Land utilization category classification method and device

Abstract

The invention discloses a land utilization category classification method and device, relates to the technical field of land classification data processing, and aims to at least solve the technical problem that the classification result is inaccurate when land classification is carried out from natural attributes. The method comprises the following steps: dividing a target region into a plurality of first contour regions according to natural attribute feature information; dividing the target area into a plurality of second contour areas according to the social attribute feature information; each first contour area corresponds to a first ground feature category; the correlation degree of the natural attribute feature information between the pixel points in the first contour area is greater than or equal to a first correlation degree threshold value; each second contour area corresponds to a second ground feature category; and based on the plurality of second contour regions, performing secondary classification on the plurality of first contour regions of the target region to obtain a target category classification result of the target region, the result comprising the first ground feature category and / or the second ground feature category.

Description

Technical Field

[0001] The present application relates to the technical field of data processing for land classification, and in particular to a method and device for classifying land use categories. Background Art

[0002] Human production activities and population growth influence global land use changes. These dynamic changes in land use, in turn, influence and guide human production and living practices by altering the climate and ecological environment. Land use feature classification is a key indicator of urban morphology and the basis for extracting two- and three-dimensional urban morphological characteristics. Therefore, conducting land use classification research is fundamental to studying two- and three-dimensional urban morphological characteristics. However, current land use classification methods, typically based on natural attributes, fail to fully reflect the specific role of land in socioeconomic activities and therefore fail to fully characterize the actual land use situation and demand. Summary of the Invention

[0003] The present invention provides a land use classification method and device to at least address the technical problem in related technologies where land classification based on natural attributes fails to fully characterize the actual conditions and utilization requirements of the land. The technical solution of the present invention is as follows:

[0004] According to a first aspect of an embodiment of the present invention, a land use category classification method is provided, the method comprising: extracting natural attribute feature information of a target area from point cloud data and image data, and extracting social attribute feature information of the target area from point of interest data of the target area; dividing the target area into a plurality of first contour areas according to the natural attribute feature information; and dividing the target area into a plurality of second contour areas according to the social attribute feature information; each first contour area corresponds to a first land feature category, the first land feature category representing the natural attribute land feature category; the correlation degree of the natural attribute feature information between pixel points in the first contour area is greater than or equal to a first correlation threshold; each second contour area corresponds to a second land feature category, the second land feature category representing the social attribute land feature category; based on the plurality of second contour areas, performing secondary classification on the plurality of first contour areas of the target area to obtain a target category classification result of the target area; the target category classification result includes the first land feature category and / or the second land feature category.

[0005] According to a second aspect of an embodiment of the present invention, a land use category classification device is provided, which includes: an extraction unit for extracting natural attribute feature information of a target area from point cloud data and image data, and extracting social attribute feature information of the target area from point of interest data of the target area; a first classification unit for dividing the target area into multiple first contour areas according to the natural attribute feature information; and dividing the target area into multiple second contour areas according to the social attribute feature information; each first contour area corresponds to a first land feature category, and the first land feature category represents the natural attribute land feature category; the correlation degree of the natural attribute feature information between pixel points in the first contour area is greater than or equal to a first correlation degree threshold; each second contour area corresponds to a second land feature category, and the second land feature category represents the social attribute land feature category; a second classification unit for performing secondary classification on the multiple first contour areas of the target area based on the multiple second contour areas to obtain a target category classification result of the target area; the target category classification result includes the first land feature category and / or the second land feature category.

[0006] According to a third aspect of an embodiment of the present invention, a land use category classification system is provided, which is configured to execute the land use category classification method of the first aspect and any possible implementation thereof.

[0007] According to a fourth aspect of an embodiment of the present invention, a computer device is provided, comprising: a processor and a memory for storing processor-executable instructions; wherein the processor is configured to execute the executable instructions to implement a land use category classification method such as the first aspect and any possible implementation thereof.

[0008] According to a fifth aspect of an embodiment of the present invention, a computer-readable storage medium is provided, on which instructions are stored. When the instructions in the computer-readable storage medium are executed by a processor of a computer device, the computer device is enabled to execute the land use category classification method such as the first aspect and any possible implementation thereof.

[0009] According to the sixth aspect of the embodiment of the present application, a computer program product is provided, which includes computer instructions. When the computer instructions are executed on a computer device, the computer device executes the land use category classification method of the first aspect and any possible implementation thereof.

[0010] The technical solution provided by the embodiments of the present invention brings at least the following beneficial effects: extracting natural attribute feature information of the target area from point cloud data and image data to classify the target area according to the natural attribute features and divide the target area into multiple first contour areas; at the same time, extracting social attribute feature information of the target area from the point of interest data of the target area to classify the target area according to the social attribute features and divide the target area into multiple second contour areas. Furthermore, the multiple first contour areas are reclassified based on the multiple second contour areas, so that the final target classification result can not only accurately count the area, distribution and quality status of various types of land, timely grasp the dynamic changes of land resources, but also reflect the specific role of land in social and economic activities, so that the land classification result can more comprehensively characterize the actual situation and utilization needs of the land.

[0011] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the specification are used to explain the principles of the present application, and do not constitute an improper limitation on the present application.

[0013] Figure 1 is a flow chart showing a method for classifying land use categories according to an exemplary embodiment;

[0014] Figure 2 is a diagrammatic representation of a result of fusion of data of different feature dimensions according to an exemplary embodiment;

[0015] Figure 3 is a schematic block diagram of a land use classification device according to an exemplary embodiment;

[0016] Figure 4 The figure is a schematic diagram showing a computer device according to an exemplary embodiment. DETAILED DESCRIPTION

[0017] In order to enable ordinary people in the art to better understand the technical solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0018] It should be noted that the terms "first," "second," and the like in the specification and claims of this application and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate so that the embodiments of the application described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. Instead, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0019] Before giving a detailed introduction to the land use category classification method provided in the embodiment of the present application, a brief introduction to the application scenarios involved in the embodiment of the present application is first given.

[0020] Human production activities and population growth influence global land use changes. These dynamics, in turn, influence and guide human production and living practices through alterations in the climate and ecological environment. Land use feature classification is a key indicator of urban morphology and the basis for extracting two- and three-dimensional urban morphological features. Therefore, conducting land use classification research is fundamental to studying two- and three-dimensional urban morphological features. With the accelerated urbanization process in my country, the disorderly expansion of construction land, the encroachment of green space resources, and environmental pollution have emerged, posing significant challenges to urban planning and development. Therefore, conducting land use classification research to accurately understand the utilization of land resources provides fundamental data and decision-making support for scientific urban development planning, rational land resource utilization, and improved living environments. This research has important theoretical and practical value. However, current land use classification methods, typically based on natural attributes, fail to fully reflect the specific role of land in socioeconomic activities and therefore fail to fully characterize the actual land use situation and demand.

[0021] Furthermore, categorizing land solely based on its natural attributes fails to capture the specific role of land in socioeconomic activities. For example, a piece of construction land can be a commercial zone, an industrial zone, or a residential zone, each with vastly different economic outputs and employment opportunities. However, categorizing land based on its natural attributes might simply classify it as construction land, failing to distinguish its distinct economic functions and values.

[0022] Based on this, the following research was conducted. Satellite remote sensing technology can obtain earth surface coverage information on a large scale, quickly and accurately. It has the advantages of high efficiency, precision, comprehensiveness, real-time and periodicity. Remote sensing sensors can sense and record the spectral reflectance of different surface cover features from visible light to infrared bands and from medium to extremely high spatial resolution, which is conducive to better image interpretation and processing. It is an important technical means to grasp the current land use status.

[0023] Currently, high-spatial-resolution images, multi- / hyperspectral images, and LiDAR have been widely used in land use classification research. However, with the continuous development of cities, land feature types are becoming increasingly diverse and complex. A single remote sensing data source can no longer meet the requirements for extracting complex land use feature categories in cities. The fusion of multi-source remote sensing data can effectively complement each other's strengths. Multi- / hyperspectral remote sensing images can obtain more or continuous spectral information of land features, but the spatial resolution of most images is not high, making it difficult to meet the requirements of fine land use classification. Laser scanners can obtain high-precision three-dimensional surface spatial information without being affected by lighting conditions, but LiDAR data lacks sufficient spectral information, making it difficult to identify land features with similar height characteristics. Airborne aerial images can obtain high-resolution images in the visible light band, which can be collected and acquired simultaneously with LiDAR point cloud data.

[0024] Therefore, the fusion of airborne LiDAR and image data not only complements each other's strengths but also eliminates the problem of data acquisition time asynchrony, making high-resolution urban land use classification research possible. However, whether it is based on remote sensing imagery or airborne LiDAR point cloud data, the features of land objects only describe the natural properties of the land objects and lack a description of their socioeconomic attributes. Land use classification is the result of the interaction between the natural world and human activities.

[0025] Therefore, when considering the fusion of multi-source data in fine-grained land use classification research, it is also necessary to introduce socioeconomic data that can represent human activities. With the rapid development of information science and technology, urban points of interest (POI) based on big data are a representative of the rapid development of Internet technology and the increasing updating of online electronic maps. They describe the spatial and attribute information of geographic entities, greatly enhancing the ability to obtain entity locations and connecting people's daily life trajectories with the world of geographic information. With strong computing and expression capabilities, they can better reflect human activity information in cities and have important reference value for geospatial data mining and location-based regional population analysis.

[0026] Based on this, we will further investigate land use classification by fusing high-resolution remote sensing data with point of interest (POI) data. This classification study will not only examine the differences in natural attributes between land use types, but also consider the differential impacts of human socioeconomic activities on different land feature types. We will fully exploit the direct and indirect features of airborne LiDAR point cloud data, the spectral and spatial features of aerial imagery, and extract features from POI data. We will design classification schemes for both single-source and multi-source data fusion, compare the impact of different input features on classification accuracy, and analyze the advantages and disadvantages of different data fusion classification schemes. We will also conduct classification experiments using machine learning algorithms, including the K-Nearest Neighbor (KNN) algorithm, which is widely used in land feature classification, the Random Forest (RF) algorithm, which has achieved good classification results, and the newer extreme gradient boosting (XBGoost) algorithm, which is widely used in major competitions. The results of the three classifiers will be analyzed to test the stability of the classification schemes and to select the one with the highest classification accuracy from all three classifiers for the land use classification output. Finally, the classification scheme with the best accuracy is selected to realize the mapping output of land use classification.

[0027] In response to the above research findings, this application provides a land use classification method, which extracts the natural attribute feature information of the target area from the point cloud data and image data to classify the target area according to the natural attribute features, and divides the target area into multiple first contour areas; at the same time, the social attribute feature information of the target area is extracted from the point of interest data of the target area to classify the target area according to the social attribute features, and divides the target area into multiple second contour areas. Furthermore, the multiple first contour areas are reclassified based on the multiple second contour areas, so that the final target classification result can not only accurately count the area, distribution and quality status of various types of land, timely grasp the dynamic changes of land resources, but also reflect the specific role of land in social and economic activities, so that the land classification result can more comprehensively characterize the actual situation and utilization needs of the land.

[0028] The land use classification method provided in the embodiment of the present application can be applied to a land use classification system or a computer device for land use classification. For ease of understanding, the land use classification method provided in the present application is described in detail below with reference to the accompanying drawings.

[0029] Figure 1 is a flow chart showing a land use classification method according to an exemplary embodiment. Figure 1 As shown, the land use category classification method includes the following steps.

[0030] S11, extracting natural attribute feature information of the target area from the point cloud data and the image data, and extracting social attribute feature information of the target area from the point of interest data of the target area.

[0031] In some embodiments, natural attribute feature information includes indirect feature information, direct feature information, spectral feature information, and image spatial feature information. Direct feature information includes elevation feature information and intensity feature information; indirect feature information includes spatial geometric feature information of the point cloud and texture feature information used to describe the correlation between the grayscale of two points within a preset distance and in a preset direction; spatial geometric feature information includes the undulation of the ground object on the surface, the flatness of the ground object surface, and the degree of dispersion of the point cloud; and image spatial feature information includes: morphological building index, morphological shadow index, and urban complexity.

[0032] The step S11 is specifically implemented through the following steps.

[0033] First, as a data acquisition method, point cloud data is obtained from the laser radar scanner of the airborne LiDAR measurement system, and image data is obtained from the airborne digital camera.

[0034] In order to ensure the accuracy of the image data, the resolution of the onboard digital camera is set to be higher than the resolution threshold.

[0035] Second, indirect and direct feature information for each pixel at each location in the target area is extracted from the point cloud data, and spectral and spatial feature information for each pixel at each location is extracted from the image data. It should be understood that the point cloud data and image data described above are image data for the land area to be classified.

[0036] In some embodiments, the point cloud data is airborne LiDAR point cloud data, and the image data may also be referred to as airborne image data. The point cloud data includes ground points and non-ground points and is discrete point cloud data.

[0037] In some embodiments, an airborne LiDAR measurement system can simultaneously acquire LiDAR point cloud data and aerial image data. The airborne laser scanning system is integrated with hardware such as a RIEGL VUX-1LR LiDAR scanner, a PHASE ONE IXU1000-R high-resolution digital camera, and a POS system consisting of an IMU and differential GPS. The entire system is mounted on a flight platform or aircraft (such as a manned aircraft or drone) for data acquisition. During data acquisition, the LiDAR scanner is used to obtain the X, Y, and Z three-dimensional coordinates and intensity information of the target object in real time, and the digital camera is used to obtain high-resolution image data of the target object in the visible light band. The POS system uses combined navigation using IMU inertial navigation information and differential GPS positioning information to obtain the carrier's motion posture and position information (i.e., POS trajectory information) for subsequent point cloud fusion and solution, as well as image correction.

[0038] The point cloud data used above has undergone preliminary preprocessing, including the acquisition of original data, calculation of POS trajectory, and fusion of POS and laser point cloud data to generate point cloud data in the geographic coordinate system, which will not be repeated here.

[0039] In an embodiment of the present application, the orthorectification of airborne image data is achieved by using model key points extracted from point cloud ground points based on point cloud data as ground control points. Therefore, the geographic coordinates of various feature information extracted based on point cloud data and image data are corresponding and unified, so as to facilitate the unified definition of the coordinate systems of feature information of each dimension.

[0040] S12, dividing the target area into a plurality of first contour areas according to the natural attribute feature information; and dividing the target area into a plurality of second contour areas according to the social attribute feature information.

[0041] Each first contour area corresponds to a first land feature category, and the first land feature category represents a land feature category with natural attributes.

[0042] The correlation degree of the natural attribute feature information between the pixel points in the first contour area is greater than or equal to the first correlation threshold; each second contour area corresponds to a second land feature category, and the second land feature category represents a social attribute land feature category.

[0043] S13, performing secondary classification on the multiple first contour areas of the target area based on the multiple second contour areas to obtain a target category classification result of the target area.

[0044] The target category classification result includes the first feature category and / or the second feature category.

[0045] Through the above-described implementation, natural attribute feature information of the target area is extracted from the point cloud data and image data to classify the target area according to the natural attribute features, dividing the target area into multiple first contour areas. Simultaneously, social attribute feature information of the target area is extracted from the point of interest data of the target area to classify the target area according to the social attribute features, dividing the target area into multiple second contour areas. Furthermore, the multiple first contour areas are further classified based on the multiple second contour areas, so that the final target classification result can not only accurately count the area, distribution, and quality status of various types of land, timely grasp the dynamic changes of land resources, and reflect the specific role of land in social and economic activities, but also make the land classification result more comprehensive in representing the actual situation and utilization needs of the land.

[0046] As an implementation manner, in the above step S12, the target area is divided into a plurality of second contour areas according to the social attribute feature information. The specific process is as follows.

[0047] First, determine the kernel density of the point of interest data at each location in the target area.

[0048] POI point data, consisting of a series of point data with attributes such as name, address, coordinates, and category, possesses strong computational and representation capabilities and is widely used in urban spatial analysis and visualization. Kernel density analysis is used to calculate the unit density of point and line feature measurements within a specified neighborhood. It intuitively reflects the distribution of discrete measurements within a continuous area, ultimately producing a smooth surface with large median values and small peripheral values. The grid value represents the unit density. Kernel density analysis effectively represents the spatial distribution of POI point data and is a commonly used method for expressing POI data. The specific expression is shown in Formula (1-1).

[0049]

[0050] In the above formula (1-1), i = 1, 2, ..., n are the points in the input point data. If these points are within the radius distance of the (x, y) position, only the points in the sum are included; D(x, y) is the density prediction value of the new (x, y) point; r is the search radius; pop i is the weight value of the point (can be ignored and assigned a value of 1); d i is the distance between point i and (x, y).

[0051] Specifically, it can be understood as spatially transforming POI data that express human activities, social, economic and other phenomena, and using the kernel density analysis method to transform discrete point data into raster data that can be spatially analyzed and expressed.

[0052] In some implementations, the process of extracting features from POI data of a target area is as follows.

[0053] First, the POI data needed to be reclassified. The target area's POI data types primarily included 20 categories: automotive services, car sales, car repair, motorcycle services, catering services, shopping services, lifestyle services, sports and leisure services, healthcare services, accommodation services, scenic spots, commercial residences, government agencies and social groups, science, education, and culture services, transportation facilities services, financial and insurance services, companies and enterprises, road ancillary facilities, place names and addresses, and public facilities. These data were primarily distributed in built-up areas. During the reclassification process, these 20 categories were reclassified based on their social functional attributes into residential areas, commercial areas, industrial areas, public services, and open spaces.

[0054] Then, kernel density analysis is performed on the POI data containing functional area category attributes, and the search radius is set to a preset length (such as 500m).

[0055] Finally, the classified POI data including POI data features are rasterized to ensure that the generated POI data has a 1-meter resolution raster image that is consistent with the resolution of the airborne image data, also known as a POI distribution heat map. Based on the range of the block, the mean (mn), standard deviation (std), and sum (sum) of the kernel density of residential areas, commercial areas, industrial areas, public services, and open spaces in each block are extracted to obtain 15 POI kernel density features.

[0056] Secondly, referring to the preset kernel density ranges corresponding to different second feature categories, the second target feature categories corresponding to the target kernel density range to which the kernel density of the point of interest data at each position in the target area belongs are divided.

[0057] The second target feature category is any feature category in the second feature category.

[0058] Thirdly, the candidate region consisting of positions in the target region that belong to the same second target object category and are continuous in position is determined as the second contour region corresponding to the same second object category.

[0059] Optionally, in the above step S22, the target area is divided into multiple first contour areas according to the natural attribute feature information, specifically including: determining the natural attribute feature information of each position in the target area in multiple dimensions; dividing multiple positions in the target area whose similarity of the natural attribute feature information in multiple dimensions is higher than the similarity threshold corresponding to each dimension and whose positions are continuous into the same first contour area, and determining the first target land feature category corresponding to the same first contour area based on the natural attribute feature information of the same first contour area; the first target land feature category is any land feature category in the first land feature category.

[0060] Optionally, in order to improve classification accuracy, the above step S22 of dividing the target area into a plurality of first contour areas according to the natural attribute feature information can be specifically implemented in the following manner: One pixel represents one position.

[0061] First, a plurality of different land use classification models are used to classify the land object category to which each pixel point belongs according to the feature similarity between the indirect feature information, direct feature information, spectral feature information, visible light vegetation index and image space feature information of each pixel point in the image corresponding to the target area being higher than the corresponding similarity threshold, and obtain a plurality of first classification results; the land use classification model characterizes the correlation between each feature information of the point cloud data and the image data and the land object category to which each pixel point belongs.

[0062] Secondly, based on the frequency of the object category to which each pixel point in the multiple first classification results belongs, the object category to which each pixel point in the multiple first classification results belongs is evaluated and analyzed, and the object category to which each pixel point with the highest evaluation result belongs is determined as the first target object category of each pixel point, thereby obtaining a second classification result including each first target object category to which each pixel point belongs.

[0063] Third, the image including each pixel point is segmented according to a preset segmentation scale, and the elevation feature information in the first multidimensional feature information and the spectral feature information in the second multidimensional feature information are used as the basis for determining the contour features of the object. Based on the preset merging conditions, the segmented areas in the image are merged to obtain an image including multiple contour areas corresponding to multiple target objects; the multiple pixel points included in a contour area belong to one target object.

[0064] The above-mentioned preset merging conditions include one or more of the following: minimizing object heterogeneity or minimizing image smoothness heterogeneity, minimizing image compactness heterogeneity, minimizing image shape heterogeneity, and minimizing the average heterogeneity of all objects; the above-mentioned preset segmentation scale is determined according to the object type of the segmented object.

[0065] Fourth, based on the target object to which the pixel in the contour region belongs, the first target object category to which the pixel in the second classification result corresponds is corrected, thereby obtaining a third classification result including the second target object category of each pixel. The third classification result includes multiple first contour regions.

[0066] This classification method determines feature classification based on the similarity between the first multi-dimensional feature information of each pixel in the point cloud data and the second multi-dimensional feature information of each pixel in the image data, both from multiple data sources. This ensures that each first classification result is based on data from multiple data sources with different dimensions, thereby ensuring the accuracy of each first classification result. Each first classification result is then evaluated, and the best-evaluated feature classification result is selected as the second classification result. This reduces the impact of instability in the application of each single classification algorithm on the first classification result, thereby ensuring a more accurate second classification result. Furthermore, to avoid the "salt and pepper" phenomenon in the classification results caused by over-considering data dimensions (i.e., the fragmented distribution of two- and three-dimensional target features), an object-oriented classification method is used to further classify the feature categories of the image based on two dimensional information: elevation feature information and spectral feature information. This determines the target object to which multiple pixels within the contour area belong. Based on this target object classification result, any potential "salt and pepper" phenomenon in the second classification result is corrected, ultimately resulting in a more accurate third classification result.

[0067] Optionally, in order to improve the representation dimension of the classification result, in the above step S23, multiple first contour areas of the target area are secondary classified based on multiple second contour areas to obtain the target category classification result of the target area, which specifically includes the following steps.

[0068] First, it is determined that there is an overlapping area between the first contour area and the second contour area of the target area.

[0069] Secondly, if the category attributes of the first target object category of the first contour area corresponding to the repeated area and the second target object category of the second contour area are the same, and the granularity of the first target object category is higher than the granularity of the second target object category, then the repeated area in the first contour area is classified according to the second target object category of the repeated area.

[0070] The larger the granularity of the feature category, the lower the accuracy of its land division, and the higher the granularity priority of the corresponding feature category.

[0071] Third, if the first target object category of the first contour area and the second target object category of the second contour area corresponding to the repeated area have the same category attributes, and the granularity of the first target object category is lower than that of the second target object category, the first target object category of the repeated area in the first contour area is retained.

[0072] Fourthly, if the first target object category of the first contour area and the second target object category of the second contour area corresponding to the repeated area have different category attributes, the repeated area in the first contour area is classified according to the second target object category of the repeated area.

[0073] The first feature category includes one or more of the following: building land, bare land, cultivated land, grassland, road, woodland, and water bodies; the second feature category includes one or more of the following: residential land, commercial land, industrial land, public service and administrative land, educational and scientific research land, green space, and square land. The first target feature category is any of the first feature categories. The second target feature category is any of the second feature categories.

[0074] Building land shares the same category attributes as residential land, commercial land, and industrial land. The granularity of building land is higher than that of residential land, commercial land, and industrial land. Therefore, if the first target feature category of an overlapping region is building land, and the second target feature category of the overlapping region is residential land, commercial land, or industrial land, the overlapping region is reclassified into the second target feature category accordingly.

[0075] Grassland, woodland, and water bodies share the same category attributes as green space and squares. The granularity of grassland, woodland, and water bodies is lower than that of green space and squares. Therefore, if the first target feature category of an overlapping area is grassland, woodland, or water, and the second target feature category of the overlapping area is green space or square, the overlapping area is reclassified as the first target feature category.

[0076] Bare land and public service and management land have different category attributes. If the first target feature category of the overlapping area is bare land and the second target feature category of the overlapping area is public service or management land, the overlapping area is divided into the second target feature category accordingly.

[0077] As another land classification method, to ensure rapid classification, land classification is performed according to a trained land classification model. Specifically, the natural attribute characteristic information and social attribute characteristic information of each location in the target area are input into multiple different land use classification models to obtain different land feature classification results. The different land feature classification results are evaluated based on the frequency of the land feature category belonging to the same location in the different land feature classification results. The land feature category classification result with the highest frequency in the evaluation results is determined as the target classification result.

[0078] The above-mentioned multiple different land use classification models include a K-nearest neighbor classifier algorithm model, a random forest classifier algorithm model, and an extreme gradient boosting classifier algorithm model.

[0079] As a classification model, the above-mentioned multiple different land use classification models are different machine learning models. The specific process of obtaining the land use classification model includes the following steps.

[0080] First, construct multiple different initial models that represent the relationship between the various feature information of point cloud data, image data, and POI data and the ground feature category to which each pixel belongs; these initial models include a K-nearest neighbor classifier algorithm model, a random forest classifier algorithm model, and an extreme gradient boosting classifier algorithm model;

[0081] Second, extracting historical indirect feature information and historical direct feature information of each historical pixel from multiple sets of historical point cloud data, extracting historical spectral feature information and historical image spatial feature information of each historical pixel from multiple sets of historical image data corresponding to the multiple sets of historical point cloud data, and obtaining social attribute feature information from POI data, and obtaining multiple sets of historical images corresponding to the multiple sets of historical point cloud data, wherein the feature category to which each historical pixel belongs has been marked;

[0082] Third, according to the preset ratio rule, the training sample data set and the validation sample data set are selected from multiple sets of historical point cloud data, multiple sets of historical image data, historical POI data and multiple sets of historical images; the preset ratio rule is to select the number of samples for each feature category based on the ratio of each feature category in the historical image area;

[0083] Fourthly, multiple initial models are trained separately based on the training sample data set, and the validation sample data set is used to validate each trained initial model, so that each initial model with the best validation result is used as multiple different land use classification models.

[0084] As a specific implementation method, the selection of training sample data sets mainly includes the following steps: First, we randomly create sample points covering the entire study area. These samples do not contain any feature attribute information; second, we combine high-resolution Google Earth images and aerial true color images with similar imaging time to classify the sample points; then, we use the random sampling method again to divide the sample points into 80% training data sets and 20% test data sets (there is no overlap between the training data sets and the test data sets); finally, according to the synthetic images generated in the data fusion scheme, the feature information (band information) contained in the synthetic images of each classification scheme is extracted to the corresponding training sample points and test sample points as their attribute values. Based on the above steps, the data fusion results of different feature dimensions are obtained, as shown below. Figure 2 The diagram shows sample data sets for fourteen classification schemes, where each sample point contains the corresponding input feature information.

[0085] Optionally, the target area is divided into a plurality of first contour areas according to the natural attribute feature information, which is specifically implemented in the following manner.

[0086] According to the data features of the point cloud data and image data of the target area, a first target model is determined from multiple first preset classification models; the natural attribute feature information of the target area is input into the first target model, and the target area is divided into multiple first contour areas.

[0087] Optionally, the target area is divided into multiple second contour areas according to the social attribute characteristic information, which is specifically implemented in the following manner: based on the data characteristics of the points of interest in the target area, a second target model is determined from multiple second preset classification models; the social attribute characteristic information of the target area is input into the second target model, and the target area is divided into multiple second contour areas.

[0088] The plurality of first preset classification models and the plurality of second preset classification models may be, respectively, a K-nearest neighbor classifier algorithm model (KNN), a random forest classifier algorithm model (RF), and an extreme gradient boosting classifier algorithm model (XGBoost). The acquisition and training methods for the plurality of first preset classification models and the plurality of second preset classification models are the same as the training process for the land use classification model described above and are not further described here.

[0089] The above-mentioned data characteristics include the amount of data indicating the size of the data in the target area, the number of feature dimensions indicating whether the data includes feature information, the randomness that characterizes whether the data distribution is regular, and the computing resources required for data processing.

[0090] The randomness of whether the data distribution is regular can be understood as fitting the data to obtain multiple data distribution patterns. The higher the fitting degree, the more regular the data distribution is, and the corresponding randomness is lower.

[0091] In order to ensure that the model selection is more accurate and relevant to the current data, a correspondence between data features and preset classification models is established.

[0092] Exemplarily, if the data size of the target area is less than or equal to the first data size, the K-nearest neighbor classifier algorithm model is used. If the data size of the target area is greater than the first data size, the random forest classifier algorithm model and the extreme gradient boosting classifier algorithm model are used.

[0093] If the number of feature dimensions of the data in the target area is greater than or equal to the first dimension, the Random Forest Classifier and the Extreme Gradient Boosting Classifier are used. If the number of feature dimensions of the data in the target area is less than the first dimension, the K-Nearest Neighbor Classifier is used.

[0094] If the randomness of the data in the target area is greater than or equal to the first randomness, the K-nearest neighbor classifier algorithm model is used. If the randomness of the data in the target area is less than the first randomness, the random forest classifier algorithm model and the extreme gradient boosting classifier algorithm model are used.

[0095] If the computing resources required for data processing in the target area are greater than the first preset resource amount, the extreme gradient boosting classifier algorithm model is adopted; if the computing resources required for data processing are greater than the second preset resource amount and less than or equal to the first preset resource amount, the random forest classifier algorithm model is adopted; if the computing resources required for data processing are less than or equal to the second preset resource amount, the K nearest neighbor classifier algorithm model is adopted.

[0096] This application selects three classifier algorithms, KNN, RF and XGBoost in machine learning, to perform land use classification in order to further illustrate the classification model involved in this application.

[0097] On the one hand, by analyzing the classification accuracy of the three classifier algorithms, the consistency of the impact of feature addition on classification accuracy is tested; on the other hand, by comparing the classification results of the three classifier algorithms, the classification scheme with the highest classification accuracy is selected to output the classification results of the land use classification of the study area.

[0098] The first one uses KNN as a classifier as follows.

[0099] The KNN algorithm is a powerful nonparametric learning algorithm widely used in pattern recognition. Due to its simple implementation, ease of understanding, good predictive performance, insensitivity to outliers, and ability to determine the uniformity of sample distribution based on the algorithm's accuracy, KNN is often used alongside other machine learning algorithms in remote sensing image classification research. KNN is also an instance-based lazy algorithm model; it does not require pre-training a large number of samples to generate a classifier. Instead, it stores all available instances and measures the similarity between samples based on calculated distances. KNN's implementation principle: For a set of training samples, each sample contains features and a target variable identified as a categorical value. For a newly input sample without a categorical value, the features of that sample are compared with the features of every sample in the training set. The K most similar (close) samples are found, and the category with the most occurrences in these samples is used as the classification value for the new input prediction.

[0100] Specifically, KNN classification mainly includes the following steps.

[0101] Step 1: Sample data preparation.

[0102] Training sample set T = {(x1, y1), (x2, y2), ..., (x i ,y i),…,(x n ,y n )}, where n is the number of training samples, and the test sample set {(x1, y1), (x2, y2), …, (x j ,y j ),…,(x m ,y m )},x i ,x j ∈F d represents the eigenvector, y i ,y j ∈C={c1,c2,…,c l} represents the classification label, where the categories in the test sample set are used for post-classification cross-validation. Before training, we first quantize all features of all samples to make them comparable, quantizing non-numeric features to numeric values. Secondly, all features are normalized. Since there are multiple input feature parameters, each with its own domain and range of values, this will have a certain impact on the subsequent distance calculation. For example, parameters with larger values will have a greater impact than parameters with smaller values. Therefore, formula 2-1 is used to normalize all feature values of the samples.

[0103]

[0104] Step 2: Calculate the distance between each sample in the test set and the training set. KNN distance calculations typically use Euclidean distance or Manhattan distance. This article uses the Euclidean distance algorithm (Formula 2-2) to calculate the distance between sample data in a multidimensional feature space. After calculating the distance from all test samples to the training samples, sort all sample data in ascending order of distance.

[0105]

[0106] Step 3: Determine the K value. The selection of the K value is crucial in the classification process. A large or small K value will lead to over-normalization of the model or highlight local differences.

[0107] Step 4: Determine the category. Use the optimal K value as the number of nearest neighbors, count the frequency of each category among the K points, and use the category with the highest frequency as the category of the unknown point (test point).

[0108] The second method is to use RF as a classifier.

[0109] The RF classifier algorithm is a non-parametric pattern recognition method. It is an ensemble classifier based on decision trees and bagging. The classification prediction result is determined by voting based on the classification results of multiple decision trees. RF can handle input samples with high-dimensional features without the need for dimensionality reduction. It does not require prior assumptions about the data distribution. Even in the case of a limited number of samples, it can efficiently operate on large datasets and obtain good classification results. At the same time, it can evaluate the importance of input variables. Due to these advantages, random forests perform well in remote sensing data classification, such as multispectral data, hyperspectral data, LiDAR data, multi-source remote sensing data, etc.

[0110] Suppose there are N samples and M feature variables. RF randomly and with replacement extracts 2N / 3 independent sample data from the original training samples through the bootstrap sampling method to construct decision trees and generate a random forest; then randomly selects m features (m < M) with replacement as the basis for the branches of this tree, and determines the splitting of each node based on the Gini criterion. The node is split by the variable that provides the best split.

[0111] The remaining sample data is called out-of-bag (OOB) data, which is used to evaluate the error rate of the random forest and calculate the importance of each feature. Through multiple iterations, OOB gradually removes relatively poor features and selects the best forest. After the OOB predicts the results of all samples and compares them with the true values, the out-of-bag error rate (OOB estimate of error rate) of this forest can be obtained. The classification of new sample data is determined by the majority vote in the classification results of all constructed decision trees.

[0112] The following specifically explains the best split of each node. RF generates classification decision trees based on the Classification and Regression Tree (CART) algorithm. The CART algorithm mainly constructs a binary tree recursively. In terms of feature selection, it adopts the Gini Index (GI) minimization criterion. GI is an impurity splitting method, which represents the probability that a randomly selected sample in the sample set is misclassified. Suppose a given sample dataset T has a total of K classes, and the probability that a sample belongs to the k-th class is p k , then the GI of this probability distribution is:

[0113]

[0114] The formula satisfies the condition

[0115] The smaller the GI is, the smaller the uncertainty of the sample category is and the higher the purity is. When GI(T) = 0, it means that all samples at this node belong to the same category, indicating that the uncertainty of the sample is 0 and the tree splitting will stop.

[0116] The sample data set T of a child node i Divided into j parts T i ={1, 2…, j}, where N i For child node T i The number of samples, N is the number of samples in the set T, then the feature m at the child node i The GI of each split node will be calculated. The split GINI coefficient of the node is:

[0117]

[0118] For the sample set T, calculate the minimum GINI coefficient of each feature:

[0119]

[0120] The basic idea of GINI minimum segmentation is: for each feature, traverse all possible segmentation methods, and if the minimum GINI split , that is, the purity of the sample is the largest at this time, and this is used as the segmentation criterion at the node. The subtree divided by this feature is the optimal branch; then continue to split according to the specification with the smallest purity until the branch stopping rule is met and the growth stops. By setting the leaf node purity threshold, the leaf node stops growing when it is greater than or equal to the threshold.

[0121] Two key parameters in RF are the number of decision trees (ntree) and the number of features randomly selected by each tree at each node split (mtry). The ntree value is generally considered more important because a larger number of decision trees increases model complexity and reduces efficiency. For most RF applications, a good range for ntree is 0 to 1000, while the mtry value defaults to the square root of the number of input features.

[0122] The third method is to use XGBoost as a classifier as follows.

[0123] XGBoost is a type of boosting algorithm, belonging to the category of gradient boosted tree models. It is a gradient boosting algorithm based on decision trees. The XGBoost algorithm undergoes multiple rounds of data iteration, generating a weak classifier with each iteration. Each classifier is trained based on the classification residuals obtained in the previous iteration. Weak classifiers are required to be simple, have low variance, and high bias (such as the CART classifier, but linear classifiers are also supported). The training process continuously improves classification accuracy by reducing bias. The final classifier is generated by weighted summation of the weak classifiers obtained from each round of training using an additive model. The XGBoost algorithm has advantages such as regularization to reduce overfitting, parallel processing, customizable optimization objectives and evaluation criteria, and the ability to handle sparse and missing values, making it very valuable for remote sensing data processing. XGBoost has demonstrated high prediction accuracy and processing efficiency in remote sensing data processing and classification applications. Some researchers have also used XGBoost for land use / land cover classification, but there is still room for further research in classification within the context of multi-source remote sensing data fusion.

[0124] The implementation process of XGBoost is as follows:

[0125] For a given dataset with n samples and m features, the dataset D can be defined as {(x i ,y i )}(|D|=n,x i ∈R m ,y i ∈R),x i represents the i-th sample, y i Represents the label of the i-th sample. The algorithm uses K trees to calculate the predicted value of each tree for the sample, and adds the predicted values of each tree as the predicted value of the sample. The tree integration algorithm function is defined as follows:

[0126]

[0127] Where F is the space of decision tree F={f(x)=ω q(x)}(q:R m →T,ω∈R T ), q represents the model of the tree, that is, input a sample, map the sample to the leaf node according to the model and output the prediction score; ω q (x) represents the set of all leaf node scores of tree q; T is the number of leaf nodes in tree q. The regularized objective function defined by the learning model f(x) is as follows:

[0128]

[0129] Where, Represents the model prediction value; yi Represents the category label of the i-th sample; For sample x i Training error; Ω(f k ) represents the regularization term of the kth tree; T represents the number of leaf nodes of each tree; ω represents the set of scores of leaf nodes of each tree; γ and λ represent the regularization coefficients. In practical applications, these two parameters need to be adjusted. The goal is to obtain the corresponding model of L(φ) and f(x). The model is learned by using additive training, that is, the original model is kept unchanged each time, and the predicted value is added to a new f(x) in each iteration to reduce the objective function as much as possible. Let To calculate the predicted value of the i-th instance at the t-th iteration, we need to add f t To minimize the following objectives:

[0130]

[0131] Represents sample x i The final prediction value is the sum of the prediction value of the t-th tree and the prediction value of the first t-1 trees. Next, the first-order derivative g i and the second-order inverse h i The solution is as follows:

[0132]

[0133] Substituting into (Formula 2-12), we have:

[0134]

[0135] The constant term has no effect on the change of the minimum value of the objective function. We remove the constant term to simplify the calculation, and the objective function is simplified to the following formula:

[0136]

[0137] Next, define I j ={i|q(x i )=j} is an instance of leaf node j. Based on the above inference, we have the following formula:

[0138]

[0139] Where, ω j is the score of each tree's leaf node j, and ω of each tree j Add up to get the final prediction score, and find the optimal ω by minimizing the objective function j value, and ω in the above formula j Find the partial derivative and calculate the optimal weight value:

[0140]

[0141] Then the optimal value corresponding to tree q is:

[0142]

[0143] The above formula (2-15) can be used to evaluate the quality q of the tree, which is similar to the impurity evaluation score of the decision tree. In general, it is impossible to calculate all possible tree structures. XGBoost uses a greedy algorithm, starting with a single leaf node, and continuously adding branches to the tree through iteration. Suppose the set I is divided into the left leaf node I L and right leaf node I R , then the loss function after dividing the left and right leaf nodes is as follows:

[0144]

[0145] In practical applications, this formula is often used to evaluate the quality of a split tree structure. The XGBoost algorithm sorts feature samples, divides the features from small to large, compares the magnitude of the objective function after each division, and finds the feature with the largest decrease in the objective function, which is used as the optimal split point.

[0146] XGBoost is often used in conjunction with cross-validation and grid search, two crucial aspects of machine learning. The combination of cross-validation and grid search is the most common method for model optimization and parameter evaluation. In machine learning, the same dataset is used for both model training and estimation, leading to inaccurate model error estimates. Cross-validation can address this issue by estimating the generalization error closer to the true model performance. This is especially true when sufficient sample data is available, making the true generalization error estimate more accurate. However, in real-world applications, sample data is often insufficient, necessitating data reuse. Cross-validation randomly divides sample data into training, validation, and test sets. The training set is used for model training, the validation set is used for model estimation and optimization, and the test set is used to evaluate the final learning method. In practical applications, we often use the k-fold cross-validation method. Its basic principle is to evenly divide the original data into k groups. Each sub-dataset is used as validation data in turn, and the corresponding k-1 sub-datasets are used as training sets, thereby obtaining k models. Finally, the average classification accuracy of the validation sets of these k models is used as the performance indicator of the classifier under this k-fold cross-validation. K-fold cross-validation only needs to be repeated k times, which can greatly reduce the complexity of model calculations. In practical applications, k = 10 is generally a good empirical value. The grid search algorithm uses cross-validation to find the optimal model parameters. It is an exhaustive search method. Its implementation principle is to find the optimal result in an array. That is, among all candidate parameter choices, it iterates through all possibilities and tries every possibility. The best performing parameter is the final result, which can achieve automatic parameter adjustment.

[0147] In order to realize the above functions, the land use category classification device includes hardware structures and / or software modules corresponding to the execution of each function. It should be easily appreciated by those skilled in the art that, in combination with the algorithm steps of each example described in the embodiments disclosed herein, the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0148] The present application also provides a Figure 3 The land use category classification device shown includes: an extraction unit 301, a first classification unit 302, and a second classification unit 303.

[0149] The extraction unit 301 is used to extract the natural attribute feature information of the target area from the point cloud data and the image data, and to extract the social attribute feature information of the target area from the point of interest data of the target area.

[0150] The first classification unit 302 is used to divide the target area into multiple first contour areas according to natural attribute feature information; and to divide the target area into multiple second contour areas according to social attribute feature information; each first contour area corresponds to a first land feature category, and the first land feature category represents the natural attribute land feature category; the correlation degree of the natural attribute feature information between the pixel points in the first contour area is greater than or equal to the first correlation degree threshold; each second contour area corresponds to a second land feature category, and the second land feature category represents the social attribute land feature category.

[0151] The second classification unit 303 is used to perform secondary classification on the multiple first contour areas of the target area based on the multiple second contour areas to obtain a target category classification result of the target area; the target category classification result includes the first ground object category and / or the second ground object category.

[0152] As an embodiment, the first classification unit 302 is specifically used to determine the kernel density of the point of interest data at each position in the target area; refer to the preset kernel density range corresponding to different second feature categories, and divide the second target feature category corresponding to the target kernel density range to which the kernel density of the point of interest data at each position in the target area belongs; the second target feature category is any feature category in the second feature category; and the candidate area composed of each position in the target area that belongs to the same second target feature category and has continuous positions is determined as the second contour area corresponding to the same second feature category.

[0153] As another embodiment, the first classification unit 302 is specifically used to determine the natural attribute feature information of each position in the target area in multiple dimensions; divide multiple positions in the target area whose similarity of natural attribute feature information in multiple dimensions is higher than the similarity threshold corresponding to each dimension and whose positions are continuous into the same first contour area, and determine the first target land feature category corresponding to the same first contour area based on the natural attribute feature information of the same first contour area; the first target land feature category is any land feature category in the first land feature category.

[0154] As another embodiment, the second classification unit 303 is specifically used to determine whether there is an overlapping area between the first contour area and the second contour area of the target area; if the category attributes of the first target object category of the first contour area corresponding to the overlapping area and the second target object category of the second contour area are the same, and the granularity priority of the first target object category is higher than that of the second target object category, then the overlapping area in the first contour area is classified according to the second target object category of the overlapping area; if the category attributes of the first target object category of the first contour area corresponding to the overlapping area and the second target object category of the second contour area are the same, and the granularity priority of the first target object category is lower than that of the second target object category, then the first target object category of the overlapping area in the first contour area is retained; if the category attributes of the first target object category of the first contour area corresponding to the overlapping area and the second target object category of the second contour area are different, then the overlapping area in the first contour area is classified according to the second target object category of the overlapping area.

[0155] As another embodiment, the second classification unit 302 is also used to input the natural attribute feature information and social attribute feature information of each location in the target area into multiple different land use classification models to obtain different land feature category classification results; according to the frequency of the land feature category belonging to the same location in different land feature category classification results, the different land feature category classification results are evaluated to determine the land feature category classification result with the highest frequency in the evaluation results as the target category classification result; wherein, the first land feature category includes one or more of the following: building land, bare land, cultivated land, grassland, road, woodland and water body; the second land feature category includes one or more of the following: residential land, commercial land, industrial land, public service and management land, education and scientific research land, green land and square land.

[0156] As another embodiment, the first classification unit 302 is specifically used to: determine a first target model from multiple first preset classification models based on data features of point cloud data and image data of the target area; input natural attribute feature information of the target area into the first target model, and divide the target area into multiple first contour areas.

[0157] As another embodiment, the first classification unit 302 is specifically used to: determine a second target model from multiple second preset classification models based on the data characteristics of the points of interest in the target area; input the social attribute feature information of the target area into the second target model, and divide the target area into multiple second contour areas.

[0158] As another embodiment, data features include the amount of data indicating the size of the data, the number of feature dimensions indicating the high or low feature dimensions of the feature information included in the data, the degree of randomness indicating whether the data distribution is regular, and the computing resources required for data processing.

[0159] As another embodiment, the natural attribute feature information includes indirect feature information, direct feature information, spectral feature information and image space feature information; the extraction unit 301 is specifically used to: obtain point cloud data of the target area from the laser radar scanner of the airborne LiDAR measurement system, and image data of the target area from the airborne digital camera; the resolution of the airborne digital camera is higher than the resolution threshold; extract the indirect feature information and direct feature information of each pixel point corresponding to each position in the target area from the point cloud data, and extract the spectral feature information and image space feature information of each pixel point corresponding to each position from the image data; wherein the direct feature information includes elevation feature information and intensity feature information; the indirect feature information includes spatial geometric feature information of the point cloud and texture feature information used to describe the correlation between the grayscale of two points within a preset distance and in a preset direction; the spatial geometric feature information includes the undulation of the ground object on the surface, the flatness of the ground object surface and the discreteness of the point cloud; the image space feature information includes: morphological building index, morphological shadow index and urban complexity.

[0160] Regarding the device in the above embodiment, the specific manner in which each unit module performs the operation has been described in detail in the embodiment of the method, and will not be elaborated here.

[0161] Figure 4 This is a schematic diagram of a computer device provided by this application. Figure 4 The computer device 60 may include at least one processor 601 and a memory 603 for storing processor-executable instructions. The processor 601 is configured to execute the instructions in the memory 603 to implement the land use classification method in the following embodiment.

[0162] In addition, the computer device 60 may also include a communication bus 602 , at least one communication interface 604 , an input device 606 , and an output device 605 .

[0163] The processor 601 may be a central processing unit (CPU), a microprocessor unit, an ASIC, or one or more integrated circuits for controlling the execution of the program of the present application.

[0164] The communication bus 602 may include a pathway for transmitting information between the aforementioned components.

[0165] The communication interface 604 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area network (WLAN), etc.

[0166] The input device 606 is used to receive input signals and the output device 605 is used to output signals.

[0167] The memory 603 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, an optical disc storage (including a compressed optical disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program codes in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory may be independent and connected to the processing unit via a bus. The memory may also be integrated with the processing unit.

[0168] The memory 603 is used to store instructions for executing the solution of the present application, and the execution is controlled by the processor 601. The processor 601 is used to execute the instructions stored in the memory 603, thereby realizing the functions of the method of the present application.

[0169] In a specific implementation, as an embodiment, the processor 601 may include one or more CPUs, such as Figure 4 CPU0 and CPU1 in.

[0170] In a specific implementation, as an embodiment, the computer device 60 may include multiple processors, such as Figure 4 6 and 607. Each of these processors may be a single-CPU processor or a multi-CPU processor. A processor herein may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0171] The computer equipment Figure 4 The system includes a processor 601 and a memory 603 for storing executable instructions of the processor 601. The processor 601 is configured to execute the executable instructions to implement a land use classification method according to any of the above possible implementations. The system can achieve the same technical effects, and to avoid repetition, the description is omitted here.

[0172] The present application also provides a computer-readable storage medium. When the instructions in the computer-readable storage medium are executed by a processor of a control device or a control apparatus, the control device or the control apparatus can perform the land use classification method according to any of the above possible implementations. To avoid repetition, the above description is omitted.

[0173] The present application also provides a computer program product, including a computer program or instructions, which is executed by a processor to implement any of the above-described possible implementations of the land use classification method. The computer program or instructions can achieve the same technical effects, and to avoid repetition, they are not described here.

[0174] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0175] It should be understood that the present application is not limited to the exact structures described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A land use classification method, characterized in that: The method comprises: Extracting natural attribute feature information of a target area from the point cloud data and the image data, and extracting social attribute feature information of the target area from the point of interest data of the target area; According to the natural attribute feature information, the target area is divided into a plurality of first contour areas; and according to the social attribute feature information, the target area is divided into a plurality of second contour areas; each of the first contour areas corresponds to a first feature category, and the first feature category represents a natural attribute feature category; the correlation degree of the natural attribute feature information between the pixels in the first contour area is greater than or equal to a first correlation degree threshold; each of the second contour areas corresponds to a second feature category, and the second feature category represents a social attribute feature category; Based on the multiple second contour areas, the multiple first contour areas of the target area are secondary classified to obtain a target category classification result of the target area; the target category classification result includes the first ground feature category and / or the second ground feature category.

2. The method according to claim 1, characterized in that The step of dividing the target area into a plurality of second contour areas according to the social attribute feature information includes: Determine the kernel density of the point of interest data at each location in the target area; Referring to the preset kernel density ranges corresponding to different second object categories, the kernel density of the point of interest data at each position in the target area is divided into the second target object categories corresponding to the target kernel density range; the second target object category is any object category among the second object categories; A candidate area consisting of positions in the target area that belong to the same second target object category and are continuous in position is determined as the second contour area corresponding to the same second object category.

3. The method according to claim 1, characterized in that The step of dividing the target area into a plurality of first contour areas according to the natural attribute feature information includes: Determining the natural attribute feature information of each location in the target area in multiple dimensions; Multiple locations in the target area whose similarity of the natural attribute feature information in multiple dimensions is higher than the similarity threshold corresponding to each dimension and whose positions are continuous are divided into the same first contour area, and the first target land object category corresponding to the same first contour area is determined based on the natural attribute feature information of the same first contour area; the first target land object category is any land object category in the first land object categories.

4. The method according to claim 1, wherein The performing secondary classification on the plurality of first contour areas of the target area based on the plurality of second contour areas to obtain a target category classification result of the target area includes: determining that there is an overlapping area between the first contour area and the second contour area of the target area; If the first target object category of the first contour area and the second target object category of the second contour area corresponding to the repeated area have the same category attributes, and the granularity priority of the first target object category is higher than that of the second target object category, then classify the repeated area in the first contour area according to the second target object category of the repeated area; If the first target object category of the first contour area and the second target object category of the second contour area corresponding to the repeated area have the same category attributes, and the granularity priority of the first target object category is lower than that of the second target object category, then retain the first target object category of the repeated area in the first contour area; If the first target object category of the first contour area and the second target object category of the second contour area corresponding to the repeated area have different category attributes, the repeated area in the first contour area is classified according to the second target object category of the repeated area.

5. The method according to claim 1, wherein The method further comprises: Inputting the natural attribute characteristic information and the social attribute characteristic information of each location in the target area into the multiple different land use classification models to obtain classification results of different land features; Evaluate the different ground feature classification results according to the frequencies of the ground feature categories belonging to the same position in the different ground feature classification results, and determine the ground feature classification result with the highest frequency in the evaluation results as the target classification result; Among them, the first land feature category includes one or more of the following: building land, bare land, cultivated land, grassland, road, woodland and water body; the second land feature category includes one or more of the following: residential land, commercial land, industrial land, public service and management land, education and scientific research land, green land and square land.

6. The method according to claim 1, characterized in that The step of dividing the target area into a plurality of first contour areas according to the natural attribute feature information includes: Determining a first target model from a plurality of first preset classification models according to data features of the point cloud data and the image data of the target area; The natural attribute feature information of the target area is input into the first target model, and the target area is divided into the plurality of first contour areas.

7. The method according to claim 1, characterized in that The step of dividing the target area into a plurality of second contour areas according to the social attribute feature information includes: determining a second target model from a plurality of second preset classification models according to data features of the points of interest in the target area; The social attribute feature information of the target area is input into the second target model, and the target area is divided into the plurality of second contour areas.

8. The method according to claim 6 or 7, characterized in that The data characteristics include the amount of data indicating the size of the data, the number of feature dimensions indicating whether the data includes feature information, the randomness that characterizes whether the data distribution is regular, and the computing resources required for data processing.

9. The method according to any one of claims 1 to 7, characterized in that The natural attribute feature information includes indirect feature information, direct feature information, spectral feature information and image space feature information; The natural attribute feature information of the target area is extracted from the point cloud data and the image data, including: Acquire point cloud data of the target area from a laser radar scanner of an airborne LiDAR measurement system, and image data of the target area from an airborne digital camera; the resolution of the airborne digital camera is higher than a resolution threshold; Extracting indirect feature information and direct feature information of each pixel point corresponding to each position in the target area from the point cloud data, and extracting spectral feature information and image space feature information of each pixel point corresponding to each position from the image data; Among them, the direct feature information includes elevation feature information and intensity feature information; the indirect feature information includes spatial geometric feature information of the point cloud and texture feature information used to describe the correlation between the grayscale of two points within a preset distance and in a preset direction; the spatial geometric feature information includes the undulation of the ground object on the surface, the flatness of the ground object surface and the discreteness of the point cloud; the image spatial feature information includes: morphological building index, morphological shadow index and urban complexity.

10. A land use classification device, characterized in that: The device comprises: An extraction unit, configured to extract natural attribute feature information of a target area from the point cloud data and the image data, and to extract social attribute feature information of the target area from the point of interest data of the target area; a first classification unit, configured to divide the target area into a plurality of first contour areas according to the natural attribute feature information; and to divide the target area into a plurality of second contour areas according to the social attribute feature information; each of the first contour areas corresponds to a first feature category, the first feature category representing a natural attribute feature category; a correlation degree of the natural attribute feature information between pixels in the first contour area is greater than or equal to a first correlation degree threshold; each of the second contour areas corresponds to a second feature category, the second feature category representing a social attribute feature category; The second classification unit is used to perform secondary classification on the multiple first contour areas of the target area based on the multiple second contour areas to obtain a target category classification result of the target area; the target category classification result includes the first ground feature category and / or the second ground feature category.

Images