Geographic information data processing method

Abstract

The invention discloses a geographic information data processing method, which belongs to the technical field of geographic information processing, and comprises the following steps: constructing a geographic information identification model by adopting a convolutional neural network, and training parameters of the geographic information identification model by adopting an improved intelligent optimization algorithm and preprocessed sample geographic information data. According to the method, the real-time geographic information data is acquired, the trained geographic information identification model is acquired, the trained geographic information identification model is adopted to identify the real-time geographic information data, and the classified geographic information data is acquired, so that data identification and extraction can be quickly realized, the data identification accuracy is improved, and the geographic information data processing efficiency is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of geographic information processing, and in particular relates to a geographic information data processing method. Background Art

[0002] Geographic information data is the representation and all useful knowledge about the properties, characteristics and motion states of geographic entities. It is an interpretation of geographic data that expresses the relationship between geographic feature elements and geographic phenomena. With the continuous growth of geographic information data, how to efficiently and accurately process this data has become a research hotspot in the field of geographic information. At present, some studies have attempted to apply deep learning technology to geographic information data collection. For example, remote sensing images are automatically classified and recognized by training deep neural network models to extract geographic information data. However, deep neural network models usually need to be optimized before they can accurately perform preset tasks, and most algorithms in the existing technology often have poor optimization effects, resulting in poor subsequent geographic information data processing effects. Summary of the invention

[0003] The present invention provides a geographic information data processing method to solve the technical problems existing in the prior art.

[0004] A geographic information data processing method, comprising: Acquiring sample geographic information data, and preprocessing the sample geographic information data to obtain preprocessed sample geographic information data; A convolutional neural network is used to construct a geographic information recognition model, and the parameters of the geographic information recognition model are trained using an improved intelligent optimization algorithm and pre-processed sample geographic information data to obtain a trained geographic information recognition model; Acquire real-time geographic information data, and use the trained geographic information recognition model to recognize the real-time geographic information data to obtain classified geographic information data; Blockchain encryption technology is used to perform distributed storage on the classified geographic information data to complete the processing of the geographic information data.

[0005] Furthermore, the sample geographic information data includes initial terrain elevation data, building shape data, ground cover shape data, building size and location data.

[0006] Further, the sample geographic information data is preprocessed to obtain the preprocessed sample geographic information data, including: Normalizing the sample geographic information data to obtain normalized sample geographic information data; The normalized sample geographic information data is subjected to K-means clustering processing to obtain the preprocessed sample geographic information data.

[0007] Furthermore, a convolutional neural network is used to construct a geographic information recognition model, including: using a U-net neural network to construct a geographic information recognition model.

[0008] Furthermore, the parameters of the geographic information recognition model are trained using the improved intelligent optimization algorithm and the preprocessed sample geographic information data to obtain the trained geographic information recognition model, including: For each category in the sample geographic information data after preprocessing, the category data set is used as training data, and the parameters of the geographic information recognition model are trained using an improved intelligent optimization algorithm to obtain the geographic information recognition model corresponding to each category.

[0009] Furthermore, the category data set is used as training data, and the parameters of the geographic information recognition model are trained using an improved intelligent optimization algorithm, including: Initializing parameters randomly between the upper and lower limits of the parameters of the geographic information recognition model multiple times to obtain multiple model parameter codes; Taking the category data set as a training sample, obtaining the loss function value corresponding to each model parameter encoding; wherein the loss function value is obtained by using the Dice loss function; According to the loss function value corresponding to each model parameter encoding, the optimal parameter encoding and the worst parameter encoding are determined, and the other parameter encodings except the optimal parameter encoding and the worst parameter encoding are set as the target model parameter encoding; The optimal parameter encoding is searched by using the joint distribution optimal region search to obtain the optimal parameter encoding after the search; A chain memory joint search is used to search each target model parameter encoding to obtain the target model parameter encoding after one search; The target model parameter encoding after the first search is searched by using the overall information fusion search improved by the golden sine, and the target model parameter encoding after the second search is obtained; The worst parameter encoding is searched by using a rotating flying search to obtain the worst parameter encoding after the search; Regroup the optimal parameter encoding after the search, the target model parameter encoding after the secondary search, and the worst parameter encoding after the search into a population; Determine whether the maximum number of training times has been reached. If so, re-obtain the optimal parameter coding based on the recombined population, decode the model parameters contained in the optimal parameter coding, and apply it to the geographic information recognition model to complete the training. Otherwise, based on the recombined population, return to the step of determining the optimal parameter coding and the worst parameter coding.

[0010] Furthermore, the optimal parameter encoding is searched by using the joint distribution optimal region search to obtain the optimal parameter encoding after the search, including: According to the loss function value corresponding to each model parameter code, determine the first target parameter code with the second smallest loss function value and the second target parameter code with the third smallest loss function value; According to the first target parameter code and the second target parameter code, a joint distribution search is performed on the optimal parameter code in the optimal region, and the optimal parameter code after the search is obtained as follows:

[0011]

[0012] in, represents the optimal parameter encoding in the t-th training process, Indicates the first target parameter encoding, Indicates the second target parameter encoding, represents the optimal parameter encoding after the search, represents the joint distribution search weight corresponding to the optimal parameter encoding, represents the joint distribution search weight corresponding to the first target parameter encoding, represents the joint distribution search weight corresponding to the second target parameter encoding, =1 / ( E u +0.00001), E u Indicates parameter encoding The corresponding loss function value is, Indicates parameter encoding The corresponding position quality, Indicates parameter encoding The corresponding joint distribution search weight, They are , or .

[0013] Furthermore, chain memory joint search is used to search each target model parameter encoding, and the target model parameter encoding after one search is obtained, including: According to the loss function value corresponding to each target model parameter code, all target model parameter codes are arranged in ascending order of the loss function value to obtain a parameter code sequence; Determine the serial number corresponding to each target model parameter code in the parameter code sequence, and determine the joint search amount of each target model parameter code as:

[0014] in, Indicates the first i The target model parameter encoding, Indicates the first i -1 target model parameter encoding, Indicates the first i The joint search volume corresponding to the target model parameter encoding; when i=1, then Set to optimal parameter encoding; According to the joint search amount of each target model parameter encoding, the target model parameter encoding is memorized and searched, and the target model parameter encoding after one search is obtained as follows:

[0015]

[0016] in, Indicates t -2nd training session i The target model parameter encoding, Indicates the target model parameter encoding after a search , Indicates t -3 training sessions i The target model parameter encoding, represents the adjustment factor, e represents natural training, and T represents the maximum number of training times.

[0017] Furthermore, the overall information fusion search improved by golden sine is used to search the target model parameter encoding after the first search, and the target model parameter encoding after the second search is obtained as follows:

[0018] in, represents the target model parameter encoding after the mth search during the tth training process, Indicates the target model parameter encoding after the second search , represents a random number between [0,2π], represents a random number between [0,π], represents the optimal parameter encoding, represents the center position of the target model parameter encoding after the secondary search, , represents the first update coefficient, represents pi, represents the golden section number, and , , represents the second update coefficient.

[0019] Furthermore, the worst parameter encoding is searched by rotating flying search to obtain the worst parameter encoding after the search, including:

[0020]

[0021] in, represents the worst parameter encoding, Indicates the worst parameter encoding after the search , Represents the code obtained through Levi flight, represents random parameter encoding, Indicates the upper limit With lower limit The random coefficients between Represents a random number between [-1,1], Represents a random number between [-1,1], represents the first random flight coefficient, represents the second random flight coefficient, represents a random number between (0,2], and and It obeys normal distribution, that is , ; represents the intermediate parameter, and , represents the gamma function.

[0022] The present invention provides a geographic information data processing method, which constructs a geographic information recognition model by adopting a convolutional neural network, and trains the parameters of the geographic information recognition model by adopting an improved intelligent optimization algorithm and pre-processed sample geographic information data to obtain a trained geographic information recognition model, and adopts the trained geographic information recognition model to recognize the real-time geographic information data to obtain classified geographic information data, thereby realizing rapid data recognition and extraction, improving data recognition accuracy, and improving geographic information data processing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0024] Figure 1 A flowchart of a geographic information data processing method is provided for an embodiment of the present invention.

[0025] The above drawings have shown clear embodiments of the present invention, which will be described in more detail below. These drawings and text descriptions are not intended to limit the scope of the present invention in any way, but to illustrate the concept of the present invention for those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0026] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0027] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0028] like Figure 1 As shown, an embodiment of the present invention provides a geographic information data processing method, including: S11, acquiring sample geographic information data, and preprocessing the sample geographic information data to obtain preprocessed sample geographic information data; Optionally, the preprocessing can be set to remove outliers, remove duplicates, normalize, and so on, so as to achieve the preprocessing of geographic information data, making it more regular and convenient for subsequent processing.

[0029] S12, using a convolutional neural network to construct a geographic information recognition model, and using an improved intelligent optimization algorithm and pre-processed sample geographic information data to train parameters of the geographic information recognition model to obtain a trained geographic information recognition model; In the embodiment of the present invention, a convolutional neural network is used to construct a geographic information recognition model, including: using a U-net neural network to construct a geographic information recognition model. However, it is worth noting that the U-net neural network is only a preferred implementation of the embodiment of the present invention, and other types of neural network models can also be selected to process geographic information.

[0030] In the prior art, the Adam optimizer is often used to train the neural network model. Although the Adam optimizer has a faster training speed, the parameters often fall into the local optimal position in the solution space during the optimization process using the Adam optimizer, resulting in poor subsequent geographic information recognition effect. Therefore, the embodiment of the present invention provides an improved intelligent optimization algorithm to solve the technical problems existing in the prior art and improve the accuracy of geographic information extraction.

[0031] S13, acquiring real-time geographic information data, and using the trained geographic information recognition model to recognize the real-time geographic information data to obtain classified geographic information data; Since the geographic information recognition model is trained by preprocessing the sample geographic information data, in order to ensure subsequent accurate recognition, the real-time geographic information data should also be in the same format as the sample geographic information data and undergo the same preprocessing.

[0032] S14. Use blockchain encryption technology to perform distributed storage on the classified geographic information data to complete the processing of the geographic information data.

[0033] In the process of accessing geographic information, data security issues often arise. Therefore, the embodiment of the present invention uses blockchain encryption technology to store classified geographic information data, which can effectively ensure the security of geographic information.

[0034] The present invention provides a geographic information data processing method, which constructs a geographic information recognition model by adopting a convolutional neural network, and trains the parameters of the geographic information recognition model by adopting an improved intelligent optimization algorithm and pre-processed sample geographic information data to obtain a trained geographic information recognition model, and adopts the trained geographic information recognition model to recognize the real-time geographic information data to obtain classified geographic information data, thereby realizing rapid data recognition and extraction, improving data recognition accuracy, and improving geographic information data processing efficiency.

[0035] In an embodiment of the present invention, the sample geographic information data includes initial terrain elevation data, building shape data, ground cover shape data, building size and location data.

[0036] It is worth noting that, in addition to the above-mentioned geographic information data, other geographic information data can also be collected for processing, which will not be elaborated here.

[0037] In an embodiment of the present invention, preprocessing the sample geographic information data to obtain the preprocessed sample geographic information data includes: Normalizing the sample geographic information data to obtain normalized sample geographic information data; The normalized sample geographic information data is subjected to K-means clustering processing to obtain the preprocessed sample geographic information data.

[0038] In an embodiment of the present invention, the parameters of the geographic information recognition model are trained using an improved intelligent optimization algorithm and pre-processed sample geographic information data to obtain a trained geographic information recognition model, including: For each category in the sample geographic information data after preprocessing, the category data set is used as training data, and the parameters of the geographic information recognition model are trained using an improved intelligent optimization algorithm to obtain the geographic information recognition model corresponding to each category.

[0039] In the embodiment of the present invention, the category data set is used as training data, and the parameters of the geographic information recognition model are trained using an improved intelligent optimization algorithm, including: Initializing parameters randomly between the upper and lower limits of the parameters of the geographic information recognition model multiple times to obtain multiple model parameter codes; Taking the category data set as a training sample, obtaining the loss function value corresponding to each model parameter encoding; wherein the loss function value is obtained by using the Dice loss function; For example, one can first use existing more complex technologies or equipment to obtain real classified geographic information data, and then obtain the loss function value based on the extracted classified geographic information data, so as to determine whether the neural network model is trained.

[0040] According to the loss function value corresponding to each model parameter encoding, the optimal parameter encoding and the worst parameter encoding are determined, and the other parameter encodings except the optimal parameter encoding and the worst parameter encoding are set as the target model parameter encoding; The optimal parameter encoding is searched by using the joint distribution optimal region search to obtain the optimal parameter encoding after the search; A chain memory joint search is used to search each target model parameter encoding to obtain the target model parameter encoding after one search; The target model parameter encoding after the first search is searched by using the overall information fusion search improved by the golden sine, and the target model parameter encoding after the second search is obtained; The worst parameter encoding is searched by using a rotating flying search to obtain the worst parameter encoding after the search; Regroup the optimal parameter encoding after the search, the target model parameter encoding after the secondary search, and the worst parameter encoding after the search into a population; Determine whether the maximum number of training times has been reached. If so, re-obtain the optimal parameter coding based on the recombined population, decode the model parameters contained in the optimal parameter coding, and apply it to the geographic information recognition model to complete the training. Otherwise, based on the recombined population, return to the step of determining the optimal parameter coding and the worst parameter coding.

[0041] In the embodiment of the present invention, the optimal parameter coding is searched by using the joint distribution optimal region search to obtain the optimal parameter coding after the search, including: According to the loss function value corresponding to each model parameter code, determine the first target parameter code with the second smallest loss function value and the second target parameter code with the third smallest loss function value; According to the first target parameter code and the second target parameter code, a joint distribution search is performed on the optimal parameter code in the optimal region, and the optimal parameter code after the search is obtained as follows:

[0042]

[0043] in, represents the optimal parameter encoding in the t-th training process, Indicates the first target parameter encoding, Indicates the second target parameter encoding, represents the optimal parameter encoding after the search, represents the joint distribution search weight corresponding to the optimal parameter encoding, represents the joint distribution search weight corresponding to the first target parameter encoding, represents the joint distribution search weight corresponding to the second target parameter encoding, =1 / ( E u +0.00001), E u Indicates parameter encoding The corresponding loss function value is, Indicates parameter encoding The corresponding position quality, Indicates parameter encoding The corresponding joint distribution search weight, They are , or .

[0044] The joint distribution optimal area search not only improves the information exchange between codes, but also strengthens the influence of other codes on the global optimal position, speeds up the information exchange in the population, searches for the optimal position to the greatest extent, and effectively reduces the possibility of premature algorithm in the late iteration.

[0045] In the embodiment of the present invention, chain memory joint search is used to search each target model parameter encoding to obtain the target model parameter encoding after one search, including: According to the loss function value corresponding to each target model parameter code, all target model parameter codes are arranged in ascending order of the loss function value to obtain a parameter code sequence; Determine the serial number corresponding to each target model parameter code in the parameter code sequence, and determine the joint search amount of each target model parameter code as:

[0046] in, Indicates the first i The target model parameter encoding, Indicates the first i -1 target model parameter encoding, Indicates the first i The joint search volume corresponding to the target model parameter encoding; when i=1, then Set to optimal parameter encoding; According to the joint search amount of each target model parameter encoding, the target model parameter encoding is memorized and searched, and the target model parameter encoding after one search is obtained as follows:

[0047]

[0048] in, Indicates t -2nd training session i The target model parameter encoding, Indicates the target model parameter encoding after a search , Indicates t -3 training sessions i The target model parameter encoding, represents the adjustment factor, e represents natural training, and T represents the maximum number of training times.

[0049] This chain memory joint search can improve the algorithm's exploration ability in the early stage of training by virtue of its memory characteristics. As the number of iterations increases, the adjustment coefficient continues to decrease, which is beneficial to the later algorithm development. In addition, the process of joint searching of all codes can enhance population diversity and improve the convergence speed of the algorithm.

[0050] In the embodiment of the present invention, the target model parameter encoding after the first search is searched by using the overall information fusion search improved by the golden sine, and the target model parameter encoding after the second search is obtained as follows:

[0051] in, represents the target model parameter encoding after the mth search during the tth training process, Indicates the target model parameter encoding after the second search , represents a random number between [0,2π], represents a random number between [0,π], represents the optimal parameter encoding, represents the center position of the target model parameter encoding after the secondary search, , represents the first update coefficient, represents pi, represents the golden section number, and , , represents the second update coefficient.

[0052] The overall information fusion search improved by the golden sine can speed up the optimization of the algorithm while maintaining the diversity of the population. As the algorithm proceeds, all codes begin to gather in the solution space, and the convergence accuracy of the algorithm begins to increase, thereby balancing the global search capability and convergence accuracy of the algorithm.

[0053] In an embodiment of the present invention, the worst parameter coding is searched by rotating flying search to obtain the worst parameter coding after the search, including:

[0054]

[0055] in, represents the worst parameter encoding, Indicates the worst parameter encoding after the search , Represents the code obtained through Levi flight, represents random parameter encoding, Indicates the upper limit With lower limit The random coefficients between Represents a random number between [-1,1], Represents a random number between [-1,1], represents the first random flight coefficient, represents the second random flight coefficient, represents a random number between (0,2], and and It obeys normal distribution, that is , ; represents the intermediate parameter, and , represents the gamma function.

[0056] The rotating flight search searches at a certain rotation angle, which avoids the generation of repeated codes to the greatest extent, and can effectively improve the algorithm's ability to develop unfamiliar areas. Long jumps occasionally occur during the random walk process to expand the group search range, which can not only enhance the algorithm's local development capabilities, but also effectively assist the algorithm to jump out of the local optimum and achieve global search, thereby improving the algorithm's search performance.

[0057] Optionally, out-of-bounds processing may be performed and a greedy algorithm may be used to control one or more search processes, thereby improving the effectiveness and performance of the algorithm.

[0058] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0059] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0060] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0061] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0062] A person of ordinary skill in the art can understand that all or part of the steps in realizing the above-mentioned facts and methods can be completed by instructing the relevant hardware through a program, and the program involved or the program described can be stored in a computer-readable storage medium. When the program is executed, it includes the following steps: At this time, the corresponding method steps are derived, and the storage medium can be ROM / RAM, a disk, an optical disk, etc.

[0063] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A geographic information data processing method, characterized in that: include: Acquiring sample geographic information data, and preprocessing the sample geographic information data to obtain preprocessed sample geographic information data; A convolutional neural network is used to construct a geographic information recognition model, and the parameters of the geographic information recognition model are trained using an improved intelligent optimization algorithm and pre-processed sample geographic information data to obtain a trained geographic information recognition model; Acquire real-time geographic information data, and use the trained geographic information recognition model to recognize the real-time geographic information data to obtain classified geographic information data; Blockchain encryption technology is used to perform distributed storage on the classified geographic information data to complete the processing of the geographic information data.

2. The geographic information data processing method according to claim 1, characterized in that: The sample geographic information data includes initial terrain elevation data, building shape data, ground cover shape data, building size and location data.

3. The geographic information data processing method according to claim 1, characterized in that: Preprocessing the sample geographic information data to obtain the preprocessed sample geographic information data includes: Normalizing the sample geographic information data to obtain normalized sample geographic information data; The normalized sample geographic information data is subjected to K-means clustering processing to obtain the preprocessed sample geographic information data.

4. The geographic information data processing method according to claim 3, characterized in that: A convolutional neural network is used to construct a geographic information recognition model, including: using a U-net neural network to construct a geographic information recognition model.

5. The geographic information data processing method according to claim 4, characterized in that: The parameters of the geographic information recognition model are trained using the improved intelligent optimization algorithm and the pre-processed sample geographic information data to obtain the trained geographic information recognition model, including: For each category in the sample geographic information data after preprocessing, the category data set is used as training data, and the parameters of the geographic information recognition model are trained using an improved intelligent optimization algorithm to obtain the geographic information recognition model corresponding to each category.

6. The geographic information data processing method according to claim 5, characterized in that: Using the category data set as training data, the improved intelligent optimization algorithm is used to train the parameters of the geographic information recognition model, including: Initializing parameters randomly between the upper and lower limits of the parameters of the geographic information recognition model multiple times to obtain multiple model parameter codes; Taking the category data set as a training sample, obtaining the loss function value corresponding to each model parameter encoding; wherein the loss function value is obtained by using the Dice loss function; According to the loss function value corresponding to each model parameter encoding, the optimal parameter encoding and the worst parameter encoding are determined, and the other parameter encodings except the optimal parameter encoding and the worst parameter encoding are set as the target model parameter encoding; The optimal parameter encoding is searched by using the joint distribution optimal region search to obtain the optimal parameter encoding after the search; A chain memory joint search is used to search each target model parameter encoding to obtain the target model parameter encoding after one search; The target model parameter encoding after the first search is searched by using the overall information fusion search improved by the golden sine, and the target model parameter encoding after the second search is obtained; The worst parameter encoding is searched by using a rotating flying search to obtain the worst parameter encoding after the search; Regroup the optimal parameter encoding after the search, the target model parameter encoding after the secondary search, and the worst parameter encoding after the search into a population; Determine whether the maximum number of training times has been reached. If so, re-obtain the optimal parameter coding based on the recombined population, decode the model parameters contained in the optimal parameter coding, and apply it to the geographic information recognition model to complete the training. Otherwise, based on the recombined population, return to the step of determining the optimal parameter coding and the worst parameter coding.

7. The geographic information data processing method according to claim 6, characterized in that: The optimal parameter encoding is searched by using the joint distribution optimal region search to obtain the optimal parameter encoding after the search, including: According to the loss function value corresponding to each model parameter code, determine the first target parameter code with the second smallest loss function value and the second target parameter code with the third smallest loss function value; According to the first target parameter code and the second target parameter code, a joint distribution search is performed on the optimal parameter code in the optimal region, and the optimal parameter code after the search is obtained as follows: in, represents the optimal parameter encoding in the t-th training process, Indicates the first target parameter encoding, Indicates the second target parameter encoding, represents the optimal parameter encoding after the search, represents the joint distribution search weight corresponding to the optimal parameter encoding, represents the joint distribution search weight corresponding to the first target parameter encoding, represents the joint distribution search weight corresponding to the second target parameter encoding, =1 / ( E u +0.00001), E u Indicates parameter encoding The corresponding loss function value is, Indicates parameter encoding The corresponding position quality, Indicates parameter encoding The corresponding joint distribution search weight, They are , or .

8. The geographic information data processing method according to claim 7, characterized in that: Chain memory joint search is used to search each target model parameter encoding to obtain the target model parameter encoding after one search, including: According to the loss function value corresponding to each target model parameter code, all target model parameter codes are arranged in ascending order of the loss function value to obtain a parameter code sequence; Determine the serial number corresponding to each target model parameter code in the parameter code sequence, and determine the joint search amount of each target model parameter code as: in, Indicates the first i The target model parameter encoding, Indicates the first i -1 target model parameter encoding, Indicates the first i The joint search volume corresponding to the target model parameter encoding; when i=1, then Set to optimal parameter encoding; According to the joint search amount of each target model parameter encoding, the target model parameter encoding is memorized and searched, and the target model parameter encoding after one search is obtained as follows: in, Indicates t -2nd training session i The target model parameter encoding, Indicates the target model parameter encoding after a search , Indicates t -3 training sessions i The target model parameter encoding, represents the adjustment factor, e represents natural training, and T represents the maximum number of training times.

9. The geographic information data processing method according to claim 8, characterized in that: The overall information fusion search improved by golden sine is used to search the target model parameter encoding after the first search, and the target model parameter encoding after the second search is obtained as follows: in, represents the target model parameter encoding after the mth search during the tth training process, Indicates the target model parameter encoding after the second search , represents a random number between [0,2π], represents a random number between [0,π], represents the optimal parameter encoding, represents the center position of the target model parameter encoding after the secondary search, , represents the first update coefficient, represents pi, represents the golden section number, and , , represents the second update coefficient.

10. The geographic information data processing method according to claim 9, characterized in that: The worst parameter encoding is searched by rotating flying search, and the worst parameter encoding after the search is obtained, including: in, represents the worst parameter encoding, Indicates the worst parameter encoding after the search , represents the code obtained through Levi flight, represents random parameter encoding, Indicates the upper limit With lower limit The random coefficients between Represents a random number between [-1,1], Represents a random number between [-1,1], represents the first random flight coefficient, represents the second random flight coefficient, represents a random number between (0,2], and and It obeys normal distribution, that is , ; represents the intermediate parameter, and , represents the gamma function.

Images