Garden plant geographical information query system and method based on information recognition

Abstract

The present application provides a garden plant geographical information query system and method based on information recognition, which relates to the field of smart garden applications. First, it extracts features from the image data of garden plants to obtain the morphological attribute features of garden plants. At the same time, it extracts features from the name and distribution location information of garden plants to obtain the distribution spatial geographical features of garden plants. Then, it fuses the morphological attribute features of garden plants and the distribution spatial geographical features of garden plants to obtain the garden plant features. Next, it optimizes the fused features. Finally, it uses a generator to generate a plant landscape spatial geographical query file from the optimized garden plant features. In this way, it can meet diversified query requirements and can also combine the generated plant landscape spatial geographical query file with a geographical information system to achieve visual query.

Description

Technical Field

[0001] The present application relates to the field of intelligent garden applications, and more specifically, to a garden plant geographical information query system and method based on information recognition. Background Art

[0002] Traditional garden plant geographical information query methods often only support queries based on the plant name text or simple classification information. When the user is not very familiar with the plant name and the attribute expression is incomplete, the query accuracy rate is relatively low, and traditional query methods often only provide query results in text or table form, which is not intuitive enough.

[0003] Therefore, an optimized garden plant geographical information query solution based on information recognition is needed. Summary of the Invention

[0004] The present application aims at the deficiencies in the prior art and provides a garden plant geographical information query system and method based on information recognition. First, it extracts the morphological attribute features of garden plants from the image data of garden plants, and at the same time extracts the distribution space geographical features of garden plants from the name and distribution location information of garden plants, and then fuses the morphological attribute features of garden plants and the distribution space geographical features of garden plants to obtain garden plant features, and then optimizes the fused features, and finally generates a plant landscape space geographical query file through a generator with the optimized garden plant features. In this way, it can meet diversified query needs, and the generated plant landscape space geographical query file can also be combined with a geographical information system to achieve visual query.

[0005] A garden plant geographical information query system based on information recognition, comprising:

[0006] A garden plant related data acquisition module for obtaining the image data of all plants in the garden and the name and distribution location information of all plants;

[0007] A garden plant related data feature encoding module for performing feature encoding on the image data of all plants and the name and distribution location information of all plants to obtain a garden plant feature matrix;

[0008] A query file generation module for generating a plant landscape space geographical query file based on the garden plant feature matrix.

[0009] A garden plant geographical information query method based on information recognition, comprising:

[0010] Obtaining the image data of all plants in the garden and the name and distribution location information of all plants;

[0011] Perform feature encoding on the image data of all the plants, as well as the names and distribution location information of all the plants, to obtain a landscape plant feature matrix.

[0012] Generate a plant landscape spatial geographic query file based on the landscape plant feature matrix.

[0013] Due to the adoption of the above technical solutions, this application has significant technical effects:

[0014] The landscape plant geographic information query system and method based on information recognition provided by this application first extracts the morphological attribute features of landscape plants by extracting the features of the image data of landscape plants, and at the same time extracts the distribution spatial geographic features of landscape plants by extracting the features of the names and distribution location information of landscape plants. Then, it fuses the morphological attribute features of landscape plants and the distribution spatial geographic features of landscape plants to obtain the features of landscape plants, and then optimizes the fused features. Finally, the optimized landscape plant features are used by a generator to generate a plant landscape spatial geographic query file. In this way, it can meet diversified query requirements, and the generated plant landscape spatial geographic query file can also be combined with a geographic information system to achieve visual query. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following-described drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. In the drawings:

[0016] Figure 1 FIG. is a system block diagram of a landscape plant geographic information query system based on information recognition according to an embodiment of this application.

[0017] Figure 2 FIG. is a block diagram of a landscape plant related data feature encoding module in a landscape plant geographic information query system based on information recognition according to an embodiment of this application.

[0018] Figure 3 FIG. is a block diagram of a plant image data feature extraction unit in a landscape plant geographic information query system based on information recognition according to an embodiment of this application.

[0019] Figure 4 FIG. is a block diagram of a query file generation module in a landscape plant geographic information query system based on information recognition according to an embodiment of this application.

[0020] Figure 5The flowchart of the method for querying the geographical information of garden plants based on information recognition according to an embodiment of the present application. Detailed implementation manners

[0021] Next, exemplary embodiments of the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited by the exemplary embodiments described herein.

[0022] As mentioned in the above background art, in the real-world scenario of querying the geographical information of garden plants, users may face various challenges. First, users may not be familiar enough with the scientific names or common names of plants, resulting in difficulty in accurately entering plant names when querying. In addition, the description of plant attributes may be ambiguous. For example, the same plant may have different names in different regions, or users may only know some characteristics of the plant rather than the whole picture. These factors may all lead to a decrease in the accuracy and relevance of query results. Second, traditional query methods usually only provide information in text or table form, and this single information display method may be difficult to meet the user's need for intuitiveness. Therefore, an optimized solution for querying the geographical information of garden plants based on information recognition is expected.

[0023] In recent years, deep learning and neural networks have been widely applied in fields such as computer vision, natural language processing, and text signal processing. In addition, deep learning and neural networks have also shown levels close to or even exceeding those of humans in fields such as image classification, object detection, semantic segmentation, and text translation. The development of deep learning and neural networks provides new solutions and ideas for querying the geographical information of garden plants based on information recognition.

[0024] Figure 1 The system block diagram of the system for querying the geographical information of garden plants based on information recognition according to an embodiment of the present application. As Figure 1 shown, in the system 100 for querying the geographical information of garden plants based on information recognition, it includes: a module 110 for collecting data related to garden plants, which is used to obtain the image data of all plants in the garden and the name and distribution location information of all plants; a module 120 for encoding the features of data related to garden plants, which is used to perform feature encoding on the image data of all plants and the name and distribution location information of all plants to obtain a feature matrix of garden plants; a module 130 for generating a query file, which is used to generate a geographical query file for the plant landscape space based on the feature matrix of garden plants.

[0025] Specifically, in the technical solution of the present application, first, image data of all plants in the garden, as well as the names and distribution location information of all the plants, are obtained. It should be well understood that the image data of plants provides intuitive morphological characteristics of plants, including leaf shape, flower color, tree shape, etc., which are key elements for plant identification. High-resolution images can reveal microscopic characteristics of plants, such as leaf texture, detailed structure of flowers, etc., and contribute to accurately identifying plant species. The name of a plant is its identity identifier, while the distribution location information describes the specific location of the plant in the garden. The name of a plant can include the scientific name, alias, common name, etc. of the plant, which is used to identify and distinguish different plant species; the distribution location information of the plant can include longitude and latitude coordinates, the area or plot where it is located, etc., which helps to understand the distribution of plants in the garden. The obtained image data, as well as the names and distribution location information of plants, are the basic data for constructing a plant landscape spatial geographic query file.

[0026] In an embodiment of the present application, an implementable way to obtain the image data of all plants in the garden, as well as the names and distribution location information of all the plants, can be as follows: a. The image data of plants can be obtained by relevant staff through a professional digital camera; b. The distribution location information of plants can be accurately obtained through GPS, and relevant garden staff record the names and distribution location information of each plant and conduct relevant information supplementation and collation.

[0027] Figure 2 It is a block diagram of a garden plant-related data feature encoding module in a garden plant geographic information query system based on information recognition according to an embodiment of the present application. As Figure 2 shown, the garden plant-related data feature encoding module 120 includes: a plant image data feature extraction unit 121, which is used to extract features from the image data of each plant in the image data of all plants to obtain a garden plant morphological attribute feature vector corresponding to the image data of each plant; a plant distribution spatial geographic data feature extraction unit 122, which is used to extract features from the names and distribution location information of each plant in the names and distribution location information of all plants to obtain a garden plant morphological attribute feature vector corresponding to the names and distribution location information of each plant; a plant feature fusion unit 123, which is used to fuse the garden plant morphological attribute feature vectors of each plant and the garden plant distribution spatial geographic feature vectors of each plant for each plant among all plants to obtain a garden plant feature vector of each plant; a data structuring unit 124, which is used to arrange the garden plant feature vectors of all plants according to the plant sample dimension to obtain the garden plant feature matrix.

[0028] In order to extract meaningful information from the original plant images, which can characterize the morphological attributes of plants, such as color, texture, shape, etc., and these information are crucial for the subsequent plant recognition and query processes, it is necessary to separately pass the image data of each plant in the image data of all plants through a plant morphological attribute feature encoder to obtain a landscape plant morphological attribute feature map corresponding to the image data of each plant. In the technical solution of this application, the essence of the plant morphological attribute feature encoder is a convolutional neural network model including a spatial attention module and a channel attention module.

[0029] The spatial attention module can focus on the key regions in the image, that is, those parts that are more important for identifying plant species and characteristics. For example, the characteristic parts such as the leaves, flowers, fruits of plants often contain more identification information. Through the spatial attention module, the network can learn these key regions and give them higher weights during the feature encoding process. The channel attention module can focus on the importance of different channels (i.e., color channels) in the image. In plant images, different color channels may contain different information. For example, the green channel may more reflect the leaf information of plants, while the red channel may more reflect the flower information. Through the channel attention module, the network can learn the importance of different channels and give them appropriate weights during the feature encoding process.

[0030] The convolutional neural network model including the spatial attention module and the channel attention module has a more complex structure and can learn richer feature representations. In contrast, a simple convolutional neural network model may only be able to learn some basic and low-level features and cannot fully capture the complex information in plant images. Therefore, adopting a more complex model structure helps to improve the accuracy and richness of feature encoding, thereby helping to improve the accuracy of the query system.

[0031] In an embodiment of this application, an implementable way to separately pass the image data of each plant in the image data of all plants through a plant morphological attribute feature encoder to obtain a landscape plant morphological attribute feature map corresponding to the image data of each plant can be: passing the image data of each plant through the spatial attention module of the plant morphological attribute feature encoder to obtain a plant morphological spatial attention map corresponding to each plant; passing the image data of each plant through the channel attention module of the plant morphological attribute feature encoder to obtain a plant morphological channel attention map corresponding to each plant; fusing the plant morphological spatial attention map corresponding to each plant and the plant morphological channel attention map corresponding to each plant to obtain the landscape plant morphological attribute feature map corresponding to the image data of each plant.

[0032] The morphological attribute feature map of garden plants contains various morphological attribute features of plants, such as the shape of leaves, the color of flowers, etc. The feature map often has a high dimension, which will increase the complexity of the model during calculation. Therefore, it is necessary to perform pooling on each feature matrix along the channel dimension for the morphological attribute feature map of garden plants corresponding to the image data of each plant to obtain the morphological attribute feature vector of garden plants corresponding to the image data of each plant. Those of ordinary skill in the art should know that pooling is a downsampling operation, usually used to reduce the size of the feature map, reduce the computational complexity and the number of parameters of the model, and at the same time prevent overfitting. In this process, the pooling operation will select a part of the data in the feature map to represent the entire area, so as to realize the compression and feature selection of the feature map. By performing the pooling operation on the morphological attribute feature map of garden plants, it can retain the key features while reducing the computational complexity, which is beneficial to improving the performance of the entire query system.

[0033] Figure 3 It is a block diagram of the plant image data feature extraction unit in the garden plant geographical information query system based on information recognition according to an embodiment of the present application. As Figure 3 shown, the plant image data feature extraction unit 121 includes: a morphological attribute feature generation subunit 1211 of garden plants, which is used to respectively pass the image data of each plant in the image data of all plants through a plant morphological attribute feature encoder to obtain a morphological attribute feature map of garden plants corresponding to the image data of each plant; a feature map pooling subunit 1212, which is used to perform pooling on each feature matrix along the channel dimension for the morphological attribute feature map of garden plants corresponding to the image data of each plant to obtain the morphological attribute feature vector of garden plants corresponding to the image data of each plant.

[0034] The name and distribution location information of plants contains rich semantic information, including important features such as the genus and species of plants and geographical locations. This information is also crucial for identifying and describing the locations of plants in the garden. In order to convert this text information of plants into vector representations so that this information can be understood and processed by algorithms, it is necessary to pass the name and distribution location information of each plant in the name and distribution location information of all plants through a plant spatial distribution attribute context encoder to obtain a garden plant distribution spatial geographical feature vector corresponding to the name and distribution location information of each plant. In the technical solution of this application, the plant spatial distribution attribute context encoder is a Transformer-based semantic encoder that includes an embedding layer. The embedding layer is used to convert discrete input data (such as text data) into continuous vector representations. The Transformer is a deep learning model based on the self-attention mechanism and has powerful semantic representation capabilities. Compared with traditional recurrent neural networks (RNNs) or convolutional neural networks (CNNs), the Transformer can better process long sequence data and capture the dependencies in the data. In the plant spatial distribution attribute context encoder, using the Transformer-based semantic encoder can more effectively extract the semantic features in the plant name and distribution location information, thus more accurately reflecting the distribution rules of different plants in the geographical space.

[0035] In an embodiment of this application, an implementable way to pass the name and distribution location information of each plant in the name and distribution location information of all plants through a plant spatial distribution attribute context encoder to obtain a garden plant distribution spatial geographical feature vector corresponding to the name and distribution location information of each plant can be: perform word segmentation on the name and distribution location information of each plant in the name and distribution location information of all plants to obtain a plant distribution keyword sequence corresponding to the name and distribution location information of each plant; pass the plant distribution keyword sequence corresponding to the name and distribution location information of each plant through the embedding layer of the plant spatial distribution attribute context encoder to obtain a sequence of plant distribution keyword embedding vectors corresponding to the name and distribution location information of each plant; pass the sequence of plant distribution keyword embedding vectors corresponding to the name and distribution location information of each plant through the Transformer module of the plant spatial distribution attribute context encoder to obtain multiple plant distribution keyword semantic feature vectors corresponding to the name and distribution location information of each plant; concatenate the multiple plant distribution word semantic feature vectors corresponding to the name and distribution location information of each plant to obtain the garden plant distribution spatial geographical feature vector corresponding to the name and distribution location information of each plant.

[0036] The characteristics of landscape plants not only include their morphological attributes (such as color, texture, shape, etc.), but also include their distribution locations in the geographical space. Considering any one of these aspects alone may lead to incomplete or biased information. Therefore, in this application, for each of the above-mentioned all plants, the landscape plant morphological attribute feature vectors of each plant and the landscape plant distribution space geographical feature vectors of each plant are fused to obtain the landscape plant feature vectors of each plant. That is, the morphological attribute feature vectors and the distribution space geographical feature vectors respectively reflect different aspects of plants, and there is complementarity between them. By fusing these two vectors, a more comprehensive and accurate representation of plant characteristics can be obtained. Since the fused feature vectors contain the morphological attributes and geographical distribution information of plants, the query requests of users can be matched more quickly. Whether the query is made by inputting the plant name, keywords or uploading a plant photo, the system can quickly match and respond according to the fused feature vectors.

[0037] In an embodiment of this application, for each of the above-mentioned all plants, an implementable way to fuse the landscape plant morphological attribute feature vectors of each plant and the landscape plant distribution space geographical feature vectors of each plant to obtain the landscape plant feature vectors of each plant can be: use the following fusion formula to fuse the landscape plant morphological attribute feature vectors of each plant and the landscape plant distribution space geographical feature vectors of each plant to obtain the landscape plant feature vectors of each plant; where, the fusion formula is: ; where, represents the landscape plant morphological attribute feature vectors of each plant, represents the landscape plant distribution space geographical feature vectors of each plant, represents the concatenation function, represents the landscape plant feature vectors of each plant.

[0038] Immediately afterwards, the landscape plant feature vectors of all the plants are arranged according to the plant sample dimension to obtain a landscape plant feature matrix. It is not difficult to understand that by arranging the feature vectors of all plants into a matrix, the information of each plant can be integrated together. This data structure that integrates the information of all plants facilitates subsequent algorithm processing and analysis, making the organization and access of data more efficient.

[0039] Figure 4 It is a block diagram of a query file generation module in a landscape plant geographical information query system based on information recognition according to an embodiment of this application. As Figure 4As shown, the query file generation module 130 includes: a feature matrix information regression unit 131, configured to perform information regression on the landscape plant feature matrix in a divergent-convergent mode to obtain an optimized landscape plant feature matrix; and a feature matrix parsing unit 132, configured to obtain a generation result by passing the optimized landscape plant feature matrix through a generator, where the generation result is a plant landscape spatial geographic query file.

[0040] Specifically, in the technical solution of the present application, the landscape plant feature matrix is obtained by integrating the morphological attribute features of plants and the feature vectors encoded by the spatial distribution attribute context. Improving the information aggregation degree of the landscape plant feature matrix is crucial for generating an accurate plant landscape spatial geographic query file. Specifically, improving the information aggregation degree of the landscape plant feature matrix can help capture the diverse attributes and features of plants. By comprehensively considering the morphological attributes and spatial distribution attributes of plants, the characteristics of each plant can be more comprehensively described, thereby increasing the diversity of the generation results. At the same time, a high information aggregation degree means that the characteristics of each plant can be more accurately characterized. By integrating the morphological attribute features and the spatial distribution attribute context encoding, the accuracy of plant characteristics can be improved, thereby generating a more reliable plant landscape spatial geographic query file. By improving the information aggregation degree of the landscape plant feature matrix, the representation ability and generalization ability of the generator can be increased. The generator can better learn the complex relationships between plant characteristics, thereby improving the accuracy and quality of the generation results. An accurate landscape plant feature matrix can provide richer and more practical information for the plant landscape spatial geographic query file. By comprehensively considering the morphological characteristics and spatial distribution attributes of plants, the generated query file can better meet the needs of users and provide more useful information. Based on this, information regression in a divergent-convergent mode is performed on the landscape plant feature matrix to obtain an optimized landscape plant feature matrix.

[0041] Specifically, performing information regression on the landscape plant feature matrix in a divergent-convergent mode to obtain an optimized landscape plant feature matrix includes: calculating the natural exponential function values at each position in the landscape plant feature matrix to obtain a divergent landscape plant feature matrix; performing cumulative summation on the divergent landscape plant feature matrix to obtain the total value of the divergent landscape plant feature matrix; multiplying the total value of the divergent landscape plant feature matrix by the absolute value of the feature value at each position in the landscape plant feature matrix and then calculating the logarithm function value with base 2 to obtain a logarithm-transformed landscape plant feature matrix; and multiplying the logarithm-transformed landscape plant feature matrix and the landscape plant feature matrix position by position to obtain the optimized landscape plant feature matrix.

[0042] In an embodiment of the present application, another implementable manner of performing information regression in an imitation divergence - aggregation mode on the landscape plant feature matrix to obtain an optimized landscape plant feature matrix may be: performing information regression in an imitation divergence - aggregation mode on the landscape plant feature matrix with the following optimization formula, where the optimization formula is: ; where is the eigenvalue at the th position of the landscape plant feature matrix, and and are the width and height of the landscape plant feature matrix , represents the logarithm to the base 2, is the eigenvalue at the th position of the optimized landscape plant feature matrix.

[0043] Here, in order to improve the information aggregation degree of the landscape plant feature matrix, in the technical solution of the present application, information regression in an imitation divergence - aggregation mode is performed on the landscape plant feature matrix. First, the natural exponential function values at each position in the landscape plant feature matrix are calculated to diverge the feature information at that position, and a wavelet - like function is used to structurally aggregate the divergence information at all positions in the landscape plant feature matrix to obtain a global divergence factor. After applying the global divergence factor to each position of the landscape plant feature matrix, the logarithm function value to the base 2 is calculated for the divergence - modulated information at that position to perform regression on the feature values at each position of the landscape plant feature matrix. In this way, through the mechanism of imitation divergence - regression, the feature values at each position of the landscape plant feature matrix are inflated and concentrated, and in this way, the information aggregation degree of the landscape plant feature matrix is improved.

[0044] Finally, the optimized landscape plant feature matrix is passed through a generator to obtain a generation result, and the generation result is a plant landscape space geographical query file. The generator can be a machine - learning model for generating a plant landscape space geographical query file according to the input optimized landscape plant feature matrix. The plant landscape space geographical query file is a comprehensive query tool library integrating information such as plant images, names, distribution locations, and morphological attribute features. Combining the plant landscape space geographical query file with the existing geographical information system can form an interactive online platform. Users can query the distribution locations of plants in the garden by inputting plant names, keywords, or uploading plant photos to achieve visual query of plant distribution. In addition to displaying the plant distribution locations, interactive functions are also provided. For example, clicking on the distribution location of the searched plant can also view the detailed information of the plant. In this way, more accurate, intuitive, and customized plant geographical information query services can be provided for users.

[0045] In summary, the garden plant geographical information query system 100 based on information recognition according to the embodiments of the present application is elucidated. First, it extracts features from the image data of garden plants to obtain the morphological attribute features of garden plants. At the same time, it extracts features from the name and distribution location information of garden plants to obtain the distribution spatial geographical features of garden plants. Then, it fuses the morphological attribute features of garden plants and the distribution spatial geographical features of garden plants to obtain garden plant features. Next, it optimizes the fused features. Finally, it uses a generator to generate a plant landscape spatial geographical query file from the optimized garden plant features. In this way, it can meet diversified query requirements and can also combine the generated plant landscape spatial geographical query file with a geographical information system to achieve visual query.

[0046] As described above, the garden plant geographical information query system 100 based on information recognition according to the embodiments of the present application can be implemented in various terminal devices, such as a server for querying the geographical information of garden plants based on information recognition. In one example, the garden plant geographical information query system 100 based on the embodiments of the present application can be integrated into a terminal device as a software module and / or a hardware module. For example, the garden plant geographical information query system 100 can be a software module in the operating system of the terminal device, or can be an application program developed for the terminal device; of course, the garden plant geographical information query system 100 can also be one of the many hardware modules of the terminal device.

[0047] Alternatively, in another example, the garden plant geographical information query system 100 and the terminal device can also be separate devices, and the garden plant geographical information query system 100 can be connected to the terminal device through a wired and / or wireless network and transmit interaction information in accordance with a predefined data format.

[0048] Figure 5 It is a flowchart of the garden plant geographical information query method based on information recognition according to the embodiments of the present application. As Figure 5 shown, in the garden plant geographical information query method based on information recognition, it includes: S110, obtaining the image data of all plants in the garden and the name and distribution location information of all the plants; S120, performing feature encoding on the image data of all the plants and the name and distribution location information of all the plants to obtain a garden plant feature matrix; S130, generating a plant landscape spatial geographical query file based on the garden plant feature matrix.

[0049] Here, those skilled in the art can understand that the specific operations of each step in the above-described method for querying the geographical information of garden plants based on information recognition have been described in detail in the description of the system for querying the geographical information of garden plants based on information recognition with reference to Figures 1 to 4 and thus, the repeated description thereof will be omitted.

[0050] In summary, the method for querying the geographical information of garden plants based on information recognition according to the embodiments of the present application is elucidated. First, the morphological attribute features of garden plants are obtained by extracting features from the image data of garden plants. At the same time, the spatial geographical features of the distribution of garden plants are obtained by extracting features from the name and distribution location information of garden plants. Then, the morphological attribute features of garden plants and the spatial geographical features of the distribution of garden plants are fused to obtain the features of garden plants. Then, the fused features are optimized. Finally, the optimized features of garden plants are used by the generator to generate a plant landscape spatial geographical query file. In this way, diverse query requirements can be met, and the generated plant landscape spatial geographical query file can be combined with a geographical information system to achieve visual query.

[0051] In addition to the above systems and methods, an embodiment of the present application may also be a computer program product, which includes computer program instructions that, when run by a processor, cause the processor to execute the steps in the method for querying the geographical information of garden plants based on information recognition according to various embodiments of the present application described in this specification.

[0052] The computer program product can be written in any combination of one or more programming languages for the program code to perform the operations of the embodiments of the present application. The programming languages include object-oriented programming languages such as Java, C++, etc., and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computing device, partially on the user's device, executed as an independent software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0053] Furthermore, an embodiment of the present application may also be a computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are run by a processor, the processor is caused to execute the steps in the method for querying the geographical information of garden plants based on information recognition according to various embodiments of the present application described in this specification.

[0054] The computer-readable storage medium may adopt any combination of one or more readable media. The readable media may be a readable signal medium or a readable storage medium. The readable storage medium may include, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0055] The basic principles of the present application have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, benefits, effects, etc. mentioned in the present application are only examples and not limitations. It cannot be considered that these advantages, benefits, effects, etc. are essential for each embodiment of the present application. In addition, the specific details of the above application are only for the purposes of illustration and easy understanding, rather than limitations. The above details do not limit the present application to necessarily adopt the above specific details for implementation.

[0056] Words such as "including", "comprising", "having", etc. in the present application are open-ended terms, meaning "including but not limited to", and can be used interchangeably with each other. The words "or" and "and" used herein refer to the word "and / or", and can be used interchangeably with each other, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to", and can be used interchangeably with each other.

[0057] In addition, as used herein, the "or" used in the listing of items starting with "at least one" indicates a separate listing, so that for example, the listing of "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). In addition, the term "exemplary" does not mean that the described examples are preferred or better than other examples.

[0058] It should also be noted that in the systems and methods of the present application, each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent solutions of the present application.

[0059] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be very apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the broadest scope consistent with the principles and novel features claimed herein.

[0060] The foregoing description has been presented for purposes of illustration and description. In addition, this description is not intended to limit embodiments of the present application to the forms disclosed herein. Although several example aspects and embodiments have been discussed above, those skilled in the art will recognize some of their variations, modifications, alterations, additions, and sub-combinations.

Claims

1. A garden plant geographic information query system based on information recognition, characterized in that: include: The garden plant related data acquisition module is used to obtain the image data of all plants in the garden and the names and distribution location information of all the plants; A garden plant related data feature encoding module, used for performing feature encoding on the image data of all the plants and the names and distribution location information of all the plants to obtain a garden plant feature matrix; A query file generation module, used to generate a plant landscape spatial geographic query file based on the garden plant feature matrix; Wherein, the query file generation module includes: A feature matrix information regression unit is used to perform information regression in a simulated divergence-aggregation mode on the garden plant feature matrix to obtain an optimized garden plant feature matrix; A feature matrix parsing unit, used for passing the optimized garden plant feature matrix through a generator to obtain a generation result, wherein the generation result is a plant landscape spatial geographic query file; Wherein, the feature matrix information regression unit is used to: Calculating the natural exponential function value of each position in the garden plant characteristic matrix to obtain a divergent garden plant characteristic matrix; Accumulating and summing the divergent garden plant feature matrix to obtain a total value of the divergent garden plant feature matrix; The sum of the divergent garden plant characteristic matrix and the absolute value of the characteristic value of each position of the garden plant characteristic matrix are multiplied, and then a logarithmic function value with base two is calculated to obtain a garden plant characteristic matrix after logarithmic change; The logarithmically changed garden plant feature matrix is ​​multiplied by the garden plant feature matrix by position to obtain the optimized garden plant feature matrix.

2. The garden plant geographic information query system based on information identification according to claim 1 is characterized in that: The garden plant related data feature coding module includes: A plant image data feature extraction unit, used for extracting features from the image data of each plant in the image data of all plants to obtain a garden plant morphological attribute feature vector corresponding to the image data of each plant; A plant distribution spatial geographic data feature extraction unit is used to extract features from the name and distribution location information of each plant in the name and distribution location information of all plants to obtain a garden plant distribution spatial geographic feature vector corresponding to the name and distribution location information of each plant; A plant feature fusion unit, for fusing the garden plant morphological attribute feature vector of each plant and the garden plant distribution spatial geographic feature vector of each plant to obtain a garden plant feature vector of each plant; The data structuring unit is used to arrange the garden plant feature vectors of all the plants according to the plant sample dimension to obtain the garden plant feature matrix.

3. The garden plant geographic information query system based on information identification according to claim 2 is characterized in that: The plant image data feature extraction unit comprises: A garden plant morphological attribute feature generating subunit, used for respectively passing the image data of each plant in the image data of all plants through a plant morphological attribute feature encoder to obtain a garden plant morphological attribute feature map corresponding to the image data of each plant; The feature map pooling subunit is used to perform pooling of each feature matrix along the channel dimension on the garden plant morphological attribute feature map corresponding to the image data of each plant to obtain the garden plant morphological attribute feature vector corresponding to the image data of each plant.

4. The garden plant geographic information query system based on information identification according to claim 3 is characterized in that: The plant distribution spatial geographic data feature extraction unit is used to: pass the name and distribution location information of each plant in the name and distribution location information of all plants through a plant spatial distribution attribute context encoder to obtain the garden plant distribution spatial geographic feature vector corresponding to the name and distribution location information of each plant.

5. The garden plant geographic information query system based on information identification according to claim 4 is characterized in that: The plant morphological attribute feature encoder is a convolutional neural network model including a spatial attention module and a channel attention module, and the plant spatial distribution attribute context encoder is a converter-based semantic encoder including an embedding layer.

6. A method for querying geographic information of garden plants based on information recognition, characterized in that: include: Obtaining image data of all plants in the garden and the names and distribution location information of all the plants; Performing feature coding on the image data of all the plants and the names and distribution location information of all the plants to obtain a garden plant feature matrix; Based on the garden plant feature matrix, generating a plant landscape spatial geographic query file; Wherein, based on the garden plant feature matrix, a plant landscape spatial geographic query file is generated, including: Performing information regression of the garden plant feature matrix in a simulated divergence-aggregation mode to obtain an optimized garden plant feature matrix; The optimized garden plant feature matrix is ​​passed through a generator to obtain a generation result, wherein the generation result is a plant landscape spatial geographic query file; Wherein, the garden plant feature matrix is ​​subjected to information regression in a simulated divergence-aggregation mode to obtain an optimized garden plant feature matrix, which is used for: Calculating the natural exponential function value of each position in the garden plant characteristic matrix to obtain a divergent garden plant characteristic matrix; Accumulating and summing the divergent garden plant feature matrix to obtain a total value of the divergent garden plant feature matrix; The sum of the divergent garden plant characteristic matrix and the absolute value of the characteristic value of each position of the garden plant characteristic matrix are multiplied, and then a logarithmic function value with base two is calculated to obtain a garden plant characteristic matrix after logarithmic change; The logarithmically changed garden plant feature matrix is ​​multiplied by the garden plant feature matrix by position to obtain the optimized garden plant feature matrix.

7. The garden plant geographic information query method based on information identification according to claim 6 is characterized in that: Performing feature coding on the image data of all the plants and the names and distribution location information of all the plants to obtain a garden plant feature matrix, including: Performing feature extraction on the image data of each plant in the image data of all plants to obtain a garden plant morphological attribute feature vector corresponding to the image data of each plant; Extracting features from the name and distribution location information of each plant in the name and distribution location information of all plants to obtain a spatial geographic feature vector of garden plant distribution corresponding to the name and distribution location information of each plant; For each of the plants, the garden plant morphological attribute feature vector of each plant and the garden plant distribution spatial geographic feature vector of each plant are integrated to obtain the garden plant feature vector of each plant; The garden plant feature vectors of all the plants are arranged according to the plant sample dimension to obtain the garden plant feature matrix.

Images