I(x,y) R ,I(x,y) G ,I(x,y) B These represent the pixel values ​​of the image in red, green, and blue, respectively.

[0047] Noise is frequently present in raw remote sensing image data and is one of the main factors affecting image processing results. The main sources of image noise are external and internal influences encountered during image acquisition, transmission, reception, and processing. For example, during image acquisition, external lighting or other weather conditions, as well as the operating status of the equipment itself, are major factors contributing to image noise. During transmission, channel interference and the influence of external lighting and humidity during wireless transmission can generate image noise and reduce image quality. Human error can occur at each stage; therefore, noise is inherent in image acquisition. Noise reduction processing includes…

[0048] I'(x,y)=∑w(x,y,i,j)*I(i,j)

[0049] w(x,y,i,j)=exp||v(x,y)-v(i,j)|| 2 / h

[0050] In the formula, (i,j) represents the entire image, w is the weight used to represent the similarity between pixels (x,y) and (i,j); v is a small window around pixel (x,y), ||v(x,y)-v(i,j)|| 2 is the Euclidean distance between windows, and h is a parameter that controls the filtering strength;

[0051] Image enhancement includes,

[0052]

[0053] Variables z and Z represent the brightness values ​​before and after image enhancement processing, respectively. i+1 and l i To enhance the stretch range before and after, M i+1 and M i This corresponds to the brightness value range before and after enhancement, where N is the number of segments.

[0054] The relevant calculation formulas for feature extraction are as follows:

[0055] Z[i,j,k]=∑∑∑X[a,b,c]*W[ia,jb,kc]+b[k]

[0056] Z′[i,j,k]=max(X[i*s:i*s+f,j*s:j*s+f,k])

[0057] Z = W i *A p +b

[0058] In the formula, Z[i,j,k] is the convolutional image; X is the input image; W is the convolution kernel; b is the bias; i and j are the spatial positions; and k is the depth position; Z′[i,j,k] is the max-pooled image; s is the stride; f is the pooling window size; Z is the fully connected image; W i It's the weight, A p 'b' is the activation value of the previous layer, and 'b' is the bias.

[0059] S2: Construct a land type recognition model, set land type recognition standards, input the enhanced land type feature image for training and analysis, and obtain the land type recognition result.

[0060] Specifically, S2.1: Create a dataset from the processed and feature-enhanced land remote sensing images, train on the constructed dataset, and remove similar data from the trained dataset.

[0061] Furthermore, the acquired land topography remote sensing images are used for training, and the trained image dataset is set as: Y Z =[Y1,···,Y n1 ] T In the formula, Y Z This represents the trained remote sensing image data, where T represents the transpose. Because the trained land topography remote sensing image feature data contains some similar data, which affects subsequent classification, similar data is removed. The relevant calculation formula is as follows:

[0062] S(y ij ,y ik )=αS A (y ij ,y ik )+βS b (y ij ,y ik )

[0063] In the formula, S(y) ij ,y ik ) represents the data y ij and y ik The degree of similarity; α is the data similarity coefficient, and β is the image similarity relationship.

[0064] S2.2: Segment the remote sensing image of the land to be measured, label each land area, extract the corresponding quantitative features of the land, and perform similarity calculation between the land to be measured and the standard land type. The relevant calculation formulas are as follows:

[0065]

[0066] In the formula, S represents the similarity, and t iTo extract land type features from remote sensing images, t j The standard land type characteristics are set.

[0067] Land type identification criteria include primary categories and secondary categories; such as

[0068] Table 1 Land Category Identification Standards

[0069]

[0070]

[0071]

[0072] S2.3: Calculate the accuracy of land type identification in the detection.

[0073] Specifically, the effectiveness of land type identification is measured by identification performance, which includes two main factors: accuracy and success rate. To calculate the accuracy and success rate of land type identification, representative study areas were selected from UAV imagery data for accuracy evaluation. Accuracy refers to the degree of consistency between the identified image and the standard reference image in each test sample image, quantified by the K-value, which can be expressed as:

[0074]

[0075]

[0076] In the formula, n is the number of land categories, N is the number of remote sensing samples, and a is the number of correctly identified samples. i+ The quantity is the total number of pixels in the i-th row, a +i The number represents the total number of pixels in the i-th column; a, b, and c represent the number of correctly identified, unidentified, and incorrectly identified samples, respectively. The K value and η... R The larger the value of η, the better the recognition effect. L and η I The larger the value, the greater the recognition error, meaning the worse the recognition effect.

[0077] After the primary category is identified by the identification model, if the similarity between the identified type and the land standard type reaches 0.85 or higher, it is initially certified as a primary category. Further identification and classification are then performed according to the primary category identification criteria to divide it into secondary categories. A preliminary review and correction is conducted on the divided secondary categories. If the review result is the same as the preliminary identification result, the identification result is classified as category 1 and recorded. If the review result differs from the preliminary identification result, the primary category identification is re-performed. If the similarity reaches 0.95 or higher, secondary category identification is performed. A preliminary second review and correction is conducted on the divided secondary categories. If the second review result is the same as the preliminary identification result, the identification result is classified as category 2 and recorded. If the second review result differs from the preliminary identification result, the identification result is classified as category 3 and recorded, and a second review and correction is performed. If the similarity between the identified type and the land standard type is below 0.85, it cannot be determined as a primary category and is directly transferred to manual review and correction for a second review and correction.

[0078] S3: Perform a second review and correction on the land type identification results, and then authenticate and manage the land type based on the corrected land type.

[0079] Specifically, the process involves manual review and judgment, on-site verification of land type identification results, and land category certification through investigation and inquiry. A second review and correction is conducted based on the secondary category classification. When the identification type result of the second review and correction is the same as the result of the first review, the identification result is identified as category 4 and the certification result is recorded. When the identification type result of the second review and correction is different from the result of the first review, the identification type result of the second review and correction is taken as the final result, identified as category 5, and recorded.

[0080] Furthermore, this embodiment also provides a land type identification and management system based on UAV assistance, including: an acquisition module for acquiring land remote sensing images via UAV and inputting the collected images into an image processing module via a communication network; an image processing module for processing and feature extraction of the collected images, processing the images into images with enhanced features, and sending them to the identification model in the identification module for identification; an identification module for identifying the images processed and feature-extracted by the image processing module, calculating similarity, performing preliminary identification and land type determination, and obtaining a preliminary category determination result; and a verification and correction module for verifying and correcting the preliminary category determination result, checking for errors, and if the correction result is the same as the preliminary land category determination result, the land type identification result is considered correct; if the correction result is different from the preliminary land category determination result, the verification and correction result shall prevail.

[0081] This embodiment also provides a computer device applicable to the land type identification and management method based on UAV-assisted methods, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement all or part of the steps of the method described in the above embodiments of the present invention.

[0082] This embodiment also provides a storage medium on which a computer program is stored. When the computer program is executed by a processor, it performs the method in any optional implementation of the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0083] The storage medium proposed in this embodiment and the data storage method proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0084] Example 2

[0085] Referring to Table 2, which is the second embodiment of the present invention, this embodiment provides a land type identification and management method based on unmanned aerial vehicle (UAV) assistance. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0086] The initial data for the experiment were prepared from two aspects: UAV remote sensing image sample data and standard feature data. The remote sensing image sample data needed to be collected on-site using UAV equipment. A hexagonal UAV was selected as the initial remote sensing image acquisition device. This device has a wheelbase of 1200mm, an empty takeoff weight of 9.5kg, and a maximum takeoff weight of 175kg. It is powered by a 22000mAh battery, providing approximately 50 minutes of flight time and a maximum flight altitude of 120m. A controller was embedded in the UAV, and the relevant flight trajectory control program was imported into it based on the geographical distribution of the study area. The UAV's trajectory control accuracy in both the horizontal and vertical directions was ±0.5m. After UAV flight and remote sensing image acquisition, four bands of remote sensing images with a spatial resolution of 0.4m or higher were selected as research samples, and some detection results were used for illustration.

[0087] Table 2 shows the detection results for partial land type identification.

[0088] Land type K value <![CDATA[η R ]]> <![CDATA[η L ]]> <![CDATA[η I ]]> paddy field 0.886 86.72 6.85 3.88 tea garden 0.921 89.62 4.21 3.27 natural pasture 0.874 90.15 3.23 2.21 Bamboo Forest 0.893 95.21 4.51 4.07

[0089] In summary, the detection model used in this method can effectively detect and identify land types, and the K value and η... R All are above 85%, indicating high detection precision and accuracy. L and η I Most of the results were below 5%, indicating that the method has a small error in the recognition results and a good recognition effect.

[0090] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A land type identification and management method based on unmanned aerial vehicle (UAV) assistance, characterized in that: include, Land remote sensing images are collected by drones, and then processed and feature extracted to enhance the land type characteristics in the images; A land type identification model is constructed, land type identification standards are set, and the model is trained and analyzed based on the enhanced land type feature image to obtain the land type identification result. The land type identification results are reviewed and corrected a second time, and the land types are then certified and managed based on the corrected land types. The land type identification model specifically includes, A dataset is created from processed and feature-enhanced land remote sensing images. The dataset is then used for training, and similar data in the trained dataset is removed. The remote sensing image of the land to be measured is segmented, and each land area is labeled. The quantitative features of the corresponding land are extracted, and similarity calculations are performed between the land to be measured and the standard land type. The relevant calculation formulas are as follows: In the formula, S represents the similarity. To extract land type features from remote sensing images, Standard land type characteristics are set; Calculate the accuracy of land type identification in the detection process; The land type identification criteria include primary categories and secondary categories; After the primary category is identified by the identification model, if the similarity between the identified type and the land standard type reaches 0.85 or higher, it is initially certified as a primary category. Further identification and classification are performed according to the primary category identification criteria to divide it into secondary categories. A preliminary review and correction is conducted on the divided secondary categories. If the review result is the same as the preliminary identification result, the identification result is classified as category 1 and recorded. If the review result is different from the preliminary identification result, the primary category identification is re-performed. If the similarity reaches 0.95 or higher, secondary category identification is performed. A preliminary second review and correction is conducted on the divided secondary categories. If the second review result is the same as the preliminary identification result, the identification result is classified as category 2 and recorded. If the second review result is different from the preliminary identification result, the identification result is classified as category 3 and recorded, and a second review and correction is performed. When the similarity between the identified type and the land standard type is below 0.85, it cannot be determined as a first-level category and is directly transferred to manual review and correction for a second review and correction.

2. The land type identification and management method based on UAV assistance as described in claim 1, characterized in that: The land remote sensing image includes unchanged patches and changed patches. The unchanged patches include suspected newly added construction patches, suspected farmland outflow patches, suspected newly added farmland patches with no changes, patches showing changes in agricultural land for construction and facilities, patches showing changes in agricultural land, and patches showing changes in unused land. The changed patches include updated patches other than the monitored patches.

3. The land type identification and management method based on UAV assistance as described in claim 2, characterized in that: The processing includes color correction, noise reduction, and image enhancement; The color correction calculation formula is as follows: In the formula, D is a diagonal matrix, and diag is a diagonalization function. The average value is the white reference. , , These represent the average pixel values ​​in red, green, and blue, respectively. The noise reduction process includes, In the formula, (i,j) represents the entire image, w is the weight used to represent the similarity between pixels (x,y) and (i,j); v is a small window around pixel (x,y). is the Euclidean distance between windows, and h is a parameter that controls the filtering strength; The image enhancement includes, Variables z and Z represent the brightness values ​​before and after image enhancement, respectively. and To enhance the stretch range before and after, and This corresponds to the brightness value range before and after enhancement, where N is the number of segments.

4. The land type identification and management method based on UAV assistance as described in claim 3, characterized in that: The relevant calculation formulas for feature extraction are as follows: In the formula, The convolutioned image is X; the input image is W, the convolution kernel is b, the bias is i, j are the positions in the spatial dimension, and k is the position in the depth dimension. The image after max pooling; s is the step size, and f is the pooling window size; The image after full connection; It's weight. 'b' is the activation value of the previous layer, and 'b' is the bias.

5. The land type identification and management method based on UAV assistance as described in claim 4, characterized in that: The secondary review and correction includes manual review and judgment, on-site verification of land type identification results, land type certification through investigation and inquiry, and secondary review and correction based on the secondary category classification. When the identification type result of the secondary review and correction is the same as the result of the first review, the identification result is identified as category 4 and the certification result is recorded. When the identification type result of the secondary review and correction is different from the result of the first review, the identification type result of the secondary review and correction is taken as the final result, identified as category 5 and recorded.

6. A land type identification and management system based on unmanned aerial vehicles (UAVs), based on the land type identification and management method based on any one of claims 1 to 5, characterized in that: include, The acquisition module is used to acquire land remote sensing images via drones and input the collected images into the image processing module via a communication network. The image processing module is used to process and extract features from the collected images, transforming them into images with enhanced features, and then sending them to the recognition model in the recognition module for recognition. The recognition module is used to recognize images processed and feature-extracted by the image processing module, calculate similarity, perform preliminary recognition and land category identification, and obtain preliminary category identification results; The review and correction module is used to review and correct the preliminary classification results to check for errors. If the correction result is the same as the preliminary land classification result, the land type identification result is considered correct. If the correction result is different from the preliminary land classification result, the review and correction result shall prevail.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the land type identification and management method based on any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the land type identification and management method based on any one of claims 1 to 5.

Images