Artificial intelligence-based text booth space automatic design system and method
Through the automatic design system of Wensheng booth space based on artificial intelligence, using neural network model and bat algorithm, the problems of cumbersome and error in the design process of the existing booth space are solved, efficient and accurate design output is achieved, and manpower investment is reduced.
Patent Information
- Application Number
- CN202411040894.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-31
- Publication Date
- 2024-11-19
AI Technical Summary
The design process of the existing booth space is cumbersome, requiring a lot of manpower, and it is difficult to avoid errors caused by manual operations.
The Wensheng booth space automatic design system based on artificial intelligence is adopted, and the booth renderings and construction drawings are generated in one-click through neural network models and bat algorithms, and the optimal construction technology is matched to reduce manpower investment and improve design efficiency and accuracy.
The design is intelligent, standardized and modularized, which reduces manpower investment, improves design efficiency and accuracy, and ensures the quality of the output booth renderings and construction drawings.
Smart Images

Figure CN118981820A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of artificial intelligence, and in particular to an artificial intelligence-based automatic exhibition booth space design system and method. Background Art
[0002] With the rapid development of artificial intelligence technology, text graph technology has gradually become a research hotspot. Text graph, that is, the technology of generating corresponding images through natural language description, has shown great application potential in the fields of art creation, advertising design, and auxiliary education. In generative artificial intelligence applications, the key to meeting the diverse needs of users is to quickly customize reference images through text graph models to generate results of reference image subjects in different semantic scenarios.
[0003] Text-generated images usually refer to the generation of corresponding images from the text descriptions entered by users. First, natural language processing is used to parse the input text, analyze and understand the vocabulary, syntactic structure and semantic relationships to extract the semantic information in the text, and then computer vision technology is used to convert the input sample image into a compressed feature map. Noise and denoise are added on the basis of the feature map, and finally decoding is performed to generate a high-quality image.
[0004] Booth designs are all the same. In the past, they all involved manual research on demand, plane planning, space plan planning, 3D modeling, CAD drawing, etc., and finally breaking them down into construction drawings for construction. This series of tedious steps required a lot of manpower investment, and it was difficult to avoid errors caused by manual operation. Summary of the invention
[0005] 1. Technical issues to be resolved
[0006] In order to address the deficiencies in the background technology, the present invention proposes an artificial intelligence-based automatic booth space design system and method. After the AI design is completed, the booth renderings and booth construction drawings are generated with one click, and the optimal construction process is matched according to the booth construction drawings. This can minimize manpower input, and the design is intelligent, standardized, modular, and highly efficient, with one click generation of designs.
[0007] (II) Technical solution
[0008] An automatic design method for Vincent booth space based on artificial intelligence, the method comprising the following steps:
[0009] S1, setting a set of item categories corresponding to an input keyword, collecting samples of all item categories corresponding to the input keyword in the set of item categories corresponding to the input keyword, recording it as a first sample set, performing image collection on all samples in the first sample set, extracting image data, and obtaining a first sample image data matrix;
[0010] S2, performing data division on the first sample image data matrix to obtain a first sample image data training matrix and a first sample image data test matrix, constructing an initial neural network model, and training and testing the initial neural network model using the first sample image data training matrix and the first sample image data test matrix to obtain a final neural network model;
[0011] S3, input the keywords into the final neural network model to design the booth, and output the booth renderings and booth construction drawings;
[0012] S4. Collect construction process data required for booth construction, and establish a construction process data set based on the collected construction process data required for booth construction;
[0013] S5. Matching the optimal construction process to the output booth construction drawing according to the construction process data set.
[0014] By collecting images of samples of all item types corresponding to the input keywords in the item type set corresponding to the collected input keywords, image data is obtained to provide a data basis for the training and testing of the neural network model; the initial neural network model is trained and tested in large quantities through the first sample image training matrix and the first sample image test matrix until the preset conditions are met, thereby ensuring the accuracy of the final output results of the neural network model; a construction process data set is established to provide a data basis for the operation of matching the booth construction drawings with the optimal construction process; the bat algorithm is used to match the output booth construction drawings with the optimal construction process, which has a good optimization effect and high optimization efficiency, while ensuring the reliability and accuracy of the matched optimal construction process.
[0015] Preferably, a set of item categories corresponding to the input keyword is set, samples of all item categories corresponding to the input keyword in the set of item categories corresponding to the input keyword are collected, recorded as a first sample set, and image collection is performed on all samples in the first sample set to extract image data to obtain a first sample image data matrix. The specific steps are as follows:
[0016] S11, set the item type set A corresponding to the input keyword = {a1, a2, ..., a i ,…,a k}, a i represents the item type corresponding to the i-th input keyword, and k represents the number of item types corresponding to the input keyword;
[0017] S12, collect the item category set A corresponding to the input keyword = {a1, a2, ..., a i ,…,a k}, and the first sample set A' is obtained as follows:
[0018]
[0019] Among them, a' ij represents the jth sample in the item category corresponding to the i-th input keyword, and a” represents the number of samples collected for the item category corresponding to each input keyword;
[0020] S13, perform image acquisition on all samples in the first sample set A', extract image data, and obtain a first sample image data matrix B' as follows:
[0021]
[0022] Among them, b' ij It represents the data set obtained by collecting the jth sample in the item category corresponding to the i-th input keyword.
[0023] By collecting samples of all the item types corresponding to the input keywords in the item type set, performing image acquisition on all samples, extracting image data, and obtaining the first sample image data matrix, by performing image acquisition on a large and comprehensive sample, extracting image data, and obtaining image data of all the items corresponding to the keywords, a data basis is provided for the training and testing of the neural network model.
[0024] Preferably, data division is performed on the first sample image data matrix to obtain a first sample image data training matrix and a first sample image data test matrix, an initial neural network model is constructed, and the initial neural network model is trained and tested by the first sample image data training matrix and the first sample image data test matrix to obtain the final neural network model. The operation steps are as follows:
[0025] S21, set the training ratio value to y1, the test ratio value to y2, and divide the first sample image data matrix according to the set training ratio value to y1 and the test ratio value to y2, to obtain the first sample image data training matrix C' and the first sample image data test matrix D' as follows:
[0026]
[0027] Among them, c' ij represents the data set of the jth sample in the item category corresponding to the i-th input keyword in the first sample image data training matrix C'; d' ij represents the data set of the jth sample in the item category corresponding to the i-th input keyword in the first sample image data test matrix D'; a1" and a2" represent the number of samples in the first sample image data training matrix C' and the first sample image data test matrix D' respectively;
[0028] S22, constructing an initial neural network model, setting the number of input nodes of the initial neural network model to be a, the number of hidden nodes to be b, the number of output nodes to be c, the initial weight to be λ, and the initial threshold to be η;
[0029] S23, setting the training error threshold as n1, inputting the first sample data image training matrix C' into the initial neural network model to train the initial neural network model, and when the training error of the initial neural network model is less than or equal to n1, stopping the training to obtain a trained initial neural network model;
[0030] S24. Set the accuracy threshold to n2, input the first sample image data test matrix D' into the trained initial neural network model for testing, and obtain the test accuracy m after the test. When m≥n2, the trained initial neural network model is used as the final neural network model; otherwise, adjust the initial weight λ and the initial threshold η, and retest until the test accuracy obtained after the test is greater than or equal to the accuracy threshold n2, then the neural network model currently being tested is used as the final neural network model.
[0031] Based on data division of the first sample image data matrix to obtain the first sample image data training matrix and the first sample image data test matrix, the initial neural network model is trained and tested to obtain the final neural network model. The initial neural network model is trained and tested extensively until the preset conditions are met, thereby ensuring the accuracy and reliability of the output results of the final neural network model.
[0032] Preferably, the operation steps of inputting keywords into the final neural network model to design the booth and outputting the booth renderings and booth construction drawings are as follows:
[0033] S31. Randomly select any keywords and input them into the final neural network model to design the booth, and output the booth renderings and booth construction drawings.
[0034] Make sure that any keywords entered can match the best booth renderings and booth construction drawings.
[0035] Preferably, the operation steps of collecting the construction process data required for booth construction and establishing a construction process data set based on the collected construction process data required for booth construction are as follows:
[0036] S41. Collect the construction process data required for booth construction;
[0037] S42: Establish a construction process data set P = (p1, p2, ..., p i ,…,p l), i=1,2,…,l,p i represents the construction process data required for the construction of the i-th booth, and l represents the total number of construction process data required for the construction of the booth.
[0038] By establishing a construction process data set, a data basis is provided for matching the booth construction drawings with the optimal construction process, ensuring the accuracy of the optimization results.
[0039] Preferably, the operation steps of matching the optimal construction process for the output booth construction drawing according to the construction process data set are as follows:
[0040] S51, construct a bat population, set the size of the bat population to l, and express the bat population as g'={g'1,g'2,…,g' i ,…,g' l}, g' i represents the i-th bat in the bat population; setting the spatial dimension of the target search to d, the bat population g'={g'1,g'2,…,g' i ,…,g' l}, the position of each i bat in the tth iteration is Speed The pulse search frequency is f i , the pulse emission rate is Loudness is Set the maximum number of iterations to t max ;
[0041] Among them, each individual bat corresponds to a construction process data, and l individual bats correspond to each construction process data in the construction process data set P;
[0042] Initialize the initial position set of the bat population to X = {x1, x2, ..., x i ,…,x l}、Initial velocity set V={v1,v2,…,v i ,…,v l}、The initial pulse transmission frequency set is and the initial impulse loudness set is Among them, x i 、v i , as well as They represent the initial position, initial velocity, initial pulse emission frequency, and initial pulse loudness of the i-th bat respectively;
[0043] S52, respectively calculate the fitness value of each individual bat in the bat population, and record the bat individual with the largest fitness as the global optimal solution. The fitness function is as follows:
[0044]
[0045] Among them, M i is the fitness value of the i-th bat individual, m i represents the construction cost of the i-th bat individual to complete the booth construction drawing;
[0046] S53, updating the pulse frequency, speed and position of each bat according to the update formula; the update formula is as follows:
[0047]
[0048] Among them, f i represents the pulse frequency of the i-th bat, f min and f max They represent the maximum and minimum values of the pulse frequency, β represents a random number uniformly distributed between [0,1], and x * represents the current global optimal solution;
[0049] S54, generate a random number rand1, if Then the optimal solution perturbation is performed according to the search formula; otherwise, enter S56; the search formula is as follows:
[0050] x=x * +εA t
[0051] new
[0052] Among them, x new represents the new solution obtained by random perturbation, A t represents the average loudness of the t-th generation of bats, and ε represents a random number uniformly distributed between [-1,1];
[0053] S55, generate a random number rand2, if And x new The fitness value is greater than x * , then accept x new and update and The update formula is as follows:
[0054]
[0055] Among them, α represents a random number uniformly distributed between [0,1], and γ represents a random number uniformly distributed between [0,+∞];
[0056] S56, calculating the fitness of the updated bat individuals, and taking the bat individual with the largest fitness as the current global optimal solution;
[0057] S57, determine whether the maximum number of iterations t is reached maxIf it is not reached, return to S53; if it is reached, stop the iteration operation and output the optimal construction process of the booth construction drawing.
[0058] The bat algorithm is used to match the optimal construction process for the output booth construction drawing. The bat algorithm simulates the echolocation behavior of bats. During the flight of bats, they will make loud sounds, and then judge whether there are obstacles around based on the received echo signals. The echo can avoid collisions with neighbors and can also quickly lock the direction and distance of prey. It has a good optimization effect and high optimization efficiency, while ensuring the reliability and accuracy of the matched optimal construction process.
[0059] The present invention also discloses an artificial intelligence-based automatic booth space design system, including a keyword corresponding object image data acquisition module, a neural network model construction module, a booth effect diagram and construction drawing output module, a construction process data set construction module, and a construction process optimization module;
[0060] The keyword corresponding item image data acquisition module acquires samples of all item types corresponding to the input keyword in the item type set corresponding to the input keyword, which is recorded as a first sample set, and performs image acquisition on all samples in the first sample set to extract image data to obtain a first sample image data matrix;
[0061] The neural network model construction module performs data partitioning on the first sample image data matrix to obtain a first sample image data training matrix and a first sample image data test matrix, and trains and tests the initial neural network model based on the first sample image data training matrix and the first sample image data test matrix to obtain a final neural network model;
[0062] The booth rendering and construction drawing output module randomly selects any keywords and inputs them into the final neural network model to perform booth design, and outputs the booth rendering and the booth construction drawing;
[0063] The construction process data set building module collects the construction process data required for the booth construction, and builds the construction process data set based on the collected construction process data required for the booth construction;
[0064] The construction process optimization module matches the optimal construction process to the output booth construction drawing through the bat algorithm.
[0065] (III) Beneficial effects
[0066] The present invention provides a system and method for automatic booth space design based on artificial intelligence, which has the following beneficial effects:
[0067] 1. The keyword-corresponding item image data acquisition module acquires samples of all item types corresponding to the input keyword in the item type set corresponding to the input keyword, performs image acquisition on all samples, obtains a first sample image data matrix, and extracts image data from a large number of comprehensive samples to obtain image data, thereby providing a data basis for the training and testing of the neural network model; the neural network model construction module performs data division on the first sample image data matrix to obtain a first sample image data training matrix and a first sample image data test matrix, trains and tests the initial neural network model to obtain the final neural network model, performs a large amount of training and testing on the initial neural network model until the preset conditions are met, thereby ensuring the accuracy and reliability of the output result of the final neural network model;
[0068] 2. The booth rendering and construction drawing output module randomly selects any number of keywords and inputs them into the final neural network model to perform booth design, and outputs booth renderings and booth construction drawings, ensuring that any number of input keywords can match the optimal booth renderings and booth construction drawings;
[0069] 3. The construction process data set building module collects the construction process data required for the booth construction, and establishes a construction process data set based on the collected construction process data required for the booth construction, thereby providing a data basis for matching the optimal construction process for the booth construction drawing and ensuring the accuracy of the optimization result; the construction process optimization module matches the optimal construction process for the output booth construction drawing through the bat algorithm, which has a good optimization effect and high optimization efficiency, while ensuring the reliability and accuracy of the matched optimal construction process. BRIEF DESCRIPTION OF THE DRAWINGS
[0070] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying creative work.
[0071] Figure 1 A flowchart of an artificial intelligence-based automatic booth space design method provided by the present invention;
[0072] Figure 2 A schematic diagram of a module of an artificial intelligence-based automatic booth space design system provided by the present invention. DETAILED DESCRIPTION
[0073] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0074] In the description of the present invention, it is necessary to understand that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention.
[0075] The embodiments of the artificial intelligence-based automatic booth space design system and method are as follows:
[0076] See also Figure 1 — Figure 2 , an automatic design method of Vincent booth space based on artificial intelligence, the method comprises the following steps:
[0077] S1, setting a set of item categories corresponding to an input keyword, collecting samples of all item categories corresponding to the input keyword in the set of item categories corresponding to the input keyword, recording it as a first sample set, performing image collection on all samples in the first sample set, extracting image data, and obtaining a first sample image data matrix;
[0078] S11, set the item type set A corresponding to the input keyword = {a1, a2, ..., a i ,…,a k}, a i represents the item type corresponding to the i-th input keyword, and k represents the number of item types corresponding to the input keyword;
[0079] S12, collect the item category set A corresponding to the input keyword = {a1, a2, ..., a i ,…,a k}, and the first sample set A' is obtained as follows:
[0080]
[0081] Among them, a' ij represents the jth sample in the item category corresponding to the i-th input keyword, and a” represents the number of samples collected for the item category corresponding to each input keyword;
[0082] S13, perform image acquisition on all samples in the first sample set A', extract image data, and obtain a first sample image data matrix B' as follows:
[0083]
[0084] Among them, b' ij It represents the data set obtained by collecting the jth sample in the item category corresponding to the i-th input keyword.
[0085] S2, performing data division on the first sample image data matrix to obtain a first sample image data training matrix and a first sample image data test matrix, constructing an initial neural network model, and training and testing the initial neural network model using the first sample image data training matrix and the first sample image data test matrix to obtain a final neural network model;
[0086] S21, set the training ratio value to y1, the test ratio value to y2, and divide the first sample image data matrix according to the set training ratio value to y1 and the test ratio value to y2, to obtain the first sample image data training matrix C' and the first sample image data test matrix D' as follows:
[0087]
[0088]
[0089] Among them, c' ij represents the data set of the jth sample in the item category corresponding to the i-th input keyword in the first sample image data training matrix C'; d' ij represents the data set of the jth sample in the item category corresponding to the i-th input keyword in the first sample image data test matrix D'; a1" and a2" represent the number of samples in the first sample image data training matrix C' and the first sample image data test matrix D' respectively;
[0090] S22, constructing an initial neural network model, setting the number of input nodes of the initial neural network model to be a, the number of hidden nodes to be b, the number of output nodes to be c, the initial weight to be λ, and the initial threshold to be η;
[0091] S23, setting the training error threshold as n1, inputting the first sample data image training matrix C' into the initial neural network model to train the initial neural network model, and when the training error of the initial neural network model is less than or equal to n1, stopping the training to obtain a trained initial neural network model;
[0092] S24. Set the accuracy threshold to n2, input the first sample image data test matrix D' into the trained initial neural network model for testing, and obtain the test accuracy m after the test. When m≥n2, the trained initial neural network model is used as the final neural network model; otherwise, adjust the initial weight λ and the initial threshold η, and retest until the test accuracy obtained after the test is greater than or equal to the accuracy threshold n2, then the neural network model currently being tested is used as the final neural network model.
[0093] S3, input the keywords into the final neural network model to design the booth, and output the booth renderings and booth construction drawings;
[0094] S31. Randomly select any keywords and input them into the final neural network model to design the booth, and output the booth renderings and booth construction drawings.
[0095] S4. Collect construction process data required for booth construction, and establish a construction process data set based on the collected construction process data required for booth construction;
[0096] S41. Collect the construction process data required for booth construction;
[0097] S42: Establish a construction process data set P = (p1, p2, ..., p i ,…,p l ), i=1,2,…,l,p i represents the construction process data required for the construction of the i-th booth, and l represents the total number of construction process data required for the construction of the booth.
[0098] S5. Matching the optimal construction process for the output booth construction drawing according to the construction process data set;
[0099] S51, construct a bat population, set the size of the bat population to l, and express the bat population as g'={g'1,g'2,…,g' i ,…,g' l}, g' i represents the i-th bat in the bat population; setting the spatial dimension of the target search to d, the bat population g'={g'1,g'2,…,g' i ,…,g' l}, the position of each i bat in the tth iteration is Speed The pulse search frequency is f i , the pulse emission rate is Loudness is Set the maximum number of iterations to t max ;
[0100] Among them, each individual bat corresponds to a construction process data, and l individual bats correspond to each construction process data in the construction process data set P;
[0101] Initialize the initial position set of the bat population to X = {x1, x2, ..., x i ,…,x l}、Initial velocity set V={v1,v2,…,v i ,…,v l}、The initial pulse transmission frequency set is and the initial impulse loudness set is Among them, x i 、v i , as well as They represent the initial position, initial velocity, initial pulse emission frequency, and initial pulse loudness of the i-th bat respectively;
[0102] S52, respectively calculate the fitness value of each individual bat in the bat population, and record the bat individual with the largest fitness as the global optimal solution. The fitness function is as follows:
[0103]
[0104] Among them, M i is the fitness value of the i-th bat individual, m i represents the construction cost of the i-th bat individual to complete the booth construction drawing;
[0105] S53, updating the pulse frequency, speed and position of each bat according to the update formula; the update formula is as follows:
[0106]
[0107] Among them, f i represents the pulse frequency of the i-th bat, f min and f max They represent the maximum and minimum values of the pulse frequency, β represents a random number uniformly distributed between [0,1], and x * represents the current global optimal solution;
[0108] S54, generate a random number rand1, if Then the optimal solution perturbation is performed according to the search formula; otherwise, enter S56; the search formula is as follows:
[0109] x=x * +εA t
[0110] new
[0111] Among them, x new represents the new solution obtained by random perturbation, A t represents the average loudness of the t-th generation of bats, and ε represents a random number uniformly distributed between [-1,1];
[0112] S55, generate a random number rand2, if And x new The fitness value is greater than x * , then accept x new and update and The update formula is as follows:
[0113]
[0114] Among them, α represents a random number uniformly distributed between [0,1], and γ represents a random number uniformly distributed between [0,+∞];
[0115] S56, calculating the fitness of the updated bat individuals, and taking the bat individual with the largest fitness as the current global optimal solution;
[0116] S57, determine whether the maximum number of iterations t is reached max If it is not reached, return to S53; if it is reached, stop the iteration operation and output the optimal construction process of the booth construction drawing.
[0117] The present invention also discloses an artificial intelligence-based automatic booth space design system, including a keyword corresponding object image data acquisition module, a neural network model construction module, a booth effect diagram and construction drawing output module, a construction process data set construction module, and a construction process optimization module;
[0118] The keyword corresponding item image data acquisition module acquires samples of all item types corresponding to the input keyword in the item type set corresponding to the input keyword, which is recorded as a first sample set, and performs image acquisition on all samples in the first sample set to extract image data to obtain a first sample image data matrix;
[0119] The neural network model construction module performs data partitioning on the first sample image data matrix to obtain a first sample image data training matrix and a first sample image data test matrix, and trains and tests the initial neural network model based on the first sample image data training matrix and the first sample image data test matrix to obtain a final neural network model;
[0120] The booth rendering and construction drawing output module randomly selects any keywords and inputs them into the final neural network model to perform booth design, and outputs the booth rendering and the booth construction drawing;
[0121] The construction process data set building module collects the construction process data required for the booth construction, and builds the construction process data set based on the collected construction process data required for the booth construction;
[0122] The construction process optimization module matches the optimal construction process to the output booth construction drawing through the bat algorithm.
[0123] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. An artificial intelligence-based automatic booth space design method, characterized in that: The steps include: S1, setting a set of item categories corresponding to an input keyword, collecting samples of all item categories corresponding to the input keyword in the set of item categories corresponding to the input keyword, recording it as a first sample set, performing image collection on all samples in the first sample set, extracting image data, and obtaining a first sample image data matrix; S2, performing data division on the first sample image data matrix to obtain a first sample image data training matrix and a first sample image data test matrix, constructing an initial neural network model, and training and testing the initial neural network model using the first sample image data training matrix and the first sample image data test matrix to obtain a final neural network model; S3, input the keywords into the final neural network model to design the booth, and output the booth renderings and booth construction drawings; S4. Collect construction process data required for booth construction, and establish a construction process data set based on the collected construction process data required for booth construction; S5. Matching the optimal construction process to the output booth construction drawing according to the construction process data set.
2. The method for automatic booth space design based on artificial intelligence according to claim 1 is characterized in that: The S1 comprises the following steps: S11, set the item type set A corresponding to the input keyword = {a1, a2, ..., a i ,…,a k }, a i represents the item type corresponding to the i-th input keyword, and k represents the number of item types corresponding to the input keyword; S12, collect the item category set A corresponding to the input keyword = {a1, a2, ..., a i ,…,a k }, and the first sample set A' is obtained as follows: Among them, a' ij represents the jth sample in the item category corresponding to the i-th input keyword, and a” represents the number of samples collected for the item category corresponding to each input keyword; S13, perform image acquisition on all samples in the first sample set A', extract image data, and obtain a first sample image data matrix B' as follows: Among them, b' ij It represents the data set obtained by collecting the jth sample in the item category corresponding to the i-th input keyword.
3. The method for automatic booth space design based on artificial intelligence according to claim 2 is characterized in that: The S2 comprises the following steps: S21, performing data division on the first sample image data matrix to obtain a first sample image data training matrix and a first sample image data test matrix; S22, constructing an initial neural network model, setting the number of input nodes of the initial neural network model to be a, the number of hidden nodes to be b, the number of output nodes to be c, the initial weight to be λ, and the initial threshold to be η; S23, setting the training error threshold as n1, inputting the first sample data image training matrix C' into the initial neural network model to train the initial neural network model, and when the training error of the initial neural network model is less than or equal to n1, stopping the training to obtain a trained initial neural network model; S24. Set the accuracy threshold to n2, input the first sample image data test matrix D' into the trained initial neural network model for testing, and obtain the test accuracy m after the test. When m≥n2, the trained initial neural network model is used as the final neural network model; otherwise, adjust the initial weight λ and the initial threshold η, and retest until the test accuracy obtained after the test is greater than or equal to the accuracy threshold n2, then the neural network model currently being tested is used as the final neural network model.
4. The method for automatic booth space design based on artificial intelligence according to claim 3 is characterized in that: The data division of the first sample image data matrix comprises the following steps: The training ratio value is set to y1, and the test ratio value is set to y2. The first sample image data matrix is divided according to the set training ratio value y1 and the test ratio value y2 to obtain the first sample image data training matrix C' and the first sample image data test matrix D' as follows: Among them, c' ij represents the data set of the jth sample in the item category corresponding to the i-th input keyword in the first sample image data training matrix C'; d' ij represents the data set of the jth sample in the item category corresponding to the i-th input keyword in the first sample image data test matrix D'; a1" and a2" represent the number of samples in the first sample image data training matrix C' and the first sample image data test matrix D' respectively.
5. The method for automatic booth space design based on artificial intelligence according to claim 4 is characterized in that: The S3 comprises the following steps: S31. Randomly select any keywords and input them into the final neural network model to design the booth, and output the booth renderings and booth construction drawings.
6. The method for automatic booth space design based on artificial intelligence according to claim 5 is characterized in that: The S4 comprises the following steps: S41. Collect the construction process data required for booth construction; S42: Establish a construction process data set P = (p1, p2, ..., p i ,…,p l ), i=1,2,…,l,p i represents the construction process data required for the construction of the i-th booth, and l represents the total number of construction process data required for the construction of the booth.
7. The method for automatic booth space design based on artificial intelligence according to claim 6 is characterized in that: The S5 comprises the following steps: S51, construct a bat population, set the size of the bat population to l, and express the bat population as g'={g'1,g'2,…,g' i ,…,g' l }, g' i represents the i-th bat in the bat population; setting the spatial dimension of the target search to d, the bat population g'={g'1,g'2,…,g' i ,…,g' l }, the position of each i bat in the tth iteration is Speed The pulse search frequency is f i , the pulse emission rate is r i t , loudness is Set the maximum number of iterations to t max ; Among them, each individual bat corresponds to a construction process data, and l individual bats correspond to each construction process data in the construction process data set P; Initialize the initial position set of the bat population to X = {x1, x2, ..., x i ,…,x l }、Initial velocity set V={v1,v2,…,v i ,…,v l }, the initial pulse transmission frequency set is R = {r1 0 ,r2 0 ,…,r i 0 ,…,r l 0 } and the initial impulse loudness set is Among them, x i 、v i 、r i 0 and A i 0 They represent the initial position, initial velocity, initial pulse emission frequency, and initial pulse loudness of the i-th bat respectively; S52, respectively calculate the fitness value of each individual bat in the bat population, and record the bat individual with the largest fitness as the global optimal solution. The fitness function is as follows: Among them, M i is the fitness value of the i-th bat individual, m i represents the construction cost of the i-th bat individual to complete the booth construction drawing; S53, updating the pulse frequency, speed and position of each bat according to the update formula; the update formula is as follows: Among them, f i represents the pulse frequency of the i-th bat, f min and f max They represent the maximum and minimum values of the pulse frequency, β represents a random number uniformly distributed between [0,1], and x * represents the current global optimal solution; S54, generate a random number rand1, if rand>r i t , then the optimal solution perturbation is performed according to the search formula; otherwise, enter S56; the search formula is as follows: x new =x * +εA t Among them, x new represents the new solution obtained by random perturbation, A t represents the average loudness of the t-th generation of bats, and ε represents a random number uniformly distributed between [-1,1]; S55, generate a random number rand2, if And x new The fitness value is greater than x * , then accept x new And update r i t and The update formula is as follows: Among them, α represents a random number uniformly distributed between [0,1], and γ represents a random number uniformly distributed between [0,+∞]; S56, calculating the fitness of the updated bat individuals, and taking the bat individual with the largest fitness as the current global optimal solution; S57, determine whether the maximum number of iterations t is reached max If it is not reached, return to S53; if it is reached, stop the iteration operation and output the optimal construction process of the booth construction drawing.
8. A system for realizing the automatic design of Vincent booth space based on artificial intelligence as described in any one of claims 1 to 7.
