An AI algorithm-based urban park vitality calculation method and system

Abstract

The application provides a kind of city park vitality calculation method based on AI algorithm, comprising the following steps: the park is divided into different characteristic monitoring areas, and the monitoring images of the monitoring areas are obtained;The monitoring images are input into the vitality element extraction model, and the vitality element index of the species in the monitoring images is extracted by the vitality element extraction model;Based on the vitality element index, the index score of the species is obtained;The weighted sum of the index scores of the species in the monitoring area is calculated to obtain the basic vitality value of the species in the monitoring area;The weighted sum of the basic vitality values of all species in the monitoring area is calculated to obtain the regional vitality value of the monitoring area;The weighted sum of the regional vitality values of all monitoring areas in the park is calculated to obtain the park vitality value.The application solves the problem that the existing space vitality monitoring method is difficult to realize multi-scene monitoring, the monitored species is single, and the detailed situation of the vitality elements and the vitality details of each item cannot be analyzed.

Description

Technical Field

[0001] This invention belongs to the field of urban planning technology, specifically relating to a method and system for calculating the vitality of urban parks based on AI algorithms. Background Technology

[0002] Urban parks are important venues for public life and social interaction, and their vitality is a significant reflection of urban quality. As urban development shifts from outward expansion to intensive growth, numerous urban renewal and redevelopment plans advocate for spatial diversity, with multi-purpose and mixed-use activity areas gradually replacing traditional single-function zones. Therefore, it is necessary to explore scientific methods for monitoring the vitality of urban parks, based on a deep understanding of their inherent characteristics, in order to accurately identify their current state and provide targeted guidance and technical pathways for creating and enhancing the vitality of urban public spaces.

[0003] As attached Figure 6 As shown, existing technologies typically use residents' GPS activity trajectories and public space traffic statistics to analyze spatial density aggregation results, thus presenting population density in space. This monitoring method primarily relies on mobile phone GPS locations and traffic statistics, resulting in coarse-grained statistical analysis that is difficult to meet the needs of multi-scenario applications. Furthermore, the activity monitoring targets are mainly humans, lacking monitoring of other ecological species, leading to a single monitored species. An example of an existing activity heatmap is attached. Figure 7 As shown, although it reflects the activity levels of different regions, it is difficult to analyze the details of the activity elements in depth and it is impossible to monitor the activity details of each sub-item. Summary of the Invention

[0004] This invention aims to at least solve the technical problems existing in the prior art, and provides a method and system for calculating the vitality of urban parks based on AI algorithms. It solves the problems of existing spatial vitality monitoring methods, which are difficult to monitor multiple scenarios, monitor only one species, and cannot analyze the detailed situation of vitality elements and the vitality details of each sub-item.

[0005] To achieve the above-mentioned objectives of the present invention, according to a first aspect of the present invention, the present invention provides a method for calculating the vitality of urban parks based on AI algorithms, comprising the following steps: dividing the park into monitoring areas with different characteristics, and acquiring monitoring images of the monitoring areas; inputting the monitoring images into a vitality element extraction model, wherein the vitality element extraction model extracts vitality element indicators of species in the monitoring images; obtaining indicator scores of species based on vitality element indicators; calculating the weighted sum of indicator scores of species within the monitoring area to obtain the basic vitality value of the species within the monitoring area; calculating the weighted sum of the basic vitality values ​​of all species within the monitoring area to obtain the regional vitality value of the monitoring area; and calculating the weighted sum of the regional vitality values ​​of all monitoring areas within the park to obtain the park's vitality value.

[0006] Furthermore, species are categorized into humans, animals, and plants. Human vitality indicators include activity types within the monitoring area, the number of people participating in each activity type, activity duration, and actual visitor flow. Human indicator scores within the monitoring area include spatial diversity index, activity duration index, and visitor flow ratio. The calculation process for human indicator scores includes: calculating the spatial diversity index based on activity types and the number of people participating in each activity type within the monitoring area; obtaining the activity duration index based on the activity duration and index conversion table; obtaining the projected visitor flow within the monitoring area; and obtaining the visitor flow ratio within the monitoring area based on the ratio of actual visitor flow to projected visitor flow.

[0007] Furthermore, the animal vitality index includes the actual number of species and the animal activity time within the monitoring area; the animal index score within the monitoring area includes the animal species ratio and the activity time ratio; the calculation process of the animal index score includes: obtaining the ratio of the actual number of species to the initial number of species within the monitoring area as the animal species ratio, and obtaining the ratio of the animal activity time to the monitoring time within the monitoring area as the activity time ratio.

[0008] Furthermore, the plant vitality indexes include the number of exotic plant species, native plant species, pest and disease area, greening area, and total plant area within the monitoring area; the plant index scores within the monitoring area include plant diversity, pest and disease rate, and greening rate; the calculation process for the plant index scores includes: obtaining the ratio of exotic plant species to native plant species within the monitoring area as plant diversity, obtaining the ratio of pest and disease area to total plant area within the monitoring area as pest and disease rate, and obtaining the ratio of greening area to total plant area within the monitoring area as greening rate.

[0009] Furthermore, in the step of calculating the weighted sum of the basic vitality values ​​of all species within the monitoring area to obtain the regional vitality value of the monitoring area, the weight of the basic vitality value of the species is set according to the density and diversity of the species within the monitoring area.

[0010] Furthermore, in the step of calculating the weighted sum of the regional vitality values ​​of all monitored areas within the park to obtain the park's vitality value, the weight of each monitored area is the ratio of the monitored area's area to the total area of ​​the park.

[0011] Furthermore, the monitoring areas include low-density, low-diversity areas, low-density, high-diversity areas, high-density, low-diversity areas, and high-density, high-diversity areas.

[0012] Furthermore, the training process of the vitality element extraction model includes: feeding sample images into the initial neural network model to extract vitality element indicators, calibrating the extracted sample images, labeling the sample images with extraction errors, feeding them into the initial model for training, until the extraction efficiency and extraction accuracy reach the set values, and deriving the trained initial neural network model as the vitality element extraction model.

[0013] To achieve the above-mentioned objectives of the present invention, according to a second aspect of the present invention, the present invention provides an urban park vitality calculation system based on an AI algorithm, which uses any of the above-mentioned urban park vitality calculation methods based on AI algorithms during operation; the system includes an import module, a vitality element extraction module, a storage module, and a calculation module; the import module is used to import monitoring images of the monitoring area; the vitality element extraction module is used to train and store a vitality element extraction model, which is used to extract vitality element indicators from the monitoring images; the storage module is used to store and record the weights of the indicator scores of species, the weights of the basic vitality values ​​of species, and the weights of the monitoring area; the calculation module is used to calculate the basic vitality value of the species, the regional vitality value, and the park vitality value.

[0014] Furthermore, it also includes a monitoring module, a login module, and a display module. The monitoring module is used to capture monitoring images of the monitoring area; the login subsystem is used to verify the user's identity, and after successful verification, it starts the import module and the vitality element extraction module; the display module is used to display the park's vitality value.

[0015] The technical principle of this invention is as follows: This solution obtains the park vitality value through the analysis and index extraction of monitoring images, followed by a progressive calculation. Monitoring images of different monitoring areas are acquired, and the vitality element indicators of species in the monitoring images are extracted. The index scores of the species are calculated based on the vitality element indicators. The basic vitality value of the species is calculated based on the index parameters and their weights. The regional vitality value within the monitoring area is calculated based on the basic vitality value and the species' weights. Finally, the park vitality value is calculated based on the regional vitality value of each monitoring area and the weights of the monitoring areas.

[0016] The beneficial effects of this invention are as follows: This invention calculates park vitality values ​​by acquiring index scores from multiple species, increasing the number of monitoring targets and improving the diversity of park vitality values; it calculates regional vitality values ​​for each monitoring area within the park separately, providing timely feedback on the vitality status of each monitoring area, enabling multi-scenario monitoring of park vitality values, and improving the accuracy and comprehensiveness of park vitality values; this invention sets vitality element indicators for each species, and compared to existing technologies where vitality calculation depends solely on the number of species, the basic vitality value, regional vitality value, and park vitality value obtained by this invention all possess diversity and in-depth analytical value, facilitating urban vitality space management. Compared to existing technologies that rely on mobile phone location positioning and traffic statistics technology, this invention relies on monitoring images and AI recognition technology to acquire species vitality elements, improving the accuracy of vitality element monitoring and reducing the implementation difficulty of vitality value calculation. Attached Figure Description

[0017] Figure 1 This is a logical schematic diagram of an AI algorithm-based method for calculating the vitality of urban parks according to the present invention;

[0018] Figure 2 This is the tool interface of the labelimg software;

[0019] Figure 3 This is a schematic diagram of parameter analysis for the vitality element extraction model of this invention;

[0020] Figure 4 This is a schematic diagram of the training logic of the vitality element extraction model of the present invention;

[0021] Figure 5 This is a schematic diagram of the structure of an urban park vitality calculation system based on AI algorithms according to the present invention;

[0022] Figure 6 It is an existing method for monitoring vitality values;

[0023] Figure 7 It is a space activity heat map obtained from existing technology monitoring. Detailed Implementation

[0024] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0025] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0026] In the description of this invention, unless otherwise specified and limited, it should be noted that the terms "installation", "connection" and "linking" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two components. They can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0027] As attached Figure 1 As shown, this invention provides a method for calculating the vitality of urban parks based on AI algorithms, including the following steps:

[0028] The park was divided into monitoring areas with different characteristics, and monitoring images of the monitoring areas were acquired; the monitoring images included pictures and videos collected in the park.

[0029] The monitoring images are input into the vitality element extraction model, which extracts the vitality element indicators of the species in the monitoring images; the index scores of the species are obtained based on the vitality element indicators.

[0030] The basic vitality value of a species within a monitoring area is obtained by calculating the weighted sum of its index scores. The regional vitality value of a monitoring area is obtained by calculating the weighted sum of the basic vitality values ​​of all species within the monitoring area. The park's overall vitality value is obtained by calculating the weighted sum of the regional vitality values ​​of all monitoring areas within the park.

[0031] Specifically, species are divided into humans, animals, and plants. Human vitality indicators include activity types within the monitoring area, number of people participating in each activity type, activity duration, and actual visitor flow. Human indicator scores within the monitoring area include spatial diversity index, activity duration index, and visitor flow ratio.

[0032] The process of calculating human indicator scores includes: calculating the spatial diversity index based on the activity types and the number of people in each activity type within the monitoring area; obtaining the activity duration index based on the activity duration and index conversion table; obtaining the expected visitor flow in the monitoring area; and obtaining the visitor flow ratio within the monitoring area based on the ratio of actual visitor flow to expected visitor flow.

[0033] Specifically, the spatial diversity index is calculated using the Shannon-Wiener index formula; the Shannon-Wiener index formula is as follows:

[0034] H=-∑|(n i / N)ln(n i / N)|

[0035] Where H represents the spatial diversity index, and n i This represents the number of people in the i-th activity type, and N represents the total number of people in the monitored images, i.e., the total number of people in all activity types.

[0036] Specifically, the index conversion table is shown in Table 1;

[0037]

[0038] Animal vitality indicators include the actual number of species and the time of animal activity within the monitoring area; the indicator scores of animals within the monitoring area include the animal species ratio and the activity time ratio; the calculation process of animal indicator scores includes: obtaining the ratio of the actual number of species to the initial number of species within the monitoring area as the animal species ratio, and obtaining the ratio of the animal activity time to the monitoring time within the monitoring area as the activity time ratio. Preferably, the monitoring time is set to 24 hours, but it can also be set to 12 hours or other values.

[0039] The indicators of plant vitality include the number of exotic plant species, native plant species, pest and disease area, green area, and total plant area within the monitoring area; the indicators of plant performance within the monitoring area include plant diversity, pest and disease rate, and greening rate; the calculation process of plant indicators includes: obtaining the ratio of exotic plant species to native plant species within the monitoring area as plant diversity, obtaining the ratio of pest and disease area to total plant area within the monitoring area as pest and disease rate, and obtaining the ratio of green area to total plant area within the monitoring area as greening rate.

[0040] Preferably, since the vitality element indices of different species are different, the weights of the index scores are all different;

[0041] Specifically, based on data from mass behavior analysis, tourist behavior surveys and analyses, etc., the spatial diversity index represents the activity level of a single activity type within the monitoring area, with a weight of 40%–50%. The activity sustainability index, identified in mass behavior analysis as the attractiveness of the environment to human activities, has a weight of 30%–40%. Since the expected visitor flow data for the monitoring area has been analyzed during the environmental design phase, the ratio of actual visitor flow to expected visitor flow will determine the park's attractiveness, with a weight of 10%–30%. Assuming a 40% weight for the spatial diversity index, a 30% weight for the activity sustainability index, and a 30% weight for visitor flow comparison, then the basic human vitality value = spatial diversity index × 40% + activity sustainability index × 30% + visitor flow comparison × 30%.

[0042] Since the construction of public spaces will inevitably lead to an increase or decrease in animal species, comparing the actual number of species in the monitoring area with the initial number of species at the time of park construction can determine the positive or negative impact of urban public space construction on the surrounding natural environment, and can effectively analyze the life patterns of animal species. The animal species ratio can effectively determine whether humans live in harmony with nature, so the weight of the animal species ratio is set at 50% to 60%. Animal activity time often represents animal vitality in animal behavior analysis, so the weight of the activity time ratio is set at 40% to 50%. Assuming that the weight of the animal species ratio is 50% and the weight of the activity time ratio is 50%, then the basic vitality value of animals = animal species ratio × 50% + activity time ratio × 50%.

[0043] According to relevant plant community surveys and landscape evaluation analyses, plant diversity can be divided into native species and introduced species. Native species should be adapted to various objective factors such as weather and soil, exhibit good growth, strong ornamental value, and possess local characteristics. Introduced species, having undergone long-term domestication, have adapted to local climate and soil conditions, and their ornamental value complements native plants. The ornamental value achieved by combining native and introduced species in landscaping should be fully considered, and the weight of plant diversity should be set at 40%–50%. Data from plant pest and disease studies show that plant infection with pests and diseases will affect the ornamental value of plants, and in severe cases, may even negatively impact their overall appearance. Considering the existing relatively complete mechanisms for plant management, plant diseases and pests can be controlled and prevented, with the disease and pest rate weighted at 30%–40%. Due to differences between the north and south, plant species vary, generally divided into evergreen and deciduous plants. According to relevant books on landscape design, urban public space design already considers green area and plant species combinations, with the greening rate weighted at 10%–30%. Assuming plant diversity has a weight of 40%, disease and pest rate has a weight of 30%, and greening rate has a weight of 30%, then the basic vitality value of a plant = plant diversity × 40% + disease and pest rate × 30% + greening rate × 30%.

[0044] In the step of calculating the weighted sum of the basic vitality values ​​of all species within the monitoring area to obtain the regional vitality value of that monitoring area, the weight of the basic vitality value of each species is set according to the density and diversity of that species within the monitoring area. People of different ages show interest in different activity areas within urban parks. For example, children's main activity areas are children's playgrounds and interactive smart facilities; the main activity areas for the elderly are fitness equipment. Therefore, the monitoring areas of urban parks will have different vitality values ​​depending on the age of visitors. Thus, different monitoring areas and human needs will lead to diversity in human behavior; for example, common human behaviors in lawn areas include family play, sightseeing, sports and leisure, and cultural performances. The density of plants and animals in the monitoring area can directly affect the diversity of human behavior.

[0045] Specifically, low-density, low-diversity areas are those with abundant vegetation. Due to the large number of plants and the relatively small area for human activities, it is considered to focus on monitoring and analyzing the vitality index of plants and animals. At the same time, park construction and weather changes will also directly affect human activities, resulting in low human density and low diversity of human behavior. Let the weight of the basic vitality value of plants in low-density, low-diversity areas be 50%, the weight of the basic vitality value of animals be 30%, and the weight of the basic vitality value of humans be 20%. Then, the regional vitality value of low-density, low-diversity areas = basic vitality value of plants × 50% + basic vitality value of animals × 30% + basic vitality value of humans × 20%.

[0046] The low-density, high-diversity area is a waterfront area. The plants along the water can provide shade for humans, and humans can carry out some activities along the waterfront. However, due to the relatively narrow activity area, the human density is low. Let the weight of the basic vitality value of plants in the low-density, high-diversity area be 40%, the weight of the basic vitality value of animals be 30%, and the weight of the basic vitality value of humans be 30%. Then, the regional vitality value of the low-density, high-diversity area = the basic vitality value of plants × 40% + the basic vitality value of animals × 30% + the basic vitality value of humans × 30%.

[0047] High-density, low-diversity areas are sports areas in urban parks and are one of the main human activity areas. While the plant coverage and animal activity areas are relatively small, the human density is high due to the relatively singular function of these activity areas. Assuming the weight of the basic plant vitality value for high-density, low-diversity areas is 15%, the weight of the basic animal vitality value is 15%, and the weight of the basic human vitality value is 70%, then the area vitality value of high-density, low-diversity areas = basic plant vitality value × 15% + basic animal vitality value × 15% + basic human vitality value × 70%.

[0048] High-density, high-diversity areas are lawn areas in urban parks, which are one of the main areas of human activity. The vegetation is mainly ground cover, with a high degree of homogeneity. Animal activity areas are mainly concentrated at high altitudes, while the area of ​​human activity is large, which can lead to increased diversity of human behavior. Human vitality can be a key area for monitoring. Let the weight of the basic vitality value of plants in high-density, high-diversity areas be 20%, the weight of the basic vitality value of animals be 10%, and the weight of the basic vitality value of humans be 70%. Then the regional vitality value of high-density, high-diversity areas = basic vitality value of plants × 20% + basic vitality value of animals × 10% + basic vitality value of humans × 70%.

[0049] Preferably, the monitoring area includes low-density low-diversity areas, low-density high-diversity areas, high-density low-diversity areas, and high-density high-diversity areas;

[0050] Preferably, in the step of calculating the weighted sum of the regional vitality values ​​of all monitored areas within the park to obtain the park vitality value, the weight of each monitored area is the ratio of its area to the total park area. For example, if low-density diversity areas account for 45% of the total park area, low-density high-diversity areas account for 15%, high-density low-diversity areas account for 10%, and high-density high-diversity areas account for 30%, then the park vitality value = (regional vitality value of low-density diversity areas × 45%) + (regional vitality value of low-density high-diversity areas × 15%) + (regional vitality value of high-density low-diversity areas × 10%) + (regional vitality value of high-density high-diversity areas × 30%).

[0051] As attached Figure 4As shown, the training process of the vitality element extraction model includes: putting sample images into the initial neural network model; the initial neural network model determining whether the sample images contain vitality element indices; placing sample images containing vitality element indices and sample images not containing vitality element indices in different folders; extracting vitality element indices from images containing vitality element indices; calibrating the extracted sample images according to the manual calibration menu; labeling and storing incorrectly extracted sample images in the corresponding folders; simultaneously putting incorrectly extracted sample images into the initial model for training until the extraction accuracy and extraction efficiency reach the set values; and exporting the trained initial neural network model as the vitality element extraction model.

[0052] Specifically, image annotation tools are used to identify and annotate target objects in sample images; preferably, the image annotation tool is labelimg software, as shown in the attached image annotation tool. Figure 2 As shown, the software can quickly identify and label the size and location of multiple target objects and directly generate XML files. The labeled sample images are divided into training, validation, and test sets in an 8:1:1 ratio. The training set is used to train the parameters of the initial neural network model; the validation set is used to verify the performance of the initial neural network model and select the optimal hyperparameters; the test set is used to evaluate a set of various metrics that objectively assess the initial neural network model during training.

[0053] Specifically, this approach uses PyTorch as the training framework to train the initial neural network model. PyTorch supports distributed and parallel training. The labeled sample images are then fed into the initial neural network model.

[0054] Preferably, the initial neural network model consists of three parts: Backbone, Neck, and Head. The Backbone refers to the feature extraction network, which extracts information about target objects from sample images. To detect the location and category of target objects from sample images, necessary feature information, such as HOG features, is first extracted from the sample images. This necessary feature information is then used to locate and classify the sample images. Commonly used Backbones include: VGG, ResNet (ResNet18, 50, 100), ResNeXt, DenseNet, SqueezeNet, Darknet (Darknet19, 53), DetNet, DetNASSpineNet, EfficientNet (EfficientNet-B0 / B7), CSPResNeXt50, and CSPDarknet53. In this scheme, to improve the robustness of the initial neural network model, CSPDarknet53 was selected as the backbone network. The network structure parameters were optimized to improve the overall feature extraction accuracy. The Head is the detection head, mainly used to predict the type and location (bounding boxes) of target objects in the sample image, making predictions using necessary feature information. The Neck module is a crucial link, fusing the necessary feature information extracted by the Backbone to make the network learn more diverse, before handing it over to the subsequent Head for detection, thereby improving the performance of the initial neural network model.

[0055] Preferably, in order to improve the training accuracy of the initial neural network model, a variety of data processing methods are used to process the sample images. These data processing methods include Mosaic, Cutout, image perturbation, changing the brightness, contrast, saturation, and hue of the sample images, adding noise, random scaling, random cropping, flipping, rotating, and random erasing. In order to improve the training speed of the initial neural network model and reduce the training time, a rectangular training method is also added during the data processing.

[0056] Preferably, in order to improve the training efficiency of the initial neural network model, multiple GPUs are used for parallel training according to the hardware situation; specifically, the initial neural network model is loaded into the memory of multiple GPUs, and the sample images and image annotation tools are loaded into the system cache of the GPUs, which increases the training time by a factor of the number of GPUs.

[0057] Preferably, training automatically stops when the set epoch is reached, and the vitality element extraction model is obtained and exported. (See attached diagram) Figure 3As shown, removing unnecessary training parameters from the vitality element extraction model reduces the model size by 60% and improves inference efficiency by 20%.

[0058] Preferably, the trained vitality element extraction model is used for inference calculations. Specifically, the trained vitality element extraction model is exported, loaded into the content through an inference program, and the sample images are resized to a uniform scale before inference. The vitality element extraction model and the input sample images support fp16 half-precision inference. According to calculations, on an RTX4000 device, the average inference time per image for the vitality element extraction model is about 40ms, and the overall processing capability is about 25 frames per second. At the same time, the vitality element extraction model supports quantization and TensorRT deployment, which can significantly improve processing efficiency.

[0059] As attached Figure 5 As shown, the present invention also provides an urban park vitality calculation system based on AI algorithms, which uses an urban park vitality calculation method based on AI algorithms as described in any one of claims 1-8 during operation; the system includes a monitoring module, a login module, an import module, a vitality element extraction module, a storage module, a calculation module, and a display module;

[0060] The monitoring module is used to capture monitoring images of the monitored area;

[0061] The login subsystem is used to verify the user's identity. After successful verification, the import module and the vitality element extraction module are started.

[0062] The import module is used to import monitoring images of the monitoring area;

[0063] The vitality element extraction module is used to train and store the vitality element extraction model, which is used to extract vitality element indices from monitoring images.

[0064] The storage module is used to store and record the weights of species index scores, the weights of species basic vitality values, and the weights of monitoring areas;

[0065] The calculation module is used to calculate the basic vitality value of species, the regional vitality value, and the park vitality value;

[0066] The display module is used to show the park's vitality value.

[0067] Preferably, the system is a visualization system based on a B / S architecture. During the initial training process of the neural network model, the system allows users to view the running parameters during training in real time through a browser, such as average mAP and loss, and also allows them to view the validation process of the test set.

[0068] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0069] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for calculating the vitality of urban parks based on AI algorithms, characterized in that, Includes the following steps: The park was divided into monitoring areas with different characteristics, and monitoring images of the monitoring areas were obtained; The monitoring images are input into the vitality element extraction model, which extracts the vitality element indicators of the species in the monitoring images. The training process of the vitality element extraction model includes: feeding sample images into an initial neural network model for vitality element indicator extraction; calibrating the extracted sample images; labeling erroneous sample images and feeding them into the initial model for training until the extraction efficiency and accuracy reach the set values; and exporting the trained initial neural network model as the vitality element extraction model. The initial neural network model consists of three parts: a backbone network, a Neck module, and a Head detection head. The species are divided into humans, animals, and plants. The indicator scores of the species are obtained based on the vitality element indicators. The human vitality element indicators include the activity type, the number of people in a single activity type, the activity duration, and the actual visitor flow within the monitoring area. The human indicator scores within the monitoring area include the spatial diversity index, the activity duration index, and the visitor flow ratio. The calculation process of the human indicator scores includes: calculating the spatial diversity index based on the activity type and the number of people in a single activity type within the monitoring area; obtaining the activity duration index based on the activity duration and the index conversion table; obtaining the expected visitor flow within the monitoring area; and obtaining the visitor flow ratio within the monitoring area based on the ratio of the actual visitor flow to the expected visitor flow. Animal vitality indicators include the actual number of species and the time spent on animal activity within the monitoring area; the indicator scores of animals within the monitoring area include the animal species ratio and the activity time ratio; the calculation process of animal indicator scores includes: obtaining the ratio of the actual number of species to the initial number of species within the monitoring area as the animal species ratio, and obtaining the ratio of the animal activity time to the monitoring time within the monitoring area as the activity time ratio. The indicators of plant vitality include the number of exotic plant species, native plant species, pest and disease area, green area, and total plant area within the monitoring area; the indicators of plant in the monitoring area include plant diversity, pest and disease rate, and greening rate; the calculation process of plant indicator scores includes: obtaining the ratio of exotic plant species to native plant species within the monitoring area as plant diversity, obtaining the ratio of pest and disease area to total plant area within the monitoring area as pest and disease rate, and obtaining the ratio of green area to total plant area within the monitoring area as greening rate. The basic vitality value of a species within the monitoring area is obtained by calculating the weighted sum of its index scores. The regional vitality value of the monitoring area is obtained by calculating the weighted sum of the basic vitality values ​​of all species within the monitoring area, wherein the weight of the basic vitality value of a species is set according to the density and diversity of that species within the monitoring area; The park's vitality value is obtained by calculating the weighted sum of the regional vitality values ​​of all monitored areas within the park, where the weight of each monitored area is the ratio of its area to the total area of ​​the park.

2. The method for calculating urban park vitality based on AI algorithm as described in claim 1, characterized in that, The monitoring areas include low-density, low-diversity areas, low-density, high-diversity areas, high-density, low-diversity areas, and high-density, high-diversity areas.

3. A city park vitality calculation system based on AI algorithms, characterized in that, The system utilizes an AI-based urban park vitality calculation method as described in any of claims 1-2 during operation; the system includes an import module, a vitality element extraction module, a storage module, and a calculation module. The import module is used to import monitoring images of the monitored area; The vitality element extraction module is used to train and store the vitality element extraction model, which is used to extract vitality element indicators from monitoring images. The storage module is used to store and record the weights of species index scores, the weights of species basic vitality values, and the weights of monitoring areas; The calculation module is used to calculate the basic vitality value of species, the regional vitality value, and the park vitality value.

4. The urban park vitality calculation system based on AI algorithm as described in claim 3, characterized in that, It also includes a monitoring module, a login module, and a display module. The monitoring module is used to capture monitoring images of the monitored area; the login subsystem is used to verify the user's identity, and after successful verification, it starts the import module and the vitality element extraction module; the display module is used to display the park's vitality value.

Images