Smart garden management system based on combination of Internet of Things and GIS technology

Abstract

The invention relates to the field of garden management, and particularly discloses a smart garden management system based on the combination of the Internet of Things and the GIS technology, and the system comprises the steps: firstly obtaining the point cloud data of a landscape garden project and the environment image of the landscape garden project obtained by the GIS, and then carrying out the feature extraction and correlation analysis of the two through employing the deep learning technology, and obtaining the feature of the landscape garden project; and finally, generating an urban garden management plant distribution 3D display interface through a generator, thereby facilitating visual understanding of the spatial layout and plant distribution conditions of the garden, and further helping to make garden management and maintenance strategies.

Description

Technical Field

[0001] The present application relates to the field of garden management, and more specifically, to an intelligent garden management system based on the combination of Internet of Things and GIS technology. Background Art

[0002] Urban gardens refer to green areas designed and maintained in urban environments, including parks, flower beds, greenways and green belts, etc. These gardens not only beautify the urban landscape, but also provide functions of leisure and entertainment, ecological balance and environmental improvement.

[0003] In the field of urban landscaping, with the acceleration of urbanization, the number and types of landscaping facilities are increasing, and the coverage is also expanding. This makes the traditional garden management methods face many challenges, such as low management efficiency, difficulty in information integration and high maintenance costs. Traditional methods usually rely on manual records and on-site inspections, which are difficult to cope with the rapidly changing urban greening needs.

[0004] Therefore, a smart garden management system based on the combination of Internet of Things and GIS technology is expected. Summary of the invention

[0005] In order to solve the above technical problems, this application is proposed. The embodiment of this application provides a smart garden management system based on the combination of Internet of Things and GIS technology, which first obtains the point cloud data of the landscape engineering and the environmental image of the landscape engineering obtained by GIS, and then uses deep learning technology to perform feature extraction and correlation analysis on the two, and finally generates a 3D display interface of urban garden management plant distribution through a generator, so as to facilitate intuitive understanding of the spatial layout and plant distribution of the garden, and then help the formulation of garden management and maintenance strategies.

[0006] According to one aspect of the present application, a smart garden management system based on the combination of Internet of Things and GIS technology is provided, which includes:

[0007] Landscape engineering data acquisition module, used to acquire point cloud data of landscape engineering and environmental images of landscape engineering acquired by GIS;

[0008] A landscape engineering data extraction module is used to extract a landscape engineering point cloud semantic feature vector and a landscape engineering environment global feature vector from the landscape engineering point cloud data and the landscape engineering environment image acquired by GIS;

[0009] The garden plant distribution 3D display generation module is used to generate a 3D display interface of urban garden management plant distribution based on the landscape engineering point cloud semantic feature vector and the landscape engineering environment global feature vector.

[0010] Compared with the prior art, the present application provides a smart garden management system based on the combination of Internet of Things and GIS technology, which first obtains the point cloud data of the landscape project and the environmental image of the landscape project obtained by GIS, and then uses deep learning technology to perform feature extraction and correlation analysis on the two, and finally uses a generator to generate a 3D display interface of plant distribution for urban garden management, so as to facilitate the intuitive understanding of the spatial layout of the garden and the distribution of plants, and thus help to formulate garden management and maintenance strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0012] Figure 1 It is a block diagram of a smart garden management system based on the combination of Internet of Things and GIS technology according to an embodiment of the present application.

[0013] Figure 2 It is a block diagram of a landscape engineering data extraction module in a smart garden management system based on the combination of Internet of Things and GIS technology according to an embodiment of the present application.

[0014] Figure 3 It is a block diagram of a garden environment image feature extraction unit in a smart garden management system based on the combination of Internet of Things and GIS technology according to an embodiment of the present application.

[0015] Figure 4 This is a block diagram of a 3D display generation module for garden plant distribution in a smart garden management system based on the combination of Internet of Things and GIS technology according to an embodiment of the present application. DETAILED DESCRIPTION

[0016] Various exemplary embodiments, features and aspects of the present application will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0017] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0018] In addition, in order to better illustrate the present application, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present application can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the subject matter of the present application.

[0019] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0020] Figure 1 FIG. 1 is a block diagram of a smart garden management system based on the combination of the Internet of Things and GIS technology according to an embodiment of the present application. Figure 1 As shown, according to the embodiment of the present application, the smart garden management system 100 based on the combination of Internet of Things and GIS technology includes: a landscape engineering data acquisition module 110, which is used to acquire the point cloud data of the landscape engineering and the environmental image of the landscape engineering acquired by GIS; a landscape engineering data extraction module 120, which is used to extract the landscape engineering point cloud semantic feature vector and the landscape engineering environment global feature vector from the point cloud data of the landscape engineering and the environmental image of the landscape engineering acquired by GIS; a garden plant distribution 3D display generation module 130, which is used to generate a 3D display interface of urban garden management plant distribution based on the landscape engineering point cloud semantic feature vector and the landscape engineering environment global feature vector.

[0021] In the above-mentioned smart garden management system 100 based on the combination of the Internet of Things and GIS technology, the landscape engineering data acquisition module 110 is used to obtain the point cloud data of the landscape engineering and the environmental image of the landscape engineering obtained by GIS. It should be understood that urban gardens refer to green areas designed and maintained in cities, such as parks, flower beds, greenways and green belts. These gardens not only improve the urban landscape, but also provide functions of leisure and entertainment, maintaining ecological balance and improving environmental quality. With the acceleration of urbanization, there are more and more garden greening facilities, the types are also increasing, and the coverage area is gradually expanding. This trend has brought challenges to traditional garden management, such as low management efficiency, difficulty in information integration and rising maintenance costs. Among them, traditional management methods rely on manual records and on-site inspections, which are difficult to adapt to the rapidly changing greening needs. Therefore, in the technical solution of the present application, the point cloud data of the landscape engineering and the environmental image of the landscape engineering obtained by GIS are obtained, and combined with deep learning technology, a three-dimensional display interface of the distribution of urban garden plants is generated to intuitively present the spatial layout and plant distribution of the garden. This visualization tool helps to understand the overall structure of the garden more clearly, thereby supporting the formulation of more effective garden management and maintenance strategies.

[0022] Specifically, in the technical solution of the present application, point cloud data is obtained through laser scanning or photogrammetry technology, which can record the three-dimensional coordinates of terrain and vegetation in detail, providing an in-depth understanding of the complexity and details of the terrain. On the other hand, the environmental images obtained through GIS (Geographic Information System) provide a visual understanding of the geographical and environmental background of the landscape project area. Among them, GIS images usually contain detailed information about terrain, land use, transportation networks and surrounding environment. Combining point cloud data and GIS environmental images, a comprehensive view can be obtained, so that more scientific decisions can be made when conducting landscape design. Ultimately, this data integration can improve the accuracy of the design, optimize resource allocation, and ensure the sustainability and environmental compatibility of the project.

[0023] In the above-mentioned smart garden management system 100 based on the combination of the Internet of Things and GIS technology, the landscape engineering data extraction module 120 is used to extract the landscape engineering point cloud semantic feature vector and the landscape engineering environment global feature vector from the point cloud data of the landscape engineering and the environmental image of the landscape engineering acquired by GIS. It should be understood that the semantic feature vector of the point cloud data provides an accurate analysis of specific terrain and objects, while the global feature vector of the GIS environmental image provides a comprehensive understanding of the macro environment. The combination of the two can achieve a comprehensive and scientific design and management of landscape projects, and improve the implementation effect and sustainability of the project.

[0024] Figure 2FIG. 1 is a block diagram of a landscape engineering data extraction module in a smart garden management system based on the combination of the Internet of Things and GIS technology according to an embodiment of the present application. Figure 2 As shown, in a specific embodiment of the present application, the landscape engineering data extraction module 120 includes: a garden point cloud data semantic feature extraction unit 121, which is used to perform semantic feature extraction on the point cloud data of the landscape engineering to obtain the semantic feature vector of the landscape engineering point cloud; a garden point cloud data spatial feature extraction unit 122, which is used to perform spatial feature extraction on the point cloud data of the landscape engineering to obtain a landscape engineering point cloud depth feature map; a garden environment image feature extraction unit 123, which is used to perform feature extraction on the environment image of the landscape engineering acquired by GIS to obtain a landscape engineering environment enhancement feature map; a landscape engineering feature aggregation unit 124, which is used to perform feature aggregation on the landscape engineering point cloud depth feature map and the landscape engineering environment enhancement feature map to obtain the landscape engineering environment global feature vector.

[0025] It should be understood that point cloud data is generated by laser scanning or photogrammetry technology, which contains rich three-dimensional spatial information, such as terrain undulations, vegetation distribution, and building layout. However, these data usually exist in the form of a large number of raw point sets, which are difficult to use directly for design and analysis. In order to extract useful information from these massive point cloud data, semantic feature extraction is required to convert these data into semantic feature vectors that can express the attributes and categories of objects. Among them, point cloud data is essentially a high-density three-dimensional point set, each point represents a position in space, but these points do not directly provide information about different object categories. By extracting semantic feature vectors, these points can be classified into different object types, such as trees, buildings, roads, etc. This classification can be based on features such as point geometry, spatial distribution, and neighborhood relationships, thereby converting complex point cloud data into more meaningful structured information. This structured data can greatly improve designers' understanding of site conditions, making the design and planning process more accurate and efficient.

[0026] Furthermore, spatial feature extraction of point cloud data involves extracting features related to the spatial structure of objects from the original three-dimensional point set data. These spatial features include information such as the density, distribution, shape, and relative position of the points, which are crucial for understanding and analyzing the terrain and objects in point cloud data. Point cloud data consists of a large number of three-dimensional points, each of which represents a position in space. By analyzing the spatial relationship of these points, detailed information about terrain undulations, object contours, and spatial structures can be obtained. The resulting point cloud depth feature map converts the spatial distribution and structure of the points into depth values ​​in the image, thereby displaying the spatial information of the point cloud data in a more intuitive and processable form.

[0027] Furthermore, feature extraction of environmental images acquired by GIS is to convert the original environmental images acquired by GIS into feature maps that can provide more useful information. These environmental images usually include information such as terrain, land use, transportation network, vegetation coverage, etc. Through feature extraction, key environmental features can be extracted from them, such as land feature boundaries, functional zoning, land type, etc. The extraction process of these features mainly includes identifying and extracting important spatial information from the image, such as the shape, location and relationship of the land features. The obtained environmental enhancement feature map can significantly improve the recognition ability of environmental factors by strengthening different features in the environmental image.

[0028] In particular, in landscape engineering, feature aggregation of point cloud deep feature maps and environmental enhancement feature maps is to integrate key information from different data sources into a unified global feature vector. Among them, the point cloud deep feature map provides detailed three-dimensional spatial information of the site, including terrain undulations, the location and shape of vegetation and buildings. The environmental enhancement feature map provides two-dimensional enhanced information about land use, transportation network and environmental elements. These two feature maps describe the spatial structure and functional characteristics of the environment from different perspectives. Aggregating these features can combine these two types of information to obtain a global feature vector that can fully reflect the characteristics of the environment. Among them, the point cloud deep feature map and the environmental enhancement feature map highlight the three-dimensional spatial structure and two-dimensional functional characteristics respectively. Using the information of any data source alone may not be able to fully depict the complexity of the entire environment. Through feature aggregation, the deep features and environmental features can be integrated into a global feature vector, which can more comprehensively capture the comprehensive information of the site, including terrain, use, facility distribution and other aspects. Such a comprehensive perspective enables designers and planners to understand the overall picture of the site more accurately, so as to formulate more scientific and practical design plans.

[0029] In a specific embodiment of the present application, the garden point cloud data semantic feature extraction unit 121 includes: passing the point cloud data of the landscape engineering project through a landscape engineering point cloud context encoder to obtain a plurality of landscape engineering point cloud feature vectors; and cascading the plurality of landscape engineering point cloud feature vectors into the landscape engineering point cloud semantic feature vector.

[0030] It should be understood that point cloud data contains rich three-dimensional spatial information, such as terrain, buildings and vegetation distribution. However, the original point cloud data exists in the form of a large number of point sets, and its spatial structure and object attributes are difficult to be directly used for analysis and decision-making. The core role of the landscape engineering point cloud context encoder is to convert these point cloud data into feature vectors with more semantic and structural information. This encoder can extract useful features by understanding the spatial contextual relationship in the point cloud data, thereby generating multiple point cloud feature vectors. These feature vectors not only contain the geometric shape of the point cloud, but also integrate the relationship between points and contextual information, making the expression of the data richer and more meaningful. Specifically, the point cloud context encoder can capture the spatial structure and local features in the point cloud data. By analyzing the relative position and neighborhood information of the points, the encoder can identify and extract the shape features, boundary information and structural features of the objects. This extraction capability enables the generated point cloud feature vector to accurately reflect the important spatial information in the point cloud, such as terrain undulations, building outlines and vegetation distribution. This understanding of spatial context provides a basis for subsequent analysis and processing. The multiple point cloud feature vectors generated by the context encoder support multi-level analysis and application. Each feature vector represents a specific aspect or level of information in the point cloud data, and multiple feature vectors can comprehensively cover various features in the point cloud data. In this way, the point cloud data can be analyzed in multiple dimensions. This multi-level information extraction not only improves the level of detail in data analysis, but also supports more complex applications, such as automated design and intelligent decision-making. Specifically, the embedding layer of the landscape engineering point cloud context encoder is used to convert the point cloud data of the landscape engineering into an embedding vector to obtain a sequence of point cloud embedding vectors; the converter-based Bert model of the landscape engineering point cloud context encoder is used to perform global context semantic encoding on the sequence of point cloud embedding vectors to obtain multiple landscape engineering point cloud feature vectors.

[0031] Furthermore, cascading multiple landscape engineering point cloud feature vectors into a landscape engineering point cloud semantic feature vector can achieve comprehensive integration and enhancement of feature information, thereby improving the understanding and application capabilities of point cloud data. Among them, the cascaded feature vectors can make full use of the rich information contained in different feature vectors. In landscape engineering, multiple point cloud feature vectors usually capture different aspects of point cloud data, such as spatial geometry, object type, positional relationship, etc. Using these feature vectors alone may not fully reflect the complexity of the data. By cascading these feature vectors, the various feature dimensions can be merged into a comprehensive semantic feature vector. This cascading method enables the final semantic feature vector to contain more information levels and provide a comprehensive description of the environment, thereby supporting more complex and detailed analysis tasks.

[0032] In a specific embodiment of the present application, the garden point cloud data spatial feature extraction unit 122 includes: spatially mapping the point cloud data of the landscape engineering project to obtain a landscape engineering point cloud projection map; passing the landscape engineering point cloud projection map through a landscape engineering point cloud projection convolutional neural network as a feature extractor to obtain a landscape engineering point cloud depth feature map.

[0033] It should be understood that spatial mapping can convert 3D point cloud data into 2D projections, making the display and analysis of data more intuitive and easier to handle. Point cloud data is essentially composed of a large number of 3D points, which contain the position, shape and other properties in space. In many cases, directly processing these 3D data can be very complicated and time-consuming. Through spatial mapping, the point cloud data is projected onto a 2D plane, which can simplify the originally complex 3D information into a 2D image that is easier to handle. This projection can effectively display the spatial distribution, shape contours and structural characteristics of the point cloud data, making the interpretation of the data more intuitive and clear.

[0034] Furthermore, the point cloud projection map of landscape engineering converts 3D point cloud data into 2D image form, simplifying the complexity of the data. Point cloud data originally has a large number of 3D points, which makes its processing and analysis complicated. The projection map obtained by spatial mapping provides a more concise 2D perspective, which is convenient for subsequent data processing. However, the 2D representation of the projection map still requires deep feature extraction to obtain useful information, which requires the help of convolutional neural networks. Convolutional neural networks (CNNs) can effectively capture local features and spatial relationships in images through their convolutional layers. In the point cloud projection map, CNN can identify key image features such as edges, textures, and shapes. These features are very important for understanding the spatial layout and structure of point cloud data. Multiple convolutional layers in the network can extract features from simple to complex layer by layer, thereby constructing detailed point cloud deep feature maps. These deep feature maps not only contain the geometric information of the projection map, but also integrate higher-level semantic features, making the analysis of the environment more comprehensive and accurate. The deep feature maps extracted by CNN provide a high-level expression of point cloud data, which is essential for various tasks in landscape engineering. Specifically, each layer of the landscape engineering point cloud projection convolutional neural network used as a feature extractor performs convolution processing, mean pooling processing based on the local feature matrix and nonlinear activation processing on the input data in the forward pass of the layer, so that the last layer of the landscape engineering point cloud projection convolutional neural network used as a feature extractor outputs the landscape engineering point cloud depth feature map, wherein the input of the landscape engineering point cloud projection convolutional neural network used as a feature extractor is the landscape engineering point cloud projection map.

[0035] Figure 3FIG. 1 is a block diagram of a garden environment image feature extraction unit in a smart garden management system based on the combination of the Internet of Things and GIS technology according to an embodiment of the present application. Figure 3 As shown, in a specific embodiment of the present application, the garden environment image feature extraction unit 123 includes: a garden environment image pixel enhancement subunit 1231, which is used to pass the environmental image of the landscape engineering acquired by GIS through a landscape engineering environment pixel enhancement generator to obtain a landscape engineering environment enhanced image; a garden environment image spatial attention encoding subunit 1232, which is used to pass the landscape engineering environment enhanced image through a landscape engineering environment enhancement feature encoder based on a spatial attention mechanism to obtain the landscape engineering environment enhancement feature map.

[0036] It should be understood that environmental images acquired by GIS are usually based on satellite or aerial photography technology, which are often affected by multiple factors such as resolution, lighting conditions, and sensor quality. These factors may cause problems such as blur, noise, or loss of details in the image, which has an adverse effect on subsequent analysis and application. The core function of the environmental pixel enhancement generator is to enhance the original environmental image through advanced image processing algorithms, thereby improving the image quality and enhancing its clarity and details. This enhancement makes the key features in the image more obvious, which helps to conduct more accurate environmental analysis and design planning. By applying image enhancement techniques such as contrast enhancement, edge sharpening, noise removal, etc., the generator is able to enhance the edges and details of objects in the image. This is especially important for landscape engineering, because the clarity of details directly affects the recognition and analysis of environmental features. Specifically, the environment image of the landscape engineering acquired by the GIS is input into the encoder of the landscape engineering environment pixel enhancement generator, wherein the encoder uses a convolution layer to perform explicit spatial encoding on the environment image of the landscape engineering acquired by the GIS to obtain image features; and the image features are input into the decoder of the landscape engineering environment pixel enhancement generator, wherein the decoder uses a deconvolution layer to perform deconvolution processing on the image features to obtain the landscape engineering environment enhanced image.

[0037] Furthermore, the landscape engineering environmental enhancement image is processed by an environmental enhancement feature encoder based on the spatial attention mechanism, and the spatial attention mechanism is used to effectively focus on the important areas in the image, thereby extracting more representative feature information. Considering that the generation of environmental enhancement images has improved the image quality through image enhancement technology, there may still be a lot of information redundancy and background noise in these images. In order to extract useful features from these enhanced images, more sophisticated feature encoding methods must be applied. The feature encoder based on the spatial attention mechanism can dynamically focus on the important areas in the image and ignore irrelevant background. This mechanism can identify key environmental features in the image, such as vegetation, terrain changes, and building structures, thereby improving the accuracy of feature extraction. Among them, the spatial attention mechanism optimizes feature extraction by weighting the importance of different regions. In the image, different regions contribute unequally to the analysis task. Through the spatial attention mechanism, the encoder can automatically adjust the degree of attention to each part of the image. For example, in landscape images, densely planted areas or key landscape areas may require more attention, while other areas can be suppressed. In this way, the feature encoder can more effectively capture these important environmental features and avoid excessive processing of irrelevant information. Specifically, the convolutional coding part of the landscape engineering environment enhancement feature encoder based on the spatial attention mechanism is used to perform deep convolutional coding on the landscape engineering environment enhancement image to obtain an initial convolutional feature map; the initial convolutional feature map is input into the spatial attention part of the landscape engineering environment enhancement feature encoder based on the spatial attention mechanism to obtain a spatial attention map; the spatial attention map is activated by the Softmax function to obtain a spatial attention feature map; the spatial attention feature map and the initial convolutional feature map are multiplied by the position point to obtain the landscape engineering environment enhancement feature map. More specifically, the initial convolutional feature map is input into the spatial attention part of the landscape engineering environment enhancement feature encoder based on the spatial attention mechanism to obtain the spatial attention map, including: performing average pooling and maximum pooling along the channel dimension on the initial convolutional feature map to obtain an average feature matrix and a maximum feature matrix; cascading and channel-adjusting the average feature matrix and the maximum feature matrix to obtain a channel feature matrix; using the convolutional layer of the spatial attention feature map to convolutionally encode the channel feature matrix to obtain the spatial attention map.

[0038] In a specific embodiment of the present application, the landscape engineering feature aggregation unit 124 includes: associating the landscape engineering point cloud depth feature map and the landscape engineering environment enhancement feature map to obtain a landscape engineering environment global feature map; pooling the landscape engineering environment global feature map to obtain a landscape engineering environment global feature vector.

[0039] It should be understood that associating these feature maps helps to enhance the accuracy and robustness of the model. In landscape engineering, the generation of global feature maps of the environment should not only accurately describe the spatial geometry, but also reflect the visual characteristics of the environment. Combining point cloud and image features can make up for the shortcomings of a single feature map and improve the ability to model complex environments. By combining information from different data sources, richer feature descriptions can be obtained when processing and analyzing the environment.

[0040] Furthermore, the global feature map of the landscape engineering environment is pooled, thereby compressing the high-dimensional and complex feature map into a lower-dimensional and more representative vector. It should be understood that the pooling operation can effectively reduce the dimension of the feature map. The global feature map of the environment usually contains rich detailed information, but this information can be very large and complex. Through pooling, the local areas in the feature map are aggregated into a simplified representation, thereby significantly reducing the feature dimension. This dimensionality reduction operation can reduce the computational burden, making subsequent processing and analysis more efficient, while reducing storage requirements.

[0041] In the above-mentioned smart garden management system 100 based on the combination of the Internet of Things and GIS technology, the garden plant distribution 3D display generation module 130 is used to generate a 3D display interface for urban garden management plant distribution based on the landscape engineering point cloud semantic feature vector and the landscape engineering environment global feature vector. It should be understood that the point cloud semantic feature vector provides detailed spatial and semantic information about plants and other landscape elements in the garden. Point cloud data captures the precise three-dimensional structure of the environment through laser scanning or three-dimensional sensors, including the spatial distribution, shape and size of plants. The semantic feature vector further parses the specific information in these point cloud data, such as plant species, health status or growth stage. The global feature vector of the environment provides a high-level overview of the overall garden environment. The global feature vector of the environment gathers key features in the environment through operations such as pooling, and describes the overall layout, spatial structure and main landscape elements of the garden. Combining these global features with point cloud data can ensure that the 3D display interface not only displays detailed information of individual plants, but also reflects the structure and style of the entire garden environment. The advantage of integrating the point cloud semantic feature vector and the global feature vector of the environment to generate a 3D display interface is that it provides a full range of environmental understanding. Through this comprehensive display, managers can visually view the distribution of garden plants in three-dimensional space, including plant species, density, health status, and their location and relationship in the overall environment. This visualization tool not only improves data accessibility, but also enhances the accuracy of decision-making.

[0042] Figure 4The block diagram of the 3D display generation module of the garden plant distribution in the smart garden management system based on the combination of the Internet of Things and GIS technology according to the embodiment of the present application is shown in FIG. Figure 4 As shown, in a specific embodiment of the present application, the garden plant distribution 3D display generation module 130 includes: an urban garden feature fusion unit 131, which is used to adjust the fine-grained information of the landscape engineering point cloud semantic feature vector and the landscape engineering environment global feature vector based on the high-dimensional manifold structure expression to obtain the urban garden display generation feature vector; an urban garden management plant distribution 3D display interface generation unit 132, which is used to pass the urban garden display generation feature vector through a generator to generate a 3D display interface for urban garden management plant distribution.

[0043] It should be understood that the fusion of the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment can combine the details of the point cloud data with the overview of the global features. This fusion not only retains the specific information of each plant and landscape element, but also integrates the position and role of these elements in the overall garden design. By combining detailed local data with a global overview, the generated feature vector can more accurately reflect the actual situation of the garden and its design intent.

[0044] In particular, in the technical solution of the present application, the point cloud data is processed by the context encoder and the convolutional neural network to obtain the three-dimensional spatial structure and depth information of the landscape. These features focus on capturing the spatial relationship, structural details and three-dimensional form of the object of the point cloud. In terms of image processing, the landscape environment image is processed by the pixel enhancement generator and the spatial attention mechanism encoder to obtain the global visual features and environmental information of the image. These features usually pay more attention to the overall texture, color distribution and environmental features of the image in the scene. That is, the semantic feature vector of the landscape engineering point cloud usually has high-precision spatial position information and detailed local structure information, while the global feature vector of the landscape engineering environment may contain higher-level abstraction and global information. Due to the differences in dimension and granularity of these features, information mismatch and difficulty in alignment between features may occur when these features are directly fused. This offset may cause the fused feature vector to be insufficient in expressing specific details. Therefore, in the technical solution of the present application, the fine-grained information adjustment based on the high-dimensional manifold structure expression of the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment is performed to obtain the feature vector for urban garden display generation.

[0045] Among them, the fine-grained information adjustment based on the high-dimensional manifold structure expression of the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment is performed to obtain the urban garden display generation feature vector, including: calculating the position difference between the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment to obtain the fine-grained information adjustment differential feature vector; calculating the position point multiplication between the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment to obtain the fine-grained information adjustment position point multiplication feature vector; calculating the position addition between the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment to obtain the fine-grained information adjustment position addition feature vector; adding the fine-grained information adjustment differential feature vector and the fine-grained information adjustment position point multiplication feature vector. The multiplication feature vector and the fine-grained information adjustment position-added feature vector are cascaded and input into a one-dimensional convolution layer and maximum pooling is performed to obtain a fine-grained information adjustment convolutional coding pooling feature vector; the square root of the two norm of the fine-grained information adjustment convolutional coding pooling feature vector is weighted and added with the one norm of the fine-grained information adjustment convolutional coding pooling feature vector to obtain a fine-grained information adjustment factor; the fine-grained information adjustment position point multiplication feature vector and the square root of the length of the landscape engineering point cloud semantic feature vector are added by position to obtain a fine-grained information adjustment intermediate vector; the fine-grained information adjustment factor is used as the first weighting coefficient and the second weighted hyperparameter is used as the second weighting coefficient to calculate the position-weighted sum of the fine-grained information adjustment position-added feature vector and the fine-grained information adjustment intermediate vector to obtain the urban garden display generation feature vector.

[0046] Among them, the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment are adjusted based on the fine-grained information expressed by the high-dimensional manifold structure to obtain the feature vector for urban garden display generation, including:

[0047]

[0048]

[0049]

[0050] Among them, V 1 Represents the semantic feature vector of landscape engineering point cloud, V 2 represents the global feature vector of the landscape engineering environment, represents positional subtraction, ⊙ represents positional multiplication, represents positional addition, conv1D represents a one-dimensional convolutional layer, MaxPool represents maximum pooling, V μRepresents fine-grained information to adjust the convolutional coding pooling feature vector, ||·|| 1 represents a norm, ||·|| 2 represents the bi-norm, [·] represents the cascade, α represents the first weighted hyperparameter, τ represents the fine-grained information adjustment factor, L represents the length of the semantic feature vector of the landscape engineering point cloud, β represents the second weighted hyperparameter, V f Represents the generated feature vector of urban garden display.

[0051] In the technical solution of the present application, in the process of fusing the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment, the feature fine-grained dimension offset between the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment is taken into consideration, which may cause local structural collapse of the fused urban garden display generated feature vector. Based on this, in the technical solution of the present application, the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment are adjusted for fine-grained information based on high-dimensional manifold structure expression.

[0052] Specifically, based on the restricted expression of the foreground-background structure distinction of the high-dimensional manifold of the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment by the amplitude of the semantic difference vector between the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment, the fine-grained correspondence of key features between the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment is modeled, and then the active fine-grained correspondence adjustment propagation is carried out through the back-propagation displacement compensation information between the corresponding eigenvalues ​​of the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment, so as to avoid the imbalance at the feature vector level of the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment in a focusing manner, thereby synchronizing the micro-dimensions of the two feature vectors, ensuring that in the fusion process, the information between the feature vectors can be effectively aligned and combined, thereby avoiding the abnormal accumulation of local structures.

[0053] Furthermore, the feature vector generated by the urban garden display combines the precise spatial information of the point cloud data with the macroscopic overview of the global environmental characteristics, and describes in detail the information such as the type, location, health status, etc. of the plants in the garden, as well as the overall layout of the garden. These feature vectors provide a comprehensive garden data foundation, but they are usually abstract numerical data themselves and do not have direct visualization effects. The role of the generator is to convert these abstract data into a visual three-dimensional interface. The generator uses the data in the feature vector to create a three-dimensional model to display the distribution, morphology and environmental information of the garden plants in a visual way. Through the generator, the information in the feature vector is converted into an intuitive 3D graphic, so that users can clearly see the actual location of each plant in the garden and its relationship with the environment. Through three-dimensional visualization, garden managers can intuitively observe the distribution of plants, analyze the health status of plants and their impact on the overall garden design. This visualization tool can help identify problem areas, optimize planning, and support management tasks such as pest control, thereby improving the maintenance and operation efficiency of the garden. Specifically, the generator may refer to a three-dimensional graphics generation model, such as a deep learning model based on a convolutional neural network (CNN) or a generative adversarial network (GAN). These models can transform the feature vectors of urban gardens into detailed 3D visualization interfaces to display the distribution of plants. Specifically, generative adversarial networks (GANs) can be used to generate high-quality 3D graphics, while deep learning models can process complex spatial data to achieve accurate 3D displays. These generators help transform abstract feature data into intuitive 3D visual effects, facilitating urban garden management and planning.

[0054] In summary, the embodiment of the present application first obtains the point cloud data of the landscape project and the environmental image of the landscape project obtained by GIS, and then uses deep learning technology to perform feature extraction and correlation analysis on the two. Finally, a generator is used to generate a 3D display interface of plant distribution for urban garden management, so as to facilitate an intuitive understanding of the spatial layout and plant distribution of the garden, and thus help to formulate garden management and maintenance strategies.

[0055] As described above, the smart garden management system 100 based on the combination of the Internet of Things and GIS technology according to the embodiment of the present application can be implemented in various terminal devices. In one example, the smart garden management system 100 based on the combination of the Internet of Things and GIS technology can be integrated into the terminal device as a software module and / or a hardware module. For example, the smart garden management system 100 based on the combination of the Internet of Things and GIS technology can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the smart garden management system 100 based on the combination of the Internet of Things and GIS technology can also be one of the many hardware modules of the terminal device.

[0056] Alternatively, in another example, the smart garden management system 100 based on the combination of the Internet of Things and GIS technology and the terminal device may also be separate devices, and the smart garden management system 100 based on the combination of the Internet of Things and GIS technology may be connected to the terminal device via a wired and / or wireless network, and transmit interactive information in accordance with an agreed data format.

[0057] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0058] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0059] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference to a figure in a claim should not be considered as limiting the claim to which it relates.

[0060] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claim can also be implemented by one unit or device through software or hardware. The second and other words are used to indicate names, but not to indicate any particular order.

[0061] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit of the technical solution of the present invention.

Claims

1. A smart garden management system based on the combination of Internet of Things and GIS technology, characterized in that: include: Landscape engineering data acquisition module, used to acquire point cloud data of landscape engineering and environmental images of landscape engineering acquired by GIS; A landscape engineering data extraction module is used to extract a landscape engineering point cloud semantic feature vector and a landscape engineering environment global feature vector from the landscape engineering point cloud data and the landscape engineering environment image acquired by GIS; The garden plant distribution 3D display generation module is used to generate a 3D display interface of urban garden management plant distribution based on the landscape engineering point cloud semantic feature vector and the landscape engineering environment global feature vector.

2. According to claim 1, the intelligent garden management system based on the combination of Internet of Things and GIS technology is characterized in that: The landscape engineering data extraction module comprises: A garden point cloud data semantic feature extraction unit, used for extracting semantic features from the point cloud data of the landscape gardening project to obtain a semantic feature vector of the point cloud of the landscape gardening project; A landscape point cloud data spatial feature extraction unit, used for performing spatial feature extraction on the point cloud data of the landscape engineering to obtain a landscape engineering point cloud depth feature map; A landscape environment image feature extraction unit, used for extracting features from the landscape environment image acquired by GIS to obtain a landscape environment enhanced feature map; The landscape engineering feature aggregation unit is used to perform feature aggregation on the landscape engineering point cloud depth feature map and the landscape engineering environment enhancement feature map to obtain the landscape engineering environment global feature vector.

3. According to claim 2, the intelligent garden management system based on the combination of Internet of Things and GIS technology is characterized in that: The garden point cloud data semantic feature extraction unit comprises: Passing the point cloud data of the landscape engineering through a landscape engineering point cloud context encoder to obtain a plurality of landscape engineering point cloud feature vectors; The plurality of landscape engineering point cloud feature vectors are concatenated into the landscape engineering point cloud semantic feature vector.

4. According to claim 3, the intelligent garden management system based on the combination of Internet of Things and GIS technology is characterized in that: The garden point cloud data spatial feature extraction unit comprises: Performing spatial mapping on the point cloud data of the landscape engineering to obtain a point cloud projection map of the landscape engineering; The landscape engineering point cloud projection map is passed through a landscape engineering point cloud projection convolutional neural network as a feature extractor to obtain the landscape engineering point cloud depth feature map.

5. According to claim 4, the intelligent garden management system based on the combination of Internet of Things and GIS technology is characterized in that: The garden environment image feature extraction unit comprises: The landscape environment image pixel enhancement subunit is used to pass the landscape environment image acquired by GIS through the landscape environment pixel enhancement generator to obtain the landscape environment enhanced image; The landscape environment image spatial attention encoding subunit is used to obtain the landscape engineering environment enhancement feature map by passing the landscape engineering environment enhancement image through a landscape engineering environment enhancement feature encoder based on the spatial attention mechanism.

6. According to claim 5, the intelligent garden management system based on the combination of Internet of Things and GIS technology is characterized in that: The landscape engineering feature aggregation unit includes: Associating the landscape engineering point cloud depth feature map with the landscape engineering environment enhancement feature map to obtain a landscape engineering environment global feature map; The global feature map of the landscape engineering environment is pooled to obtain a global feature vector of the landscape engineering environment.

7. The intelligent garden management system based on the combination of Internet of Things and GIS technology according to claim 6 is characterized in that: The garden plant distribution 3D display generation module includes: The urban garden feature fusion unit is used to adjust the semantic feature vector of the landscape engineering point cloud and the global feature vector of the landscape engineering environment based on the high-dimensional manifold structure expression to obtain the urban garden display generation feature vector; The urban garden management plant distribution 3D display interface generation unit is used to generate the urban garden display generation feature vector through a generator to generate the urban garden management plant distribution 3D display interface.

8. The intelligent garden management system based on the combination of Internet of Things and GIS technology according to claim 7 is characterized in that: The urban garden feature fusion unit includes: Calculating the position difference between the landscape engineering point cloud semantic feature vector and the landscape engineering environment global feature vector to obtain a fine-grained information adjusted differential feature vector; Calculating the position point multiplication between the landscape engineering point cloud semantic feature vector and the landscape engineering environment global feature vector to obtain a fine-grained information adjusted position point multiplication feature vector; Calculating the position-based addition between the landscape engineering point cloud semantic feature vector and the landscape engineering environment global feature vector to obtain a fine-grained information adjusted position-based addition feature vector; The fine-grained information adjustment differential feature vector, the fine-grained information adjustment position point product feature vector and the fine-grained information adjustment position addition feature vector are cascaded and input into a one-dimensional convolution layer and maximum pooling is performed to obtain a fine-grained information adjustment convolutional coding pooling feature vector; The square root value of the two-norm of the fine-grained information adjusted convolutional coding pooling feature vector is weighted and added to the one-norm of the fine-grained information adjusted convolutional coding pooling feature vector to obtain a fine-grained information adjustment factor; Adding the position-based point product feature vector of the fine-grained information adjustment to the square root of the length of the landscape engineering point cloud semantic feature vector according to the position to obtain a fine-grained information adjustment intermediate vector; The fine-grained information adjustment factor is used as the first weighting coefficient, and the second weighted hyperparameter is used as the second weighting coefficient. The position-weighted sum of the fine-grained information adjustment position-added feature vector and the fine-grained information adjustment intermediate vector is calculated to obtain the urban garden display generation feature vector.

Images