Urban greening planning method and system based on object recognition

Abstract

The invention relates to the technical field of urban planning, and discloses an urban landscaping planning method and system based on object recognition, and the system comprises a regional analysis system, an object recognition system and a planning system. The regional analysis system comprises a defense film analysis module used for collecting defense film data and a questionnaire analysis module used for setting a demand questionnaire and collecting filling data of the demand questionnaire; the object identification system comprises a structure identification module, a project identification module and a demand identification module. The structure identification module is used for identifying a green land association structure in the target area; the project identification module is used for importing a planning file and identifying a planning index in the planning file; the demand identification module is used for counting and analyzing greening demands of residents and obtaining resident intention indexes; the planning system includes a zoning module for delineating an optimization region and a staging module for generating planning suggestions. According to the method, urban greening projects can be reasonably planned, and perfect and reliable data reference can be provided for greening resource allocation.

Description

Technical Field

[0001] The present invention relates to the technical field of urban planning, and in particular to an urban greening planning method and system based on object recognition. Background Art

[0002] Urban greening is not only an important means of beautifying the urban environment, but also a key measure to improve urban ecology and enhance the quality of life of residents. Reasonable urban greening planning and implementation can effectively alleviate the urban heat island effect, purify the air, regulate the climate, play an important role in achieving the "dual carbon" process, and provide urban residents with green space for leisure and entertainment. The rationality of urban greening is reflected in the scientific planning and layout. Urban greening should not only be scattered green space distribution, but should form a systematic network.

[0003] In urban greening construction, it often takes several years of planning, layout and implementation. It is necessary to combine the distribution of green landscape resources and the needs of urban residents, and plan based on different green indicators to consider various possibilities for green space layout or optimization schemes. In traditional design, this work needs to be completed by planning and design personnel, which is costly and greatly affected by subjective experience, and lacks objectivity. With the development of new generation information technology, a few new schemes have proposed model-based intelligent planning schemes, and the automatic generation of green space layout schemes has been preliminarily realized. For example, the patent document with publication number CN116401736A discloses an automatic layout method of urban green space based on artificial intelligence, which combines classification algorithms, pathfinding algorithms, digital translation and other technologies to layout urban green space, and can quickly generate planning schemes. However, most of these existing schemes focus on the analysis and planning of the spatial elements of a piece of green space itself, but in the actual implementation of urban greening planning, there are still many factors that are not considered by the existing schemes.

[0004] The actual urban greening planning usually involves work at multiple levels. An overall green space construction goal often requires the support of multiple urban greening projects - from greenway construction, park construction, greenway maintenance, park renewal, to water body management, terrain transformation, etc. Each project has its own uniqueness and needs to be customized according to the specific conditions of the region. The implementation of urban greening projects requires a lot of resources, including financial funds, human resources and technical resources. If all projects are started at the same time, it may lead to excessive dispersion of resources, which in turn affects the quality and progress of each project. For example, the "Implementation Plan for Urban Gardening and Greening in Zhengzhou in 2024" mentioned that the annual new green space area will reach more than 3 million square meters, and it is planned to build or renovate multiple parks and greenways. Such a huge amount of engineering, if reasonable resource allocation is not carried out, is likely to lead to waste of resources or project delays. This brings new planning difficulties to urban green space planning. Summary of the invention

[0005] The present invention aims to provide an urban greening planning method and system based on object recognition, which can reasonably plan urban greening projects and provide complete and reliable data reference for greening resource allocation.

[0006] To achieve the above objectives, the present invention provides the following basic solutions.

[0007] Solution 1

[0008] An urban greening planning system based on object recognition, comprising a regional analysis system, an object recognition system and a planning system;

[0009] The regional analysis system includes a satellite image analysis module and a questionnaire analysis module; the satellite image analysis module is used to collect satellite image data and pre-process the satellite image data; the questionnaire analysis module is used to set a demand questionnaire and collect the filling data of the demand questionnaire; the demand questionnaire is used to collect residents' greening needs;

[0010] The object recognition system includes a structure recognition module, a project recognition module and a demand recognition module; the structure recognition module is used to identify the green space associated structure in the target area according to the pre-processed satellite image data; the green space associated structure includes the existing green line, the pedestrian system structure, the park green space structure, the protective green space structure, the three-dimensional greening structure, and the idle green space structure;

[0011] The project identification module is used to import the planning file, identify the planning indicators in the planning file, and arrange them in chronological order; the demand identification module is used to statistically analyze the greening needs of residents and obtain the residents' intention indicators;

[0012] The planning system includes a zoning module and a staging module; the zoning module is used to match the residents' intention indicators, planning indicators and the green space association structure of the target area, and delineate the optimization area;

[0013] The staging module is used to calculate the implementation order of each optimized area and generate planning suggestions.

[0014] Solution 2

[0015] An urban greening planning method based on object recognition, using an urban greening planning system based on object recognition as described in Scheme 1 to carry out urban greening planning; comprising the following steps:

[0016] The satellite image analysis module of the regional analysis system collects satellite image data and pre-processes the satellite image data; the questionnaire analysis module of the regional analysis system sets a demand questionnaire and collects the filling data of the demand questionnaire; the demand questionnaire is used to collect the greening needs of residents;

[0017] The structure recognition module of the object recognition system identifies the green space associated structure in the target area based on the pre-processed satellite image data; the green space associated structure includes the existing green line, the pedestrian system structure, the park green space structure, the protective green space structure, the three-dimensional greening structure, and the idle green space structure;

[0018] The project identification module of the object identification system imports the planning file, identifies the planning indicators in the planning file, and arranges them in chronological order; the demand identification module of the object identification system statistically analyzes the greening needs of residents and obtains the residents' intention indicators;

[0019] The zoning module of the planning system matches the residents' intention indicators, planning indicators and the green space correlation structure of the target area, and delineates the optimized area; the phasing module of the planning system calculates the implementation order of each optimized area and generates planning suggestions.

[0020] The working principle and advantages of the present invention are:

[0021] This scheme carries out multi-dimensional object identification. For the urban target areas to be planned, the scheme extracts relevant indicators / characteristics of the urban target areas from three dimensions: the green space association structure of the urban target areas themselves, the planning documents to be executed, and the actual intentions of the residents in the urban target areas. This can provide rich and comprehensive reference data for the project arrangement of urban greening planning. Based on these relevant indicators / characteristics, the optimized areas in the urban target areas are comprehensively selected, and their implementation order is quantified. Compared with the existing planning schemes that focus on the spatial element analysis and planning of a piece of green space (a green space project) itself, this scheme proposes a new planning perspective. From a more macro perspective, the scheme coordinates the necessity and implementation order of the overall urban green space project, which can reasonably plan urban greening projects and provide complete and reliable data reference for the allocation of greening resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 The present invention is a schematic diagram of the structure of an urban greening planning method and system embodiment based on object recognition. DETAILED DESCRIPTION

[0023] The following is a further detailed description through specific implementation methods:

[0024] The embodiment is basically as shown in the attached Figure 1 Shown: An urban greening planning system based on object recognition, including a regional analysis system, an object recognition system and a planning system.

[0025] The regional analysis system includes a satellite image analysis module and a questionnaire analysis module.

[0026] The satellite image analysis module is used to collect satellite image data and pre-process the satellite image data. Specifically, the satellite image data refers to satellite image data, and in this embodiment, specifically refers to satellite image data of the target city area to be analyzed.

[0027] The preprocessing includes delineating green areas in the satellite image data and dividing the green areas into planned areas and unplanned areas. When dividing the green areas, the collaborative division is performed based on GIS (geographic information system) technology and RS (remote sensing interpretation) technology, including the following steps:

[0028] The satellite image data is processed, including geometric correction. The existing remote sensing technology is then used to analyze the color, texture, shape and other characteristics of the satellite image data to identify areas covered by green vegetation.

[0029] Image segmentation technology is then applied to divide the image into different areas, and then green areas are distinguished according to vegetation indices (such as NDVI, Normalized Difference Vegetation Index) to obtain the first-level classification layer.

[0030] Then import the first-level classification layer into the GIS system, and perform spatial overlay analysis with other spatial data (such as land use planning maps, infrastructure distribution maps, etc.) to determine the location of the green space and its relationship with the surrounding environment. In the GIS environment, attribute information (such as area, vegetation type, etc.) can be added to each plot, and the second-level classification layer can be obtained. Then perform spatial overlay analysis with the historical greening planning map (this part of the data can be obtained from the public website or database of the urban planning department) in the GIS system, and then add attribute information (such as planned, unplanned, etc.) to each plot and obtain the third-level classification layer. Then divide it into planned areas and unplanned areas.

[0031] This setting can efficiently analyze the land layout of the target urban area and provide more reference information for subsequent object recognition. In addition, when the analysis dimension needs to be expanded, attributes can be quickly added to each plot by overlaying layers, which is more convenient.

[0032] The questionnaire analysis module is used to set demand questionnaires and collect the data filled in the demand questionnaires; the demand questionnaires are used to collect residents' greening needs. The questionnaire analysis module is preset with a questionnaire template; the questionnaire template includes a question field, an option field and other fields; the other fields are used for residents to input fields independently.

[0033] Specifically, in this embodiment, the regional analysis system also includes a questionnaire optimization module; the questionnaire optimization module is used to adjust the questionnaire template; in the questionnaire template, the question field and the option field both include fixed fields and fillable fields; wherein, the fixed field or the fillable field may be none, that is, set to a fixed question, a fixed option or a fully filled option; the questionnaire optimization module extracts corpus based on the planning indicators and the green space association structure types existing in the target area to fill in the fillable fields.

[0034] With this setting, this program can make personalized adjustments to the questionnaire template according to the actual green space conditions and planning indicator requirements, and can collect residents' needs in a more targeted manner.

[0035] Take a questionnaire template for Yuzhong District, Chongqing as an example, which contains multiple question fields, each of which corresponds to a set of option fields and a set of other fields. For example, if the question field is "your main activity area", "your main activity area" is a fixed field, and the fillable field is none; the corresponding option fields include "A. Chaotianmen Street", "B. Jiefangbei Street", "C. Nanjimen Street", "D. Qixinggang Street", etc., and the corresponding other fields are "other ___". Among them, "A.", "B.", etc. are fixed fields, and "Chaotianmen Street", "Jiefangbei Street", etc. are fillable fields. For example, if the question field is "Which urban greening promotion measures do you think are more urgent", "Which ____ do you think are more urgent" is a fixed field, and the fillable field is "urban greening promotion measures"; the corresponding option fields include "A. Greening and beautification of slopes and cliffs", "B. Greening and beautification of public building roofs", "C. Increase of tree-lined roads", etc. Among them, the fixed fields are "A.", "B.", etc., and the fillable fields are "Greening and beautification of slopes and cliffs".

[0036] The object recognition system includes a structure recognition module, an item recognition module and a requirement recognition module.

[0037] The structure recognition module is used to identify the green space associated structure in the target area according to the pre-processed satellite image data.

[0038] The green space association structure includes existing green lines, pedestrian system structure, park green space structure, protective green space structure, three-dimensional greening structure, and idle green space structure.

[0039] When identifying the green space association structure, based on the structural attributes of different green space association structures, the attributes are matched with the pre-processed satellite image data (specifically the third-level classification layer) to complete the structure identification. In this embodiment, the structural attribute feature extraction and matching are performed by the convolutional neural network model. In addition, the convolutional neural network model is pre-trained with a green space data set, which includes satellite image data of a variety of existing green space association structures and their structural attribute data; during the model training process, the satellite image data and its structural attribute data are used as input, and the green space association structure is used as output. Among them, the structural attributes of the green space association structure include physical characteristics (including area, shape, location, surrounding land use type, connectivity), vegetation characteristics (including vegetation type, vegetation coverage) and ecological functions (including carbon sink indicators), etc.

[0040] The project identification module is used to import planning documents, identify planning indicators in the planning documents, and arrange them in chronological order; specifically, the planning indicators include planning achievement time, planning limited area and planning limited numerical indicators (such as green space coverage rate, carbon sequestration, walkability, number of people served by unit green space, etc.).

[0041] In this embodiment, a preset recognition model is used to identify planning indicators. Specifically, the preset recognition model is a pre-trained NLP model. When identifying planning indicators, the preset recognition model first parses the imported planning file and converts it into text data; uses the Spacy library for text preprocessing and information extraction; defines a keyword list as planning achievement time, planning limited area, and planning limited numerical indicators; extracts relevant information based on the keyword list and stores it in the indicator list; and extracts the time string corresponding to each information (here is the planning achievement time), and uses a sorting algorithm to sort them in the order of planning achievement time.

[0042] The imported planning documents may be greening construction project plans proposed by relevant departments, such as "Yuzhong District Green Space System Planning", "Yuzhong District Pedestrian System Special Planning", etc. Optionally, planning indicators may also be directly input and set.

[0043] The demand identification module is used to statistically analyze residents' greening needs and obtain residents' intention indicators.

[0044] Specifically, the demand identification module uses statistical methods to analyze the data reported in the demand questionnaire and extracts the resident intention index from it. The resident intention index includes option fields and corresponding proportions.

[0045] The planning system includes a partitioning module, a staging module and a visualization module.

[0046] The zoning module is used to match the residents' intention indicators, planning indicators and the green space association structure of the target area, and to delineate the optimized area. The staging module is used to calculate the implementation order of each optimized area and generate planning suggestions.

[0047] Specifically, when the partitioning module delineates the optimization area, the following steps are included:

[0048] According to the planning-limited area in the planning indicators, multiple sub-areas are delineated in the target urban area as the first-level regional scope; when delineating the first-level regional scope, non-green areas in the target urban area are also pre-excluded to effectively streamline the analysis area. In this embodiment, the non-green area refers to an area that does not have greening optimization conditions or is not suitable for greening optimization, including infrastructure land (such as highways, main roads, subway stations, substations, etc.), densely built areas (such as factories, commercial centers, residential buildings, etc.), and special function areas (such as hospitals, schools, military facilities, etc.).

[0049] According to the residents' intention indicators, multiple sub-areas are identified within the first-level area as the second-level area;

[0050] And according to the proportion of residents' intention indicators, priorities are set for multiple sub-areas in the second-level area. The sub-area with a larger proportion and more residents' intention indicators that meet the requirements has a higher priority.

[0051] In this embodiment, the sub-regions in the second-level region range whose priorities are in the top a% are selected as the optimization regions.

[0052] Wherein, a reference model is also provided in the partition module; the reference model is used to assist in determining the value of a.

[0053] Specifically, the reference model is a BP neural network model. Collect environmental data of each sub-region in the second-level regional scope (in this embodiment, the environmental data is extracted through the third-level classification layer), and use the reference model to predict the greening numerical index that can be achieved by each sub-region in the second-level regional scope (this index corresponds to the planning-limited numerical index, such as when the planning-limited numerical index is the green space coverage rate, the corresponding greening numerical index includes the total green space area and per capita green space area of ​​the sub-region). Then, the greening numerical indicators of multiple sub-regions are gradually superimposed in order of priority until the planning-limited numerical index is exceeded, and the a value at this time is taken as the a value used when extracting the optimized area.

[0054] The reference model is pre-trained using historical greening project implementation data, where the historical greening project implementation data includes environmental parameters of the construction area corresponding to the greening project and the greening numerical indicators ultimately achieved.

[0055] With this setting, this plan can select optimized areas from multiple dimensions to ensure that the selected optimized areas can balance the objective planning needs and the subjective intentions of residents.

[0056] The staging module includes the following steps when calculating the implementation order of each optimized area:

[0057] Step 1: According to the planning indicators matched by each optimization area, the planning indicators are sorted according to the planning achievement time in the planning indicators to obtain the first implementation order of each optimization area;

[0058] Step 2, collecting environmental parameters at each optimized area, and calculating the construction time of each optimized area; and sorting by the length of the construction time to obtain the second implementation order of each optimized area;

[0059] Specifically, a prediction model is provided in the phase module; the prediction model is used to calculate the construction duration of each optimized area. The prediction model adopts an LSTM model and uses a specific historical data set for model training; the specific historical data set consists of historical greening project implementation data, including environmental parameters and construction duration of the construction area corresponding to the greening project. Environmental parameters are used as model input, and construction duration is used as model output.

[0060] Step 3, setting reference weights for the first implementation order and the second implementation order, and calculating the third implementation order as the final implementation order according to the weights.

[0061] This setting can provide reliable regional and time sequence references for the planning of urban greening projects and achieve reasonable project resource allocation.

[0062] The visualization module is used to generate a visualized project list. The project list is sorted by optimizing the area and by the implementation order. The content of the project list includes the optimization area, the implementation order corresponding to the optimization area, planning indicators, etc.

[0063] An urban greening planning method based on object recognition, using the above-mentioned urban greening planning system based on object recognition to carry out urban greening planning; comprising the following steps:

[0064] The satellite image analysis module of the regional analysis system collects satellite image data and pre-processes the satellite image data; the questionnaire analysis module of the regional analysis system sets a demand questionnaire and collects the filling data of the demand questionnaire; the demand questionnaire is used to collect the greening needs of residents;

[0065] The structure recognition module of the object recognition system identifies the green space associated structure in the target area based on the pre-processed satellite image data; the green space associated structure includes the existing green line, the pedestrian system structure, the park green space structure, the protective green space structure, the three-dimensional greening structure, and the idle green space structure;

[0066] The project identification module of the object identification system imports the planning file, identifies the planning indicators in the planning file, and arranges them in chronological order; the demand identification module of the object identification system statistically analyzes the greening needs of residents and obtains the residents' intention indicators;

[0067] The zoning module of the planning system matches the residents' intention indicators, planning indicators and the green space correlation structure of the target area, and delineates the optimized area; the phasing module of the planning system calculates the implementation order of each optimized area and generates planning suggestions.

[0068] The present embodiment provides an urban greening planning method and system based on object recognition, which can reasonably plan urban greening projects, provide complete and reliable data reference for greening resource allocation, and reduce planning costs.

[0069] The above is only an embodiment of the present invention. The common sense such as the known specific structure and characteristics in the scheme is not described in detail here. The ordinary technicians in the relevant field are aware of all the common technical knowledge in the technical field of the invention before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply the conventional experimental means before that date. The ordinary technicians in the relevant field can improve and implement the scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for the ordinary technicians in the relevant field to implement this application. It should be pointed out that for the technicians in this field, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent.

Claims

1. An urban greening planning system based on object recognition, characterized in that: Includes regional analysis system, object recognition system and planning system; The regional analysis system includes a satellite image analysis module and a questionnaire analysis module; the satellite image analysis module is used to collect satellite image data and pre-process the satellite image data; the questionnaire analysis module is used to set a demand questionnaire and collect the filling data of the demand questionnaire; the demand questionnaire is used to collect residents' greening needs; The object recognition system includes a structure recognition module, a project recognition module and a demand recognition module; the structure recognition module is used to identify the green space associated structure in the target area according to the pre-processed satellite image data; the green space associated structure includes the existing green line, the pedestrian system structure, the park green space structure, the protective green space structure, the three-dimensional greening structure, and the idle green space structure; The project identification module is used to import the planning document, identify the planning indicators in the planning document, and arrange them in chronological order; The demand identification module is used to statistically analyze residents' greening needs and obtain residents' intention indicators; The planning system includes a zoning module and a staging module; The zoning module is used to match the residents' intention indicators, planning indicators and the green space association structure of the target area, and to delineate the optimization area; The staging module is used to calculate the implementation order of each optimized area and generate planning suggestions.

2. The urban greening planning system based on object recognition according to claim 1 is characterized in that: The preprocessing includes delineating green areas in the satellite image data and dividing the green areas into planned areas and unplanned areas.

3. The urban greening planning system based on object recognition according to claim 1 is characterized in that: A questionnaire template is preset in the questionnaire analysis module; the questionnaire template includes a question field, an option field and other fields; the other fields are used for residents to input fields independently.

4. The urban greening planning system based on object recognition according to claim 3 is characterized in that: The regional analysis system also includes a questionnaire optimization module; the questionnaire optimization module is used to adjust the questionnaire template; in the questionnaire template, the question field and the option field both include fixed fields and fillable fields; the questionnaire optimization module extracts corpus based on the planning indicators and the green space association structure types existing in the target area to fill in the fillable fields.

5. The urban greening planning system based on object recognition according to claim 3 is characterized in that: The planning indicators include the planning achievement time, the planning restricted area and the planning restricted numerical indicators; the residents' intention indicators include option fields and corresponding proportions.

6. The urban greening planning system based on object recognition according to claim 5 is characterized in that: The staging module includes the following steps when calculating the implementation order of each optimized area: Step 1: According to the planning indicators matched by each optimization area, the planning indicators are sorted according to the planning achievement time in the planning indicators to obtain the first implementation order of each optimization area; Step 2, collecting environmental parameters at each optimized area, and calculating the construction time of each optimized area; and sorting by the length of the construction time to obtain the second implementation order of each optimized area; Step 3, setting reference weights for the first implementation order and the second implementation order, and calculating the third implementation order as the final implementation order according to the weights.

7. The urban greening planning system based on object recognition according to claim 1 is characterized in that: The planning system also includes a visualization module; the visualization module is used to generate a visualized project list. The project list is sorted by optimizing the area and by the implementation order. The content of the project list includes the optimization area, the implementation order corresponding to the optimization area, planning indicators, etc.

8. An urban greening planning method based on object recognition, characterized in that: Applying an urban greening planning system based on object recognition as described in any one of claims 1 to 7 to carry out urban greening planning; comprising the following steps: The satellite image analysis module of the regional analysis system collects satellite image data and pre-processes the satellite image data; the questionnaire analysis module of the regional analysis system sets a demand questionnaire and collects the filling data of the demand questionnaire; the demand questionnaire is used to collect the greening needs of residents; The structure recognition module of the object recognition system identifies the green space associated structure in the target area based on the pre-processed satellite image data; the green space associated structure includes the existing green line, the pedestrian system structure, the park green space structure, the protective green space structure, the three-dimensional greening structure, and the idle green space structure; The project identification module of the object identification system imports the planning file, identifies the planning indicators in the planning file, and arranges them in chronological order; the demand identification module of the object identification system statistically analyzes the greening needs of residents and obtains the residents' intention indicators; The zoning module of the planning system matches the residents' intention indicators, planning indicators and the green space correlation structure of the target area, and delineates the optimized area; the phasing module of the planning system calculates the implementation order of each optimized area and generates planning suggestions.

Images