Urban informal greenbelt identification method and device

Abstract

The embodiment of the invention provides a city informal green space identification method and device. The method comprises the following steps: determining green space quadrats in a target city, obtaining measured data of each green space quadrat, carrying out first classification on each green space quadrat, determining a first category of each green space quadrat, calculating NDVI according to remote sensing image data of the target city, determining NDVI corresponding to each green space quadrat, and determining a second category of each green space quadrat according to the NDVI corresponding to each green space quadrat. Calculating the quartile of the NDVI according to the NDVI corresponding to each informal green space quadrat, taking the quartile as a threshold value, carrying out the second classification of each green space quadrat according to the NDVI corresponding to each green space quadrat, determining the second category of each green space quadrat, and taking the remote sensing image data corresponding to each green space quadrat as a sample and the second category corresponding to each green space quadrat as a label, according to the method, a data set is constructed, an initial urban informal greenbelt identification model is trained, an urban informal greenbelt identification model is obtained, and then the urban informal greenbelt is efficiently identified based on the model.

Description

Technical Field

[0001] The present invention relates to the technical field of urban planning, and in particular to a method and a device for identifying informal urban green space. Background Art

[0002] In the past few decades, with the rapid development of globalization and urbanization, urban ecosystems are facing unprecedented challenges. As an important part of improving urban living quality, maintaining biodiversity and alleviating urban heat island effect, the effective planning and management of urban green space is becoming increasingly important. However, traditional urban green space planning often ignores the value of urban spontaneous vegetation (USV). These spontaneously growing vegetation in urban gaps and marginal areas not only enriches the biodiversity of the city, but also plays an important role in regulating urban climate, improving air quality and increasing the connectivity of urban green space. It can be seen that the current identification of informal urban green space is often done through manual field identification, which is inefficient. In view of this, how to improve the efficiency of urban informal green space identification has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] In a first aspect, an embodiment of the present invention provides a method for training an urban informal green space identification model, the method comprising:

[0004] According to the green space and ecological space distribution map of the target city, determine the green space sample plots in the target city and obtain the measured data of each green space sample plot;

[0005] According to the measured data of each green space plot, each green space plot is classified for the first time to determine the first category of each green space plot, wherein the first category includes: cultivated plant land, informal green space;

[0006] Calculate the Normalized Difference Vegetation Index (NDVI) based on the remote sensing image data of the target city, and determine the NDVI corresponding to each green space sample plot based on the location information of each green space sample plot;

[0007] According to the NDVI corresponding to each informal green space sample plot, the quartile of NDVI is calculated, and this is used as the threshold. According to the NDVI corresponding to each green space sample plot, each green space sample plot is classified for the second time to determine the second category of each green space sample plot, wherein the second category includes: cultivated plant land, low-spontaneous plant land, medium-spontaneous plant land, and high-spontaneous plant land; low-spontaneous plant land, medium-spontaneous plant land, and high-spontaneous plant land belong to informal green space;

[0008] Taking the remote sensing image data corresponding to each green space sample as samples and the second category corresponding to each green space sample as labels, a data set is constructed. The initial urban informal green space recognition model is trained according to the data set to obtain the urban informal green space recognition model.

[0009] In some possible implementations of the first aspect, determining the green space sample plot in the target city according to the green space and ecological space distribution map of the target city includes:

[0010] According to the green space and ecological space distribution map of the target city, the boundary range of each green space patch in the target city was determined, and ArcGIS was used to randomly arrange green space plots, delete the green space plots falling on buildings, roads and water surfaces, and finally determine the green space plots in the target city.

[0011] In some possible implementations of the first aspect, performing a first classification on each green space sample plot according to the measured data of each green space sample plot to determine the first category of each green space sample plot includes:

[0012] For any green space sample plot, the proportion of cultivated plant area and the proportion of spontaneous plant area are calculated based on the actual measured data. If the proportion of cultivated plant area is greater than the preset threshold, it is determined to be a cultivated plant land; if the proportion of spontaneous plant area is greater than the preset threshold, it is determined to be an informal green space.

[0013] In some possible implementations of the first aspect, the normalized difference vegetation index NDVI is calculated based on the remote sensing image data of the target city, and the NDVI corresponding to each green space sample plot is determined based on the location information of each green space sample plot, including:

[0014] The average NDVI is calculated based on multiple remote sensing image data of the target city taken during the vegetation growing season. On this basis, the NDVI of the pixels corresponding to each green space sample plot is determined according to the location information of each green space sample plot. For any green space sample plot, the average NDVI of the corresponding pixels is calculated and used as the NDVI corresponding to the green space sample plot.

[0015] In some implementations of the first aspect, according to the NDVI corresponding to each informal green space sample plot, the quartile of the NDVI is calculated, and the quartile is used as a threshold, and according to the NDVI corresponding to each green space sample plot, each green space sample plot is secondly classified to determine the second category of each green space sample plot, including:

[0016] The NDVI corresponding to each informal green space sample plot was sorted in order from small to large, and the quartiles of NDVI were calculated and used as the threshold. If the NDVI corresponding to the green space sample plot was lower than the first quartile, it was determined to be a cultivated plant land; if the NDVI corresponding to the green space sample plot was between the first quartile and the second quartile, it was determined to be a low-intensity spontaneous plant land; if the NDVI corresponding to the green space sample plot was between the second quartile and the third quartile, it was determined to be a moderate-intensity spontaneous plant land; if the NDVI corresponding to the green space sample plot was higher than the third quartile, it was determined to be a high-intensity spontaneous plant land.

[0017] In some implementations of the first aspect, the initial urban informal green space identification model is trained according to the data set to obtain the urban informal green space identification model, including:

[0018] Divide the data set into training set and validation set;

[0019] The initial urban informal green space recognition model is trained using the training set, and the performance of the trained urban informal green space recognition model is verified based on the validation set to obtain the final urban informal green space recognition model.

[0020] In a second aspect, an embodiment of the present invention provides a method for identifying informal urban green spaces, the method comprising:

[0021] Obtain remote sensing image data of the target city;

[0022] The remote sensing image data is input into the urban informal green space identification model, and the urban informal green space identification model identifies the remote sensing image data to obtain the urban informal green space identification result;

[0023] The urban informal green space identification model is obtained based on the urban informal green space identification model training method as described above.

[0024] In some implementations of the second aspect, the method further includes:

[0025] Based on the identification results of urban informal green spaces, a distribution map of urban informal green spaces is generated.

[0026] In a third aspect, an embodiment of the present invention provides a device for training an urban informal green space identification model, the device comprising:

[0027] A determination module is used to determine the green space sample plots in the target city according to the green space and ecological space distribution map of the target city, and obtain the measured data of each green space sample plot;

[0028] The classification module is used to perform a first classification on each green space sample plot according to the measured data of each green space sample plot, and determine the first category of each green space sample plot, wherein the first category includes: cultivated plant land and informal green space;

[0029] The calculation module is used to calculate the normalized difference vegetation index NDVI according to the remote sensing image data of the target city, and determine the NDVI corresponding to each green space sample plot according to the location information of each green space sample plot;

[0030] The classification module is also used to calculate the quartile of NDVI according to the NDVI corresponding to each informal green space sample plot, and use this as a threshold to perform a second classification on each green space sample plot according to the NDVI corresponding to each green space sample plot, and determine the second category of each green space sample plot, wherein the second category includes: cultivated plant land, low-degree spontaneous plant land, medium-degree spontaneous plant land, and high-degree spontaneous plant land; low-degree spontaneous plant land, medium-degree spontaneous plant land, and high-degree spontaneous plant land belong to informal green space;

[0031] The training module is used to construct a data set using the remote sensing image data corresponding to each green space sample plot as a sample and the second category corresponding to each green space sample plot as a label, and train the initial urban informal green space recognition model according to the data set to obtain the urban informal green space recognition model.

[0032] In a fourth aspect, an embodiment of the present invention provides a device for identifying informal urban green spaces, the device comprising:

[0033] Acquisition module, used to acquire remote sensing image data of the target city;

[0034] The recognition module is used to input the remote sensing image data into the urban informal green space recognition model, and the urban informal green space recognition model recognizes the remote sensing image data to obtain the urban informal green space recognition result;

[0035] The urban informal green space identification model is obtained based on the urban informal green space identification model training method as described above.

[0036] In an embodiment of the present invention, the green space plots in the target city can be determined according to the green space and ecological space distribution map of the target city, and the measured data of each green space plot can be obtained to perform a first classification on each green space plot to determine the first category of each green space plot, which includes: cultivated plant land, informal green space, and the NDVI is calculated according to the remote sensing image data of the target city, and the NDVI corresponding to each green space plot is determined according to the location information of each green space plot, and the quartile of the NDVI is calculated according to the NDVI corresponding to each informal green space plot, and used as Threshold, according to the NDVI corresponding to each green space sample plot, each green space sample plot is classified for the second time to determine the second category of each green space sample plot, including: cultivated plant land, low spontaneous plant land, moderate spontaneous plant land, high spontaneous plant land. The latter three categories belong to informal green space. The remote sensing image data corresponding to each green space sample plot is used as a sample, and the second category corresponding to each green space sample plot is used as a label to construct a data set, so as to train the initial urban informal green space recognition model, and obtain the urban informal green space recognition model, and then based on the model to efficiently identify urban informal green space.

[0037] It should be understood that the contents described in the summary of the invention are not intended to limit the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The above and other features, advantages and aspects of the embodiments of the present invention will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. The accompanying drawings are provided for a better understanding of the present invention and do not constitute a limitation of the present invention. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, wherein:

[0039] Figure 1 A flow chart of a method for training an urban informal green space identification model provided by an embodiment of the present invention is shown;

[0040] Figure 2 A schematic diagram of a random forest workflow provided by an embodiment of the present invention is shown;

[0041] Figure 3 A flow chart of a method for identifying informal urban green space provided by an embodiment of the present invention is shown;

[0042] Figure 4 A distribution map of informal urban green space provided by an embodiment of the present invention is shown;

[0043] Figure 5 A structural diagram of a model training device for identifying informal green space in an urban area provided by an embodiment of the present invention is shown;

[0044] Figure 6 A structural diagram of a device for identifying informal urban green space provided by an embodiment of the present invention is shown;

[0045] Figure 7 A structural diagram of an exemplary electronic device capable of implementing an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] In addition, the term "and / or" in the present invention is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in the present invention generally indicates that the associated objects before and after are in an "or" relationship.

[0048] In order to solve the technical problems arising from the background technology, the embodiments of the present invention provide a method and device for identifying informal green spaces in cities. Specifically, according to the green space and ecological space distribution map of the target city, the green space plots in the target city are determined, and the measured data of each green space plot is obtained, so as to perform a first classification on each green space plot, and determine the first category of each green space plot, which includes: cultivated plant land, informal green space, and calculate the NDVI according to the remote sensing image data of the target city, and determine the NDVI corresponding to each green space plot according to the location information of each green space plot, and calculate the quartile of the NDVI according to the NDVI corresponding to each informal green space plot, and use it as the threshold, According to the NDVI corresponding to each green space sample plot, each green space sample plot is classified for the second time to determine the second category of each green space sample plot, which includes: cultivated plant land, low spontaneous plant land, moderate spontaneous plant land, and high spontaneous plant land. The latter three categories belong to informal green space. The remote sensing image data corresponding to each green space sample plot is used as a sample, and the second category corresponding to each green space sample plot is used as a label to construct a data set, so as to train the initial urban informal green space recognition model and obtain the urban informal green space recognition model, and then efficiently identify urban informal green space based on this model.

[0049] In this way, machine learning can be used to efficiently identify informal urban green spaces, provide accurate data support for urban planners, environmental scientists, and policy makers, promote the optimal allocation and management decisions of urban green space resources, and thus improve the quality of the urban environment and the quality of life of residents.

[0050] In conjunction with the accompanying drawings, a method and device for identifying informal urban green spaces provided by an embodiment of the present invention will be described in detail below through specific embodiments.

[0051] Figure 1 A flowchart of a method for identifying informal urban green space provided by an embodiment of the present invention is shown. Figure 1 As shown, the urban informal green space identification method 100 may include the following steps:

[0052] S110, determining green space plots in the target city according to the green space and ecological space distribution map of the target city, and obtaining measured data of each green space plot.

[0053] In some embodiments, the boundary range of each green space patch in the target city can be determined based on the green space and ecological space distribution map of the target city and other supporting data (such as urban land use / cover remote sensing interpretation map, urban terrain DEM map), and ArcGIS can be used to randomly arrange green space plots, delete green space plots falling on buildings, roads and water surfaces, and reserve some spare green space plots to finally determine the green space plots in the target city.

[0054] For example, the number of green space plots in the target city can be 3,500, involving various land types, such as park green space, residential green space, street trees, riverside, urban greenway, abandoned land, garbage dump, etc. In addition, considering that urban green space is scattered and small in area, the spontaneous plants involved in the survey are mainly herbs, with occasional seedlings of trees and shrubs, so the size of the green space plot here can be set to 1m*1m.

[0055] In some embodiments, researchers can conduct field measurements (i.e., plant surveys) on each green space plot in a specific time period (e.g., August to September 2023) and upload the measured data. The measured indicators (i.e., plant survey indicators) involved here include:

[0056] (1) Basic information of species: record the plant species name, average height, coverage and growth status in each green space sample plot. Plant species identification is based on the Flora of China and Local Flora.

[0057] (2) Species composition information: Based on the composition and growth conditions of plants, whether they are seedlings, whether they have not been pruned or mechanically damaged, etc., "cultivated plants" and "volunteer plants" are distinguished. Tall trees and shrubs that have been artificially planted and pruned, and herbs with regular distribution are judged on site as cultivated plants, while most herbs with scattered and irregular distribution, and scattered tree or shrub seedlings are volunteer plants.

[0058] S120, performing a first classification on each green space sample plot according to the measured data of each green space sample plot, and determining a first category of each green space sample plot.

[0059] The first category includes: cultivated plant land and informal green land. The first classification described here can be specifically as follows:

[0060] For any green space sample plot, the proportion of cultivated plant area and the proportion of spontaneous plant area are calculated based on the actual measured data. If the proportion of cultivated plant area is greater than the preset threshold (for example, 50%), it is determined to be a cultivated plant land; if the proportion of spontaneous plant area is greater than the preset threshold (for example, 50%), it is determined to be an informal green space (that is, spontaneous plant land).

[0061] S130, calculating the Normalized Difference Vegetation Index NDVI according to the remote sensing image data of the target city, and determining the NDVI corresponding to each green space sample plot according to the location information of each green space sample plot.

[0062] The location information of each green space plot described here is the geographical location information recorded by a handheld high-precision GPS device (accuracy <1 meter) when the green space plot is actually measured.

[0063] In some embodiments, the average NDVI can be calculated based on multiple remote sensing image data of the target city taken during the vegetation growing season (such as Sentinel-2 satellite data). On this basis, the NDVI of the pixels corresponding to each green space sample plot is determined according to the location information of each green space sample plot. For any green space sample plot, the average NDVI of the corresponding pixels is calculated and used as the NDVI corresponding to the green space sample plot.

[0064] As an example, here we can use Sentinel-2 satellite data (10-meter resolution) in remote sensing image data to calculate the Normalized Difference Vegetation Index (NDVI). The NDVI calculation formula used here is: NDVI = (NIR-RED) / (NIR+RED), where NIR is the reflectivity in the near-infrared wavelength range, and RED is the reflectivity in the red wavelength range. It can be seen that NDVI is an important tool for assessing vegetation density, coverage, growth status, and health status, and has been widely used in the world for quality assessment of green space, agricultural land, etc. Different from other applications. Here, we analyze, compare, infer and judge the characteristics of volunteer plants and cultivated plants (including garden plants and grain cultivated plants) to distinguish them. That is, because cultivated plants need to remove biomass according to their respective purposes (for example, garden plants need to be pruned regularly, while cultivated plants need to be harvested), the NDVI of volunteer plants will be higher than that of cultivated plants during the vegetation growth season (from May to October each year). Therefore, the annual average NDVI is used in the algorithm and remote sensing image interpretation, that is, from May to October each year, four remote sensing image data covering the target city are downloaded each month, with cloud coverage less than 15%, and then they are merged and geometrically corrected, and then the cloud-covered pixels are removed. In the final analysis, the annual average NDVI is calculated at each pixel level. On this basis, according to the location information of each green space sample plot, the NDVI of the pixels corresponding to each green space sample plot is determined, and the average NDVI of the corresponding pixels is calculated, and this is used as the NDVI corresponding to the green space sample plot.

[0065] S140, according to the NDVI corresponding to each informal green space sample plot, calculate the quartile of the NDVI, and use this as a threshold, and according to the NDVI corresponding to each green space sample plot, perform a second classification on each green space sample plot to determine the second category of each green space sample plot.

[0066] Among them, the second category includes: cultivated plant land, low-volume plant land, medium-volume plant land, and high-volume plant land; low-volume plant land, medium-volume plant land, and high-volume plant land belong to informal green space. The second classification described here can be specifically as follows:

[0067] The NDVI corresponding to each informal green space sample plot was sorted in order from small to large, and the quartiles of NDVI were calculated and used as the threshold. If the NDVI corresponding to the green space sample plot was lower than the first quartile, it was determined to be a cultivated plant land; if the NDVI corresponding to the green space sample plot was between the first quartile and the second quartile, it was determined to be a low-intensity spontaneous plant land; if the NDVI corresponding to the green space sample plot was between the second quartile and the third quartile, it was determined to be a moderate-intensity spontaneous plant land; if the NDVI corresponding to the green space sample plot was higher than the third quartile, it was determined to be a high-intensity spontaneous plant land.

[0068] S150, using the remote sensing image data corresponding to each green space sample plot as a sample and the second category corresponding to each green space sample plot as a label, constructing a data set, and training an initial urban informal green space recognition model according to the data set to obtain an urban informal green space recognition model.

[0069] In some embodiments, the data set can be divided into a training set and a verification set in a ratio of 4:1, and then the initial urban informal green space recognition model is trained using the training set, and the performance of the trained urban informal green space recognition model is verified based on the verification set to obtain the final urban informal green space recognition model.

[0070] For example, the above-mentioned urban informal green space identification model may be a random forest (RF), and its principle is as follows:

[0071] Random forest is a supervised classification method. Multiple samples are extracted from the original sample. Each sample relies on the principle of the fastest entropy reduction to build a decision tree, and then the predictions of the decision tree forest are combined. Because the decision tree forest has strong anti-interference ability, the decision tree forest can avoid overfitting in complex data situations, and the optimal classification result is determined by the majority of the voting results of the decision tree forest. It is widely used in land use classification, land change detection, biological and ecological sciences, and has excellent accuracy. Figure 2 As shown in the figure, random forest classification mainly uses random forest to randomize training samples, generate a single decision tree, construct a decision tree forest, and count the results with the most votes as the classification results.

[0072] Based on the urban informal green space identification model training method 100 provided in the embodiment of the present invention, the embodiment of the present invention also provides an urban informal green space identification method, such as Figure 3 As shown, the urban informal green space identification method 300 may include the following steps:

[0073] S310, acquiring remote sensing image data of a target city.

[0074] The remote sensing image data of the target city is synthesized from multiple remote sensing image data of the target city taken during the vegetation growing season.

[0075] S320, inputting the remote sensing image data into an urban informal green space identification model, and using the urban informal green space identification model to identify the remote sensing image data to obtain an urban informal green space identification result;

[0076] The urban informal green space identification model is obtained based on the urban informal green space identification model training method as described above.

[0077] It is worth noting that, in order to facilitate intuitive display of the urban informal green space identification results, the urban informal green space identification method 300 may further include:

[0078] Based on the identification results of urban informal green spaces, a distribution map of urban informal green spaces is generated.

[0079] For further understanding, the urban informal green space identification model training method 100 and the urban informal green space identification method 300 are described in detail in conjunction with a specific embodiment, as shown below:

[0080] Taking the area formed by the Chengdu Ecological Zone around the city and the 500-meter buffer area as an example, Chengdu (30°05'-31°26'N, 102°5'-104°53'E) is located in the western part of Sichuan Province, in the hinterland of the Chengdu Plain, and is one of the central cities and transportation hubs in western my country. It is surrounded by Longquan Mountain in the east and Longmen Mountain Range in the west. The terrain in the territory is high in the west and low in the east, mainly hills and mountains, with an altitude of mostly between 1000-3000m. The whole territory has a subtropical humid monsoon climate with distinct four seasons, an average annual temperature of 16℃, and an average annual precipitation of 895.6mm.

[0081] Driven by the construction of a park city, Chengdu has built a functional complex with the concept of "landscape, scenic area, accessible and participatory" through diversified integration, and constructed an ecological "green vein" that blends mountains and waters, connects urban and rural areas, and covers the entire region, realizing green welfare for the entire region. The seven wedge-shaped plots of Jincheng Greenway form a ring. The full-ring first-level greenway is 100 kilometers long, spanning 12 districts, connecting 121 characteristic ecological parks, and has 78 "one bridge, one scene" landscape bridges, involving 133 square kilometers of ecological land. As the outer edge of Chengdu's central urban area, the ring area has rich land use types, including farmland, parks, residential areas, water systems, etc., providing a typical area for the identification of informal urban green spaces.

[0082] The dataset here is collected on the Google Earth Engine platform. The samples in the dataset are the basis of supervised classification of remote sensing images. The accuracy of supervised classification is related to the quality of the sample points. Samples with good quality can significantly increase the classification accuracy. 3,500 samples were collected from the Chengdu Ring Road area. The specific process is as follows: first enter GEE to adjust the Sentinel-level-2A remote sensing image data as the base map, click Add Mark, and create a new layer. This study created 8 layers, entered the sample point name and assigned response attributes to the layer, such as highly spontaneous plant land, moderate spontaneous plant land, low spontaneous plant land, cultivated plant land (green land, cultivated plant land (agricultural land), development and construction land, roads, and water bodies. Because the spatial resolution of Sentinel-2 is 10m, the sample data has obvious texture characteristics, and most of them are selected through visual interpretation. Points, for the three categories of highly spontaneous vegetation land (also known as highly spontaneous vegetation land), moderately spontaneous vegetation land (also known as moderately spontaneous vegetation land), and low spontaneous vegetation land (also known as low spontaneous vegetation land), they are matched with reference data obtained through field sampling, and then classified by the method mentioned in this article, which improves the work efficiency and provides convenience for large-scale remote sensing. When training the model, 80% are selected as training samples and 20% are selected as verification samples. After the final training, the overall classification accuracy of the model is greater than 0.87, and the confusion matrix classification error is less than 0.15. After the model training is completed, the remote sensing image data of Chengdu can be identified based on the urban informal green space recognition model to obtain the urban informal green space recognition results of Chengdu. Then, based on the urban informal green space recognition results, the urban informal green space distribution map of Chengdu is generated. Specifically, it can be as follows Figure 4 shown.

[0083] In summary, according to the embodiments of the present invention, at least the following technical effects are achieved:

[0084] The present invention can not only efficiently and accurately identify and evaluate the distribution of spontaneous plants in cities, but also provide important support for urban ecological planning, green space management and biodiversity protection. By implementing the present invention, the management efficiency of informal urban green spaces can be significantly improved, and the allocation of urban greening resources and the optimal layout of green infrastructure can be promoted.

[0085] Ultimately, the scientific and rational use and management of informal urban green spaces can not only promote the recovery and improvement of the urban ecological environment, but also provide strong support for responding to climate change and enhancing the resilience of cities to cope with future challenges. This requires the joint efforts and participation of all parties including the government, research institutions, social organizations and the public to achieve the long-term goal of sustainable urban development.

[0086] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described order of actions, because according to the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0087] The above is an introduction to a method embodiment. The following is a further explanation of the solution of the present invention through an apparatus embodiment.

[0088] Figure 5 FIG. 4 shows a structural diagram of a model training device for identifying informal green space in an urban area provided by an embodiment of the present invention. Figure 5 As shown, the urban informal green space identification model training device 500 may include:

[0089] The determination module 510 is used to determine the green space sample plots in the target city according to the green space and ecological space distribution map of the target city, and obtain the measured data of each green space sample plot.

[0090] The classification module 520 is used to perform a first classification on each green space sample plot according to the measured data of each green space sample plot, and determine the first category of each green space sample plot, wherein the first category includes: cultivated plant land and informal green space.

[0091] The calculation module 530 is used to calculate the normalized difference vegetation index NDVI according to the remote sensing image data of the target city, and determine the NDVI corresponding to each green space sample plot according to the location information of each green space sample plot.

[0092] The classification module 520 is also used to calculate the quartile of NDVI according to the NDVI corresponding to each informal green space sample plot, and use this as the threshold to perform a second classification on each green space sample plot according to the NDVI corresponding to each green space sample plot, and determine the second category of each green space sample plot, wherein the second category includes: cultivated plant land, low spontaneous plant land, medium spontaneous plant land, and high spontaneous plant land; low spontaneous plant land, medium spontaneous plant land, and high spontaneous plant land belong to informal green space.

[0093] The training module 540 is used to construct a data set using the remote sensing image data corresponding to each green space sample plot as a sample and the second category corresponding to each green space sample plot as a label, and train the initial urban informal green space recognition model according to the data set to obtain the urban informal green space recognition model.

[0094] Understandably, Figure 5Each module / unit in the urban informal green space identification model training device 500 has the function of realizing Figure 1 The functions of the various steps in the urban informal green space identification model training method 100 shown in the figure can achieve their corresponding technical effects, and for the sake of brevity, they will not be repeated here.

[0095] Figure 6 FIG. 4 shows a structural diagram of a device for identifying informal urban green space provided by an embodiment of the present disclosure, such as Figure 6 As shown, the urban informal green space identification device 600 may include:

[0096] The acquisition module 610 is used to acquire remote sensing image data of a target city.

[0097] The recognition module 620 is used to input the remote sensing image data into the urban informal green space recognition model, and the urban informal green space recognition model recognizes the remote sensing image data to obtain the urban informal green space recognition result.

[0098] The urban informal green space identification model is obtained based on the urban informal green space identification model training method as described above.

[0099] Understandably, Figure 6 The various modules / units in the urban informal green space identification device 600 shown have the function of realizing Figure 3 The functions of the various steps in the urban informal green space identification method 300 shown in the figure can achieve their corresponding technical effects, and for the sake of brevity, they will not be repeated here.

[0100] Figure 7 A block diagram of an exemplary electronic device capable of implementing an embodiment of the present invention is shown. Electronic device 700 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic device 700 may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown in the present invention, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required in the present invention.

[0101] like Figure 7As shown, the electronic device 700 may include a computing unit 701, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 702 or a computer program loaded from a storage unit 708 into a random access memory (RAM) 703. In the RAM 703, various programs and data required for the operation of the electronic device 700 may also be stored. The computing unit 701, the ROM 702, and the RAM 703 are connected to each other via a bus 704. An input / output (I / O) interface 705 is also connected to the bus 704.

[0102] Multiple components in the electronic device 700 are connected to the I / O interface 705, including: an input unit 706, such as a keyboard, a mouse, etc.; an output unit 707, such as various types of displays, speakers, etc.; a storage unit 708, such as a disk, an optical disk, etc.; and a communication unit 709, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 709 allows the electronic device 700 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0103] The computing unit 701 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 701 performs the various methods and processes described above, such as method 100 or method 300. For example, in some embodiments, method 100 or method 300 may be implemented as a computer program product, including a computer program, which is tangibly contained in a computer-readable medium, such as a storage unit 708. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 700 via ROM 702 and / or communication unit 709. When the computer program is loaded into RAM 703 and executed by the computing unit 701, one or more steps of the method 100 or method 300 described above may be performed. Alternatively, in other embodiments, the computing unit 701 may be configured to execute the method 100 or the method 300 in any other appropriate manner (eg, by means of firmware).

[0104] The various embodiments described above in the present invention can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs, which may be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general programmable processor, which may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0105] The program code for implementing the method of the present invention can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code can be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0106] In the context of the present invention, a computer-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. A more specific example of a computer-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0107] It should be noted that the present invention also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute method 100 or method 300 and achieve the corresponding technical effects achieved by executing the methods in the embodiments of the present invention. For the sake of concise description, they will not be repeated here.

[0108] In addition, the present invention also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the method 100 or the method 300 is implemented.

[0109] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and the present invention is not limited here.

[0110] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for training a model for identifying informal urban green space, characterized in that: The method comprises: According to the green space and ecological space distribution map of the target city, determine the green space sample plots in the target city, and obtain the measured data of each green space sample plot; According to the measured data of each green space sample plot, each green space sample plot is first classified to determine the first category of each green space sample plot, wherein the first category includes: cultivated plant land and informal green space; Calculate the Normalized Difference Vegetation Index (NDVI) based on the remote sensing image data of the target city, and determine the NDVI corresponding to each green space sample plot based on the location information of each green space sample plot; According to the NDVI corresponding to each informal green space sample plot, the quartile of NDVI is calculated, and the quartile of NDVI is used as the threshold. According to the NDVI corresponding to each green space sample plot, each green space sample plot is classified for the second time to determine the second category of each green space sample plot, wherein the second category includes: cultivated plant land, low-spontaneous plant land, medium-spontaneous plant land, and high-spontaneous plant land; low-spontaneous plant land, medium-spontaneous plant land, and high-spontaneous plant land belong to informal green space; The remote sensing image data corresponding to each green space plot is used as a sample, and the second category corresponding to each green space plot is used as a label to construct a data set. The initial urban informal green space recognition model is trained according to the data set to obtain the urban informal green space recognition model.

2. The method according to claim 1, characterized in that Determining the green space sample plots in the target city based on the green space and ecological space distribution map of the target city includes: According to the green space and ecological space distribution map of the target city, the boundary range of each green space patch in the target city is determined, and green space plots are randomly arranged using ArcGIS, and green space plots falling on buildings, roads and water surfaces are deleted, and finally the green space plots in the target city are determined.

3. The method according to claim 1, characterized in that The first classification of each green space sample plot is performed according to the measured data of each green space sample plot to determine the first category of each green space sample plot, including: For any green space sample plot, the proportion of cultivated plant area and the proportion of spontaneous plant area are calculated based on the actual measured data. If the proportion of cultivated plant area is greater than the preset threshold, it is determined to be a cultivated plant land; if the proportion of spontaneous plant area is greater than the preset threshold, it is determined to be an informal green space.

4. The method according to claim 1, characterized in that: The step of calculating the Normalized Difference Vegetation Index (NDVI) according to the remote sensing image data of the target city, and determining the NDVI corresponding to each green space sample plot according to the location information of each green space sample plot, includes: The average NDVI is calculated based on multiple remote sensing image data of the target city taken during the vegetation growing season. On this basis, the NDVI of the pixels corresponding to each green space sample plot is determined according to the location information of each green space sample plot. For any green space sample plot, the average NDVI of the corresponding pixels is calculated and used as the NDVI corresponding to the green space sample plot.

5. The method according to claim 1, characterized in that The NDVI corresponding to each informal green space plot is used to calculate the quartile of the NDVI, and the quartile is used as a threshold. The second classification of each green space plot is performed according to the NDVI corresponding to each green space plot to determine the second category of each green space plot, including: The NDVI corresponding to each informal green space sample plot was sorted in order from small to large, and the quartiles of NDVI were calculated and used as the threshold. If the NDVI corresponding to the green space sample plot was lower than the first quartile, it was determined to be a cultivated plant land; if the NDVI corresponding to the green space sample plot was between the first quartile and the second quartile, it was determined to be a low-intensity spontaneous plant land; if the NDVI corresponding to the green space sample plot was between the second quartile and the third quartile, it was determined to be a moderate-intensity spontaneous plant land; if the NDVI corresponding to the green space sample plot was higher than the third quartile, it was determined to be a high-intensity spontaneous plant land.

6. The method according to claim 1, characterized in that The initial urban informal green space identification model is trained according to the data set to obtain the urban informal green space identification model, including: Dividing the data set into a training set and a validation set; The initial urban informal green space recognition model is trained using the training set, and the performance of the trained urban informal green space recognition model is verified according to the verification set to obtain a final urban informal green space recognition model.

7. A method for identifying informal urban green spaces, characterized in that: The method comprises: Obtain remote sensing image data of the target city; Inputting the remote sensing image data into an urban informal green space identification model, and using the urban informal green space identification model to identify the remote sensing image data to obtain an urban informal green space identification result; The urban informal green space identification model is obtained based on the urban informal green space identification model training method described in any one of claims 1 to 6.

8. The method according to claim 7, characterized in that The method further comprises: According to the urban informal green space identification results, a distribution map of urban informal green space is generated.

9. A training device for identifying an informal urban green space model, characterized in that: The device comprises: A determination module is used to determine the green space sample plots in the target city according to the green space and ecological space distribution map of the target city, and obtain the measured data of each green space sample plot; A classification module, for performing a first classification on each green space sample plot according to the measured data of each green space sample plot, and determining a first category of each green space sample plot, wherein the first category includes: cultivated plant land, informal green space; A calculation module, used to calculate the normalized difference vegetation index NDVI according to the remote sensing image data of the target city, and determine the NDVI corresponding to each green space sample plot according to the location information of each green space sample plot; The classification module is further used to calculate the quartile of NDVI according to the NDVI corresponding to each informal green space sample plot, and use this as a threshold, and perform a second classification on each green space sample plot according to the NDVI corresponding to each green space sample plot, and determine the second category of each green space sample plot, wherein the second category includes: cultivated plant land, low-degree spontaneous plant land, medium-degree spontaneous plant land, and high-degree spontaneous plant land; low-degree spontaneous plant land, medium-degree spontaneous plant land, and high-degree spontaneous plant land belong to informal green space; The training module is used to construct a data set using the remote sensing image data corresponding to each green space sample plot as a sample and the second category corresponding to each green space sample plot as a label, and train the initial urban informal green space recognition model according to the data set to obtain the urban informal green space recognition model.

10. A device for identifying informal urban green space, characterized in that: The device comprises: Acquisition module, used to acquire remote sensing image data of the target city; An identification module is used to input the remote sensing image data into an urban informal green space identification model, and the urban informal green space identification model identifies the remote sensing image data to obtain an urban informal green space identification result; The urban informal green space identification model is obtained based on the urban informal green space identification model training method described in any one of claims 1 to 6.

Images