A high-strength piece area building form and three-dimensional visible green quantity collaborative optimization method, system, terminal and storage medium based on multi-view window scene green vision evaluation

Abstract

The application belongs to the technical field of data processing, and discloses a high-intensity plot area building form and three-dimensional visible green quantity collaborative optimization method, system, terminal and storage medium based on multi-view window view green evaluation, the method comprises the following steps: performing semantic classification and three-dimensional model standardization processing on three-dimensional scene data of a target plot area to obtain a standardized three-dimensional model; performing grid processing on an effective building facade to obtain a window view point; constructing a parameterized camera and a pixel ray according to the window view point, performing ray detection and classified statistics on the standardized three-dimensional model to obtain a window view green evaluation result; generating multiple optimization scenarios according to the window view green evaluation result, re-executing multi-view window view green evaluation until a preset window view green target is met, and outputting an optimal collaborative optimization scheme. The application improves the collaborative regulation and control capability of three-dimensional green quantity and building form in a high-intensity plot area, improves the calculation efficiency of plot-level multi-view, and realizes the tracing and mapping of cross-platform data.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, system, terminal, and computer-readable storage medium for the collaborative optimization of high-intensity regional building form and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation. Background Technology

[0002] High-intensity development zones are characterized by high development intensity, high building density, complex spatial hierarchy, scarce green space resources, and a high concentration of vertical users. Within these zones, significant obstruction and visual coupling exist between natural elements such as greenery, water bodies, and the sky, and building volumes, podiums, towers, platforms, and circulation spaces. The same green space area can result in drastically different window views and natural visual experiences depending on its location, height, and building form. Therefore, it is necessary to use building window views as a visual medium to establish a calculable relationship between standardized viewpoints on building facades and three-dimensional scene geometry, viewing distance structure, proportions of natural elements, and spatial optimization measures.

[0003] However, existing window view and green view evaluation methods mostly use the entire image as the statistical object, usually directly calculating the area proportion of elements such as greenery, sky, buildings, or water bodies in the field of view, without further identifying the viewing distance type of these elements, i.e., lacking refined viewing distance analysis. Using only the overall element proportion cannot fully characterize the spatial structure of the window view, making it difficult to accurately identify the specific reasons for insufficient window view quality, and also making it difficult to propose targeted spatial improvement measures, resulting in insufficient ability to coordinate and regulate three-dimensional green volume and building form in high-intensity areas. Furthermore, existing window view and green view evaluation methods also suffer from low computational efficiency at the area-level multi-viewpoint level and difficulties in continuous data transfer across platforms.

[0004] Therefore, existing technologies still need to be improved and developed. Summary of the Invention

[0005] The main objective of this invention is to provide a method, system, terminal, and computer-readable storage medium for the coordinated optimization of building form and three-dimensional visible green volume in high-intensity areas based on multi-viewpoint window view green vision evaluation. This aims to solve the problem that existing window view and green vision evaluation methods lack the identification of viewing distance type, resulting in insufficient ability to coordinate and regulate three-dimensional green volume and building form in high-intensity areas.

[0006] To achieve the above-mentioned objectives, this invention provides a method for the coordinated optimization of building form and three-dimensional visible green volume in high-intensity areas based on multi-viewpoint window view green vision evaluation. The method includes: Acquire 3D scene data of the target area, perform semantic classification and 3D model standardization processing on the 3D scene data, and obtain a standardized 3D model with semantic category identifiers; Based on the standardized 3D model, effective building facades are selected and then gridded to obtain multiple window viewpoints with unique viewpoint numbers. Based on the window viewpoint, a parametric camera and pixel rays are constructed. Based on the parametric camera and pixel rays, ray detection and classification statistics are performed on the standardized 3D model to obtain the window view green view evaluation results. Based on the window view green vision evaluation results, multiple optimization scenarios are generated. The multi-viewpoint window view green vision evaluation is re-executed for multiple optimization scenarios until the preset window view green vision target is met or the maximum number of iterations is reached, and the best collaborative optimization scheme is output.

[0007] Optionally, the step of acquiring the 3D scene data of the target area, performing semantic classification and 3D model standardization processing on the 3D scene data to obtain a standardized 3D model with semantic category identifiers specifically includes: Acquire 3D scene data of the target area, wherein the target area is a high-intensity area; The three-dimensional scene data is subjected to semantic classification and three-dimensional model standardization processing to obtain a standardized three-dimensional model with semantic category identifiers. The semantic category identifiers include buildings, plants, water bodies, ground, and sky. The buildings include the observation building and surrounding buildings.

[0008] Optionally, the step of selecting effective building facades based on the standardized 3D model, and then meshing the effective building facades to obtain multiple window viewpoints with unique viewpoint numbers, specifically includes: Based on the standardized three-dimensional model, multiple building surfaces of the observed building are obtained; Based on the normal direction, planar features, and geometric dimensions of the multiple building surfaces, the multiple building surfaces are screened, and the building surfaces that simultaneously meet the vertical sampling requirements, planar sampling requirements, and dimensional sampling requirements are selected as valid building facades; The effective building facade is gridded according to the set facade grid size to obtain multiple grids and multiple window viewpoints corresponding to the multiple grids. A unique viewpoint number is established for each window viewpoint. ; in, Indicates the viewpoint number. Indicates the building number. Indicates the facade number. Indicates the floor number. Indicates the grid number.

[0009] Optionally, the step of constructing a parametric camera and pixel rays based on the window viewpoint, and performing ray detection and classification statistics on the standardized 3D model based on the parametric camera and pixel rays to obtain the window view green view evaluation result specifically includes: Using the three-dimensional spatial coordinates of the window viewpoint as the camera position, a unified parametric camera is constructed. The parameters of the parametric camera include image resolution, image aspect ratio, focal length, and maximum viewing distance. Based on the parametric camera, a frustum target plane is set at the maximum viewing distance, and the frustum target plane is divided into multiple pixel grids according to the image resolution. The vectors pointing from the window viewpoint to the center of each pixel corresponding to each pixel grid constitute pixel rays. The mesh is split into triangular faces with semantic category identifiers. An axis-aligned bounding box is created for each triangular face. Based on the position and scale of the axis-aligned bounding box, a bounding box hierarchy is recursively constructed. The nodes of the bounding box hierarchy store the bounding box boundary, the left and right child node indices, or the triangular face index range contained in the leaf node. Based on the bounding box hierarchy, a parallel computing device is used to perform the nearest hit detection on the pixel ray in parallel to obtain the nearest hit result. The nearest hit result includes the hit category and the hit distance. The hit category is the semantic category identifier of the triangle face hit by the pixel ray. Pixel rays that do not hit any triangle face are treated as pixel rays with the hit category of sky; Based on the number of pixel rays that hit the category of building, the number of pixel rays that hit the category of plant, the number of pixel rays that hit the category of water, the number of pixel rays that hit the category of ground, the number of pixel rays that hit the category of sky, and the total number of pixel rays, the building ratio, plant ratio, water ratio, ground ratio, and sky ratio are calculated respectively, and the plant ratio is used as the window view green view rate. Based on the hit distance, near distance threshold and far distance threshold corresponding to each pixel ray, the viewing distance zone corresponding to each pixel ray is determined. The viewing distance zone includes near, medium and far. Pixel rays that do not hit any triangular face belong to the far. Based on the number of pixel rays corresponding to the foreground, the number of pixel rays corresponding to the midground, the number of pixel rays corresponding to the background, and the total number of pixel rays, the proportions of foreground, midground, and background are calculated respectively. The proportions of near view, medium view, and far view are compared with the effective proportion thresholds to obtain the window view distance type. Based on the green view rate of the window view and the suitable range of green view rate corresponding to the window view distance type, the green view quality value of the window view is calculated, and the green view level of the window view is obtained based on the green view quality value of the window view. The green view quality value and the green view level of the window view constitute the green view evaluation result of the window view.

[0010] Optionally, the step of calculating the window view green view quality value based on the window view green view rate and the suitable range of green view rate corresponding to the window view viewing distance type specifically includes: Based on the green view rate of the window view and the suitable range of green view rate corresponding to the window view distance type, the green view quality value of the window view is calculated: ; ; ; in, Indicates the viewpoint of the window. The green view quality value of the window view. Indicates the viewpoint of the window. Green view rate of the window view Indicates the type of window view distance The corresponding suitable range of green visibility The lower limit, Indicates the type of window view distance The corresponding suitable range of green visibility The upper limit.

[0011] Optionally, the step of generating multiple optimized scenarios based on the window view green view evaluation results, re-executing the multi-viewpoint window view green view evaluation on multiple optimized scenarios until the preset window view green view target is met or the maximum number of iterations is reached, and outputting the best collaborative optimization scheme, specifically includes: The window view green view evaluation results are aggregated by area, building, facade, floor and viewing distance type to identify priority improvement locations for low-quality window views; For the priority improvement locations, a set of candidate measures including building form adjustment and three-dimensional visible green volume configuration is established. Based on the set of candidate measures, multiple optimization scenarios are generated under planning control and limited green volume constraints. The multi-viewpoint window view green view evaluation was re-executed for multiple optimized scenarios to obtain multiple optimized window view green view evaluation results. If all the optimized window view green view evaluation results fail to meet the preset window view green view target and the maximum number of iterations is not reached, then the optimized scene is regenerated and the multi-viewpoint window view green view evaluation is re-executed. If the optimized window view green view evaluation result meets the preset window view green view target or reaches the maximum number of iterations, the iteration is terminated and the best collaborative optimization scheme is output.

[0012] Optionally, the method for synergistic optimization of high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation further includes: Based on the standardized 3D model with semantic category identifiers, a semantically tinted window view image corresponding to the window viewpoint is generated; The pixel colors of the semantically tinted window image are converted from the RGB color space to the HSV color space to obtain the converted pixel colors; Based on the converted pixel color and the preset semantic colors of buildings, plants, water bodies and ground, determine the semantic category of the pixel corresponding to the converted pixel color, and classify the pixels that do not match any semantic category into the semantic category of sky; Based on the number of pixels corresponding to buildings, plants, water bodies, ground, sky, and the total number of pixels, the reference building proportion, reference plant proportion, reference water body proportion, reference ground proportion, and reference sky proportion are calculated respectively. The accuracy of the reference building proportion, reference plant proportion, reference water body proportion, reference ground proportion, and reference sky proportion is verified by performing a viewpoint-by-view alignment comparison with the building proportion, plant proportion, water body proportion, ground proportion, and sky proportion.

[0013] To achieve the above-mentioned objectives, this invention also provides a high-intensity regional building form and three-dimensional visible green volume collaborative optimization system based on multi-viewpoint window view green vision evaluation. The high-intensity regional building form and three-dimensional visible green volume collaborative optimization system based on multi-viewpoint window view green vision evaluation includes: The preprocessing module is used to acquire the three-dimensional scene data of the target area, perform semantic classification and three-dimensional model standardization processing on the three-dimensional scene data, and obtain a standardized three-dimensional model with semantic category identifiers. The gridding module is used to filter out effective building facades based on the standardized 3D model, and to perform gridding processing on the effective building facades to obtain multiple window viewpoints with unique viewpoint numbers. The window view green view evaluation module is used to construct a parametric camera and pixel rays based on the window viewpoint, and to perform ray detection and classification statistics on the standardized 3D model based on the parametric camera and pixel rays to obtain the window view green view evaluation result. The collaborative optimization module is used to generate multiple optimization scenarios based on the window view green view evaluation results, re-execute the multi-viewpoint window view green view evaluation on the multiple optimization scenarios until the preset window view green view target is met or the maximum number of iterations is reached, and output the best collaborative optimization scheme.

[0014] To achieve the above-mentioned objectives, the present invention also provides a terminal, the terminal comprising: a memory, a processor, and a program for co-optimizing high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window green view evaluation, stored in the memory and executable on the processor. When the processor executes the program for co-optimizing high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window green view evaluation, it implements the steps of the method for co-optimizing high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window green view evaluation as described above.

[0015] To achieve the above-mentioned objectives, the present invention also provides a computer-readable storage medium storing a program for the collaborative optimization of high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window green view evaluation. When the program is executed by a processor, it implements the steps of the method for the collaborative optimization of high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window green view evaluation as described above.

[0016] In this invention, 3D scene data of a target area is acquired, and the 3D scene data is semantically classified and standardized to obtain a standardized 3D model with semantic category identifiers. Effective building facades are selected based on the standardized 3D model, and these facades are meshed to obtain multiple window viewpoints with unique viewpoint numbers. Parametric cameras and pixel rays are constructed based on these window viewpoints. Ray detection and classification statistics are performed on the standardized 3D model based on the parametric cameras and pixel rays to obtain window view greening evaluation results. Multiple optimization scenarios are generated based on the window view greening evaluation results. Multi-viewpoint window view greening evaluation is re-executed for these multiple optimization scenarios until a preset window view greening target is met or the maximum number of iterations is reached, outputting the optimal collaborative optimization scheme. This invention improves the collaborative control capability between 3D green volume and building form in high-intensity areas through view distance type recognition, achieving a quantitative improvement in visual benefits under limited green volume conditions. It improves the computational efficiency of multi-viewpoints at the area level through parallel ray detection and bounding box hierarchical structure. Unique viewpoint numbers enable cross-platform data traceability and mapping. Attached Figure Description

[0017] Figure 1 This is a flowchart of a preferred embodiment of the present invention, which is a method for synergistic optimization of high-intensity regional building form and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation. Figure 2 This is another flowchart of a preferred embodiment of the present invention, which is a method for synergistic optimization of high-intensity regional building form and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation. Figure 3This is a structural diagram of a preferred embodiment of the high-intensity regional building form and three-dimensional visible green volume collaborative optimization system based on multi-viewpoint window view green vision evaluation of the present invention; Figure 4 This is a structural diagram of a preferred embodiment of the terminal of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] As urban development shifts from incremental expansion to improving existing infrastructure, high-intensity development zones are gradually becoming the main spatial carriers for the highly integrated functions of business offices, public services, transportation hubs, and mixed-use living spaces. These zones are characterized by high development intensity, high building density, complex spatial hierarchy, scarce green space resources, and a high concentration of vertical users. While the ground-level green space ratio reflects the area of ​​green space on a planar scale, it cannot directly indicate how much natural scenery users can see from different buildings, with different orientations, on different floors, and at different viewing distances, nor can it reflect the visual service benefits of the same amount of green space in three-dimensional space.

[0020] In high-density development areas, natural elements such as greenery, water bodies, and the sky exhibit significant shading and visual coupling with building volumes, podiums, towers, platforms, and circulation spaces. The same green space area can result in completely different window-view greening rates and natural visual experiences depending on its location, height, and building form. Therefore, it is necessary to use building window views as a visual medium to establish a calculable relationship between standardized viewpoints on building facades and three-dimensional scene geometry, viewing distance structure, proportions of natural elements, and spatial optimization measures.

[0021] Existing urban spatial evaluation and green space planning technologies mainly include two-dimensional planning indicator control, three-dimensional model visualization, single-point view analysis, and image semantic recognition. Two-dimensional planning indicators typically use parameters such as green space ratio, plot ratio, building density, building height, and open space ratio for scheme control; three-dimensional models or urban information models can be used to express the spatial location of buildings and green spaces; GIS (Geographic Information System) and related plugins can perform local view analysis, sunlight, ventilation, or landscape visibility analysis; image segmentation or color classification technologies can extract the proportions of elements such as greenery, sky, and buildings from real-world photographs and rendered images.

[0022] The aforementioned technologies can address issues related to planar control, 3D representation, single-point visibility, or image element recognition, but they are typically independent of each other. There is a lack of direct mapping between 2D indicators and the actual window views of facade users; single-view tools struggle to reliably handle a large number of building facade viewpoints at the area scale; and 3D simulation results often exhibit data gaps with subsequent image analysis and spatial optimization platforms.

[0023] In particular, existing methods for evaluating window views and green views often use the entire image as the statistical object, typically calculating the area proportion of elements such as greenery, sky, buildings, or water bodies in the field of view without further determining the viewing distance type of these elements. Using only the overall element proportion cannot fully characterize the spatial structure of a window view. Even if two window views have the same greenery proportion, the resulting occlusion relationships, spatial hierarchy, and natural visual effects may differ significantly depending on whether the greenery is concentrated in the foreground, middle ground, or background. For example, near-field greenery may have a higher image proportion but may also create continuous occlusion; far-field greenery may have a relatively lower image proportion but may form a more complete spatial hierarchy together with the open sky and urban background. Therefore, without first determining the viewing distance structure and type, directly evaluating based on the greenery proportion will make it difficult to accurately identify the specific reasons for insufficient window view quality and to propose targeted spatial improvement measures.

[0024] Therefore, the core deficiency of existing technologies lies in the lack of a unified technological closed loop that uses multiple window viewpoints as basic units to simultaneously characterize building form obstruction, viewing distance type, three-dimensional visible green space, and window view quality, and further drives spatial adjustment. This is mainly manifested in the following aspects: (1) Existing methods mainly rely on two-dimensional plane indicators such as green space ratio and plot ratio, which make it difficult to quantify the dynamic interaction effect between green space volume and building form under different building, facade, floor and viewing distance conditions; (2) Existing image analysis methods usually directly count the proportion of elements such as greenery, sky, buildings and water in the whole image, lacking the identification of the visual distance of the elements and unable to determine the compositional relationship of foreground, middle ground and background. Moreover, existing evaluation methods fail to set or match evaluation criteria according to different viewing distance types, thus making it difficult to distinguish window views with "similar proportions of elements but different spatial levels"; (3) Single three-dimensional view tools are usually designed for a small number of observation points. The computational efficiency of multi-building, multi-facade and large number of window viewpoints at the area level is insufficient, and it is difficult to form stable and traceable building-facade-floor-viewpoint data numbering. (4) Ray analysis, image segmentation, 3D modeling and statistical evaluation are usually scattered on different platforms. The data format, viewpoint direction, camera parameters and result number are not uniform, which can easily lead to repeated modeling, repeated correction and the inability of ray results to correspond to image results point by point. (5) Under limited green space conditions, existing technologies lack a unified quantitative basis for the window view visual benefits produced by different greening locations, heights, forms and configurations, making it difficult to compare the improvement efficiency of unit green space input on different building, floor and viewing distance types.

[0025] To address the aforementioned technical problems, this invention provides a method for the collaborative optimization of high-intensity regional building morphology and 3D visible green volume based on multi-viewpoint window view green vision evaluation. The method involves acquiring 3D scene data of the target region, performing semantic classification and 3D model standardization on the 3D scene data to obtain a standardized 3D model with semantic category identifiers, filtering effective building facades based on the standardized 3D model, and then meshing these effective building facades to obtain multiple window viewpoints with unique viewpoint numbers. Parametric cameras and pixel rays are constructed based on these window viewpoints. Ray detection and classification statistics are performed on the standardized 3D model based on the parametric cameras and pixel rays to obtain window view green vision evaluation results. Multiple optimization scenarios are generated based on the window view green vision evaluation results. The multi-viewpoint window view green vision evaluation is re-executed on these multiple optimization scenarios until a preset window view green vision target is met or the maximum number of iterations is reached, outputting the optimal collaborative optimization scheme. This invention enhances the ability to coordinate the three-dimensional green volume and building form in high-intensity areas by recognizing the view distance type, thereby quantitatively improving the visual benefits under limited green volume conditions; it improves the computational efficiency of multi-viewpoints at the area level through parallel ray detection and bounding box hierarchical structure; and it enables cross-platform data traceability and mapping through unique viewpoint numbering.

[0026] The application content will be further explained below with reference to the accompanying drawings and the description of the embodiments.

[0027] A preferred embodiment of the present invention is a method for the coordinated optimization of high-intensity regional building morphology and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation, such as... Figure 1 and Figure 2 As shown, it specifically includes: S1. Obtain the 3D scene data of the target area, perform semantic classification and 3D model standardization processing on the 3D scene data, and obtain a standardized 3D model with semantic category identifier.

[0028] In one implementation of this embodiment, the step of acquiring the 3D scene data of the target area, performing semantic classification and 3D model standardization processing on the 3D scene data to obtain a standardized 3D model with semantic category identifiers specifically includes: Acquire 3D scene data of the target area, wherein the target area is a high-intensity area; The three-dimensional scene data is subjected to semantic classification and three-dimensional model standardization processing to obtain a standardized three-dimensional model with semantic category identifiers. The semantic category identifiers include buildings, plants, water bodies, ground, and sky. The buildings include the observation building and surrounding buildings.

[0029] Specifically, this invention first acquires the runtime configuration, the target area's 3D scene, and window view analysis parameters. Then, it performs validity checks on the input paths, model units, and scene geometry, and completes semantic classification and mesh standardization of environmental elements. The runtime configuration includes at least a task segmentation directory, an image storage directory, a script or calculation program directory, a 3D modeling program path, and semantic colors for environmental elements. These paths are used for data exchange between different calculation stages, while semantic colors are used to establish the correspondence between 3D elements such as buildings, plants, water bodies, and ground, and the pixel categories of the screenshots. The target area's 3D scene includes at least the observation building, scene buildings (i.e., surrounding buildings), plants, water bodies, and ground. The observation building is the building whose facade needs to generate window viewpoints; the scene buildings are all buildings involved in ray occlusion calculations and screenshot representation. When the observation building itself needs to act as an occlusion object for other viewpoints, it is simultaneously included in the scene building set. Plants may include trees, shrubs, hedges, and other greening volumes; water bodies may include rivers, lakes, ponds, and artificial water features; and ground may include roads, squares, platforms, and terrain bases.

[0030] S2. Based on the standardized three-dimensional model, select the effective building facades, and perform grid processing on the effective building facades to obtain multiple window viewpoints with unique viewpoint numbers.

[0031] In one implementation of this embodiment, the step of selecting effective building facades based on the standardized 3D model, and then performing meshing on the effective building facades to obtain multiple window viewpoints with unique viewpoint numbers, specifically includes: Based on the standardized three-dimensional model, multiple building surfaces of the observed building are obtained; Based on the normal direction, planar features, and geometric dimensions of the multiple building surfaces, the multiple building surfaces are screened, and the building surfaces that simultaneously meet the vertical sampling requirements, planar sampling requirements, and dimensional sampling requirements are selected as valid building facades; The effective building facade is gridded according to the set facade grid size to obtain multiple grids and multiple window viewpoints corresponding to the multiple grids. A unique viewpoint number is established for each window viewpoint. ; in, Indicates the viewpoint number. Indicates the building number. Indicates the facade number. Indicates the floor number. Indicates the grid number.

[0032] Specifically, this invention filters the effective facades of observed buildings, grids the facades according to a set scale, and generates window viewpoints with stable spatial numbers in batches.

[0033] Furthermore, this invention uses standardized viewpoints on building facades as the basic computational units for window view analysis. For each building surface observed, its normal direction, planar features, and geometric dimensions are first calculated, and valid facades that are approximately vertical, approximately planar, and whose dimensions meet the sampling requirements are selected. Roofs, bases, surfaces with excessive tilt, obvious distortion, or effective dimensions smaller than the target grid size are not included in regular sampling. For valid facades that meet the conditions, they are meshed according to the set facade grid size. After obtaining the grid center, floor clustering is performed based on the center point elevation, and within the same floor, they are sorted according to the local horizontal direction of the facade. Each window viewpoint is assigned a unique number, which is simultaneously written into the ray analysis results, screenshot file name, facade coloring record, quality evaluation data, and optimized location list, so that any window view result can be traced back to a specific building, facade, floor, and window pane.

[0034] S3. Construct a parametric camera and pixel rays based on the window viewpoint. Based on the parametric camera and pixel rays, perform ray detection and classification statistics on the standardized 3D model to obtain the window view green view evaluation result.

[0035] In one implementation of this embodiment, the step of constructing a parametric camera and pixel rays based on the window viewpoint, and performing ray detection and classification statistics on the standardized 3D model based on the parametric camera and pixel rays to obtain the window view green view evaluation result specifically includes: Using the three-dimensional spatial coordinates of the window viewpoint as the camera position, a unified parametric camera is constructed. The parameters of the parametric camera include image resolution, image aspect ratio, focal length, and maximum viewing distance. Based on the parametric camera, a frustum target plane is set at the maximum viewing distance, and the frustum target plane is divided into multiple pixel grids according to the image resolution. The vectors pointing from the window viewpoint to the center of each pixel corresponding to each pixel grid constitute pixel rays. The mesh is split into triangular faces with semantic category identifiers. An axis-aligned bounding box is created for each triangular face. Based on the position and scale of the axis-aligned bounding box, a bounding box hierarchy is recursively constructed. The nodes of the bounding box hierarchy store the bounding box boundary, the left and right child node indices, or the triangular face index range contained in the leaf node. Based on the bounding box hierarchy, a parallel computing device is used to perform the nearest hit detection on the pixel ray in parallel to obtain the nearest hit result. The nearest hit result includes the hit category and the hit distance. The hit category is the semantic category identifier of the triangle face hit by the pixel ray. Pixel rays that do not hit any triangle face are treated as pixel rays with the hit category of sky; Based on the number of pixel rays that hit the category of building, the number of pixel rays that hit the category of plant, the number of pixel rays that hit the category of water, the number of pixel rays that hit the category of ground, the number of pixel rays that hit the category of sky, and the total number of pixel rays, the building ratio, plant ratio, water ratio, ground ratio, and sky ratio are calculated respectively, and the plant ratio is used as the window view green view rate. Based on the hit distance, near distance threshold and far distance threshold corresponding to each pixel ray, the viewing distance zone corresponding to each pixel ray is determined. The viewing distance zone includes near, medium and far. Pixel rays that do not hit any triangular face belong to the far. Based on the number of pixel rays corresponding to the foreground, the number of pixel rays corresponding to the midground, the number of pixel rays corresponding to the background, and the total number of pixel rays, the proportions of foreground, midground, and background are calculated respectively. The proportions of near view, medium view, and far view are compared with the effective proportion thresholds to obtain the window view distance type. Based on the green view rate of the window view and the suitable range of green view rate corresponding to the window view distance type, the green view quality value of the window view is calculated, and the green view level of the window view is obtained based on the green view quality value of the window view. The green view quality value and the green view level of the window view constitute the green view evaluation result of the window view.

[0036] In one implementation of this embodiment, the step of calculating the window view green view quality value based on the window view green view rate and the suitable green view rate range corresponding to the window view viewing distance type specifically includes: Based on the green view rate of the window view and the suitable range of green view rate corresponding to the window view distance type, the green view quality value of the window view is calculated: ; ; ; in, Indicates the viewpoint of the window. The green view quality value of the window view. Indicates the viewpoint of the window. Green view rate of the window view Indicates the type of window view distance The corresponding suitable range of green visibility The lower limit, Indicates the type of window view distance The corresponding suitable range of green visibility The upper limit.

[0037] Specifically, this invention constructs a parameterized camera, a frustum target plane, and pixel rays (a set of pixel-by-pixel rays) based on a unified image resolution, image aspect ratio, focal length, and maximum viewing distance; it triangulates the semantic scene mesh, constructs a spatial acceleration index, and obtains the element category and true hit distance at each viewpoint through parallel ray detection; it statistically analyzes the proportions of elements such as buildings, plants, water bodies, ground, and sky, as well as the proportions of near, middle, and far views, and determines the window view distance type based on an effective proportion threshold; it matches the view distance type with the proportion of plant elements point by point, calls the appropriate green view rate interval corresponding to the view distance type, and determines the window view quality value and quality level.

[0038] Furthermore, to ensure consistent observation conditions across different viewpoints and between the ray casting and cropping methods, this invention constructs a unified parametric camera model. Camera parameters include at least image resolution W, image aspect ratio R, focal length F, and maximum viewing distance D. Using the window viewpoint O as the camera position, the outward normal to the elevation or the specified viewing direction as the forward vector S, and combining this with the vertical direction in world coordinates, a stable local camera coordinate system is formed. A frustum target plane is set at distance D. The horizontal half-width and vertical half-height of the target plane are calculated based on the focal length, image aspect ratio, and maximum viewing distance, and the target plane is divided into a pixel grid consistent with the image resolution. A ray is formed between the center of each pixel and the window viewpoint, thus obtaining a set of pixel-by-pixel rays at that viewpoint.

[0039] To improve computational efficiency for multiple buildings, facades, and numerous window viewpoints at the area scale, this invention decomposes the building, vegetation, water, and ground meshes into triangular faces with category identifiers, and establishes an axis-aligned bounding box for each triangular face. Subsequently, a bounding volume hierarchy (BVH) is recursively constructed based on the position and scale of the bounding boxes. Each BVH node stores at least the bounding box boundary, left and right child node indices, or leaf node triangular face index ranges. During ray calculation, it first determines whether the ray intersects with the bounding box of a BVH node, and only traverses intersecting nodes; when a leaf node is reached, the intersection operation between the ray and candidate triangular faces is performed, thereby reducing the number of triangular faces that need to be traversed for each ray. For rays originating from the same viewpoint, if a ray intersects with multiple triangular faces, only the valid hit result closest to the viewpoint is retained. This result includes at least the hit category and the true hit distance (i.e., the hit distance). Rays that do not hit any entity triangular faces are categorized as "sky".

[0040] The viewpoint, pixel rays, triangles, and BVH node arrays are uploaded to the GPU (Graphics Processing Unit) or other parallel computing devices for batch processing. When the parallel device is unavailable, the process falls back to the CPU (Central Processing Unit) to perform the same intersection process. Different computing devices use the same data structure and the nearest-hit principle to ensure consistency of results.

[0041] After the ray analysis is completed, a statistical histogram is generated for each viewpoint, consisting of the number of rays from buildings, plants, water bodies, ground, sky, and foreground, midground, and background. This provides basic data for subsequent calculation of element proportions and identification of view distance types.

[0042] Let the total number of effective rays at each viewpoint be... The number of rays from buildings, plants, water bodies, the ground, and the sky are respectively , , , and The proportions of each type of element are as follows: ; ; ; ; .

[0043] in, , , , and These represent the proportions of buildings, vegetation, water, ground, and sky, respectively. (Vegetation proportion) As the green view rate of a window, the comprehensive proportion of natural elements can be obtained by summing the proportions of plants, water bodies, and sky.

[0044] While obtaining the element category, this invention also considers the actual hit distance of the ray. Distinguish viewing distance zones (including close-up, medium-range, and long-range). Let the critical distance between close-up and medium-range (i.e., the close-up distance threshold) be... The critical distance between mid-range and far-field (i.e., the far-field distance threshold) is: ,but: when At that time, the ray is placed in the foreground; when At that time, the ray is placed in the middle ground; when At that time, the ray is included in the background; Sky rays that do not match the entity geometry are preferably assigned to the distant view zone.

[0045] This invention does not directly use the maximum distance zone ratio to determine a single viewing distance type. Instead, it sets an effective ratio threshold U. Only when the ratio of a certain viewing distance zone (i.e., the ratio of the number of rays in a certain viewing distance zone to the total number of rays, including the proportion of near view, medium view, and far view) is not less than U, is that zone written into the effective viewing distance feature of that viewpoint. For example, if the proportions of near view and medium view at a viewpoint are both not less than U, then both near view and medium view are written into the effective viewing distance feature of that viewpoint, and the viewing distance type corresponding to that viewpoint is near-medium view. Near view, medium view, and far view can be corresponding to binary bits 001, 010, and 100, respectively. By combining bit flags, viewing distance types such as near view, medium view, near-medium view, far view, near-far view, medium-far view, and full view can be obtained.

[0046] Furthermore, based on the research objectives, the above-mentioned subcategories can be categorized into three types: near-field dominance, mid-field transition, and distant openness. By setting an effective proportion threshold, the interference of a small number of edge objects, occasional distant views, or local small elements on the overall viewing distance type judgment can be reduced, while preserving multi-distance combinations with actual spatial composition significance.

[0047] This invention first determines the window view distance type, and then calls the corresponding green view rate evaluation threshold (i.e., the upper and lower limits of the appropriate range of green view rate) according to the type, thereby avoiding the direct application of a uniform greening ratio standard to window views with different view distance structures.

[0048] Set viewpoint The merged view distance category is Its window view green view rate , and category The corresponding suitable range for green visibility is ;when When the value exceeds the upper limit, the green view quality value of the window view is... According to the preset over-density attenuation rule, it decreases as the green visibility increases; when When located within this range, the green view quality value of the window view is... Take the maximum value; when When below the lower limit, the green view quality value of the window view is... The green view quality increases with increasing green view rate. The suitable ranges corresponding to different viewing distance types can be determined by real-world window view experiments, subjective evaluations, physiological response data, or statistical fitting results, and are incorporated into this invention as configurable evaluation parameters. Based on the window view green view quality value... The window view can be further classified into quality levels such as excellent, good, medium, and poor. For low-quality window views, the green view rate, viewing distance type, building ratio, sky ratio, and main obstruction distance are recorded to distinguish different causes such as insufficient green space, near-view obstruction, mid-view gaps, distant view closure, or excessive green space.

[0049] S4. Generate multiple optimized scenarios based on the window view green vision evaluation results, re-execute multi-viewpoint window view green vision evaluation on multiple optimized scenarios until the preset window view green vision target is met or the maximum number of iterations is reached, and output the best collaborative optimization scheme.

[0050] In one implementation of this embodiment, the step of generating multiple optimized scenarios based on the window view green view evaluation results, re-executing the multi-viewpoint window view green view evaluation on the multiple optimized scenarios until a preset window view green view target is met or the maximum number of iterations is reached, and outputting the best collaborative optimization scheme, specifically includes: The window view green view evaluation results are aggregated by area, building, facade, floor and viewing distance type to identify priority improvement locations for low-quality window views; For the priority improvement locations, a set of candidate measures including building form adjustment and three-dimensional visible green volume configuration is established. Based on the set of candidate measures, multiple optimization scenarios are generated under planning control and limited green volume constraints. The multi-viewpoint window view green view evaluation was re-executed for multiple optimized scenarios to obtain multiple optimized window view green view evaluation results. If all the optimized window view green view evaluation results fail to meet the preset window view green view target and the maximum number of iterations is not reached, then the optimized scene is regenerated and the multi-viewpoint window view green view evaluation is re-executed. If the optimized window view green view evaluation result meets the preset window view green view target or reaches the maximum number of iterations, the iteration is terminated and the best collaborative optimization scheme is output.

[0051] Specifically, the calculation results are mapped back to the building facade, and structured data is output according to the viewpoint number; based on the building, facade, floor and window position of the low-quality window view, a scenario for adjusting the building form and three-dimensional green volume configuration is constructed, the analysis is re-executed and the optimization effect is compared until the preset quality target is met or the iteration termination condition is reached.

[0052] Furthermore, after completing the quality evaluation, the element proportion, viewing distance type, quality value, and quality level of each viewpoint are mapped back to the original building facade according to the viewpoint number PointIndex, generating a hierarchical color grid, text annotations, and structured data tables.

[0053] The evaluation results can be summarized by window pane, floor, facade, building, and area to reveal the spatial variation of three-dimensional visible green space across different buildings, orientations, floors, and viewing distances, and to identify low-quality window view locations that require priority improvement. For window views with significant obstruction from nearby buildings, scenarios can be constructed that optimize building setbacks, volume openings, podium height, or tower location; for window views with insufficient mid-ground greenery, the location of ground greenery, tree canopy width, and greening continuity can be adjusted; for window views on high floors or with insufficient distant natural elements, platform greening, rooftop greening, or vertical greening can be implemented.

[0054] The 3D scene of the target area is updated based on the optimized scenario, and viewpoint sampling, ray analysis, viewing distance discrimination, and quality evaluation are re-executed. By comparing the average window view quality, the improvement rate of low-quality viewpoints, the proportion of qualified viewpoints, and the quality improvement brought by the unit increase in green volume before and after optimization, the building form and 3D green volume configuration scheme are determined. When the optimization results do not meet the preset quality target, the building form or green volume configuration parameters are adjusted and recalculated; when the quality target is achieved, the visual benefit per unit of green volume no longer significantly improves, or the maximum number of iterations is reached, the iteration is terminated and the final optimized scheme is output. This provides a quantitative basis for window view spatial differentiation analysis, building form adjustment, and 3D visible green volume configuration in high-intensity urban areas.

[0055] In one implementation of this embodiment, the method for synergistic optimization of high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation further includes: Based on the standardized 3D model with semantic category identifiers, a semantically tinted window view image corresponding to the window viewpoint is generated; The pixel colors of the semantically tinted window image are converted from the RGB (Red Green Blue) color space to the HSV (Hue Saturation Value) color space to obtain the converted pixel colors; Based on the converted pixel color and the preset semantic colors of buildings, plants, water bodies and ground, determine the semantic category of the pixel corresponding to the converted pixel color, and classify the pixels that do not match any semantic category into the semantic category of sky; Based on the number of pixels corresponding to buildings, plants, water bodies, ground, sky, and the total number of pixels, the reference building proportion, reference plant proportion, reference water body proportion, reference ground proportion, and reference sky proportion are calculated respectively. The accuracy of the reference building proportion, reference plant proportion, reference water body proportion, reference ground proportion, and reference sky proportion is verified by performing a viewpoint-by-view alignment comparison with the building proportion, plant proportion, water body proportion, ground proportion, and sky proportion.

[0056] Specifically, when verification is required, screenshot segmentation and image classification are performed to align the screenshot method results with the ray-mapping method results point by point. Ray-mapping detection constitutes the main analysis path of this invention, while screenshot analysis constitutes an optional independent verification path. Both paths use the same viewpoint position, observation direction, camera parameters, and semantic classification rules, and are associated with data through unique viewpoint numbers, thereby ensuring spatial traceability and comparability of different analysis results.

[0057] Furthermore, to verify the element proportions obtained from ray tracing analysis, this invention can also set up a separate screenshot verification path. This path uses the same viewpoint position, observation direction, focal length, image scale, viewing distance, and semantic classification rules as the ray tracing analysis.

[0058] Let the total number of viewpoints to be processed be V, and each task slice contain at most S viewpoints, where V and S are both positive integers. Then the required number of task slices N is: ; Each task slice must at least store the slice number, camera parameters, viewpoint position, viewing direction, PointIndex, and global sequential index. Multiple background 3D modeling processes can handle different slices separately through the atomic assignment mechanism of task files, thereby avoiding the same slice being executed repeatedly.

[0059] In one specific implementation, an external scheduling script can launch one or more 3D modeling program processes, and each process can read screenshot task files and generate semantically shading window images in batches. This scheduling method can be implemented by using PowerShell to call Rhino programs, but is not limited to a specific scripting language or 3D modeling platform.

[0060] After the screenshot is taken, the image pixels are converted from the RGB color space to the HSV color space, and the pixel color is compared with the preset semantic colors of buildings, plants, water bodies, and the ground. When the color difference is within the set tolerance, the pixel is assigned to the corresponding category; pixels that do not match an entity category are assigned to the sky. This process calculates the percentage of element pixels under the screenshot method and checks for missing images, failed segments, and discontinuous numbering.

[0061] Screenshot verification is an optional technique for checking the results of the ray casting method, and it does not replace the line-of-sight type identification based on three-dimensional hit distance.

[0062] To address the shortcomings of existing methods for coordinated regulation of 3D green volume and building form in high-intensity areas, such as insufficient multi-viewpoint computational efficiency at the area level, lack of refined viewing distance analysis, and difficulty in continuous data transfer across platforms, this invention proposes a method for coordinated optimization of building form and 3D visible green volume in high-intensity areas based on multi-viewpoint window view green vision evaluation. This method begins with the acquisition of runtime configuration, 3D scene, and analysis parameters, and ends with the output of window view evaluation results, dual-path consistency verification, and optimization of building form and 3D green volume configuration, forming a closed-loop technical route of "data input—geometric preprocessing—viewpoint sampling—view field construction—ray analysis—view distance discrimination—quality evaluation—result verification—feedback optimization".

[0063] This invention first performs semantic classification and 3D model standardization on the observed buildings, surrounding buildings, vegetation, water bodies, and ground in the target area. Mesh sampling is then performed on the effective building facades, and a unique viewpoint number containing the building, facade, floor, and mesh number is generated for each window. Subsequently, pixel rays are constructed based on unified camera parameters. Through semantic triangular meshes, bounding box hierarchies, and GPU parallel nearest-hit calculations, the proportions of buildings, vegetation, water bodies, ground, and sky, as well as the proportions of near, mid, and far views, are obtained. The viewing distance type is determined based on an effective viewing distance threshold. Then, the viewing distance type is matched with the corresponding suitable green view rate range to form a single-window view quality result, which can be verified using image segmentation results under the same camera parameters. This invention achieves batch generation and numbering of building window viewpoints, window view distance identification, and window view green view quality evaluation, thereby revealing the spatial differentiation between buildings, facades, floors, viewing distances, and areas, providing a basis for iterative optimization of the coordinated configuration of building form and 3D green volume.

[0064] Based on this, this invention aggregates single-window results by floor, facade, building, viewing distance type, and area to identify priority improvement locations for insufficient green views, near-view shading, or low quality. It establishes a candidate measure set including building height, location, volume, setback, and ground, platform, roof, and vertical greening. Under planning control and limited green space constraints, multiple optimization scenarios are generated. Window view measurement and quality evaluation are repeatedly performed, comparing and selecting based on window view quality gain, priority viewpoint coverage, and visual benefits per unit of green space, outputting a collaborative optimization scheme for building form and three-dimensional green space. This invention transforms the traditional "two-dimensional index control—manual three-dimensional adjustment" process into a closed loop that allows for batch measurement, problem location, result verification, and iterative optimization, achieving a quantitative improvement in visual benefits under limited green space conditions.

[0065] In summary, this invention uses standardized window pane viewpoints on building facades as the basic unit to generate unique viewpoint numbers that simultaneously contain information about the building, facade, floor, and window panes. These numbers are integrated into 3D ray casting results, image segmentation results, multi-scale aggregation results, and optimization measures. Under unified camera parameters, semantic triangulation meshes, BVH spatial acceleration structures, GPU parallel nearest hits, and histogram statistics are combined for the synchronous calculation of the proportion of natural elements and near-to-far view distance structures at the area level. Different view distance types are matched with corresponding optimal green view rate intervals to form window view quality. A dual-mode analysis using ray casting and screenshot methods with the same viewpoint, direction, focal length, and aspect ratio is employed to achieve result alignment, error diagnosis, and parameter correction. Single-window results are calculated and analyzed by floor, facade, building, view distance type, and area. Under green volume budget and planning control constraints, window view measurement and quality evaluation are repeatedly performed on the combination scenarios of building form and 3D green volume, with synergistic optimization based on quality gain, priority viewpoint coverage, and visual benefits per unit of green volume.

[0066] Other variations of the present invention: (1) The viewpoint can be generated from the center of the window pane, or from the actual window opening center, the center of the curtain wall module, the floor equidistant point, the indoor observation position, or the point specified by the user.

[0067] (2) Environmental elements can be classified into five categories: buildings, plants, water bodies, ground and sky. Alternatively, roads, vehicles, mountains, facilities, glass or other semantic categories can be added, and the histogram can be expanded accordingly.

[0068] (3) The viewing distance can be calculated using three levels: near, medium and far, or two, four or continuous distance statistics.

[0069] (4) Image verification can be performed using HSV color tolerance, RGB distance, Lab color distance, traditional image segmentation or deep learning semantic segmentation; the image can be from Rhino, other renderers, real-world photos or digital twin platforms.

[0070] (5) Spatial acceleration structures can be BVH, octree, KD (K-Dimensional) tree, uniform grid or other spatial indexes; computing devices can be GPU, CPU, multi-GPU, cloud cluster or hybrid devices.

[0071] (6) The quality function can be the interval optimal, piecewise linear, quadratic function, logistic function, probability model or machine learning model; the threshold can be set by experiments, questionnaires, physiological data, standards or experts.

[0072] (7) The optimization algorithm can be exhaustive, greedy, genetic algorithm, particle swarm optimization, simulated annealing, Bayesian optimization, reinforcement learning or designer interactive selection, and can output single-objective optimal or multi-objective Pareto solution.

[0073] (8) This invention is not limited to Rhino, but can be implemented in GIS, BIM (Building Information Modeling), CIM (City Information Modeling), game engines, digital twin platforms, web page 3D engines or independent computing services.

[0074] Furthermore, based on the aforementioned method for the coordinated optimization of high-intensity area building form and three-dimensional visible green volume using multi-viewpoint window view green vision evaluation, this invention also provides a system for the coordinated optimization of high-intensity area building form and three-dimensional visible green volume using multi-viewpoint window view green vision evaluation. A preferred embodiment of this system is as follows: Figure 3 As shown, it specifically includes: The preprocessing module 01 is used to acquire the three-dimensional scene data of the target area, perform semantic classification and three-dimensional model standardization processing on the three-dimensional scene data, and obtain a standardized three-dimensional model with semantic category identifiers. The meshing module 02 is used to filter out effective building facades based on the standardized three-dimensional model, and to perform meshing processing on the effective building facades to obtain multiple window viewpoints with unique viewpoint numbers. The window view green view evaluation module 03 is used to construct a parametric camera and pixel rays based on the window viewpoint, and to perform ray detection and classification statistics on the standardized 3D model based on the parametric camera and pixel rays to obtain the window view green view evaluation result. The collaborative optimization module 04 is used to generate multiple optimization scenarios based on the window view green view evaluation results, re-execute the multi-viewpoint window view green view evaluation on the multiple optimization scenarios until the preset window view green view target is met or the maximum number of iterations is reached, and output the best collaborative optimization scheme.

[0075] Furthermore, based on the aforementioned method and system for collaborative optimization of high-intensity area building form and three-dimensional visible green volume using multi-viewpoint window view green vision evaluation, this invention also provides a terminal, wherein a preferred embodiment of the terminal is as follows: Figure 4 As shown, it specifically includes a processor 10, a memory 20, and a display 30. Figure 4 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0076] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory. In other embodiments, the memory 20 may be an external storage device of the terminal, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, or flash card. Furthermore, the memory 20 may include both internal and external storage units. The memory 20 is used to store application software and various types of data installed on the terminal, such as the terminal's program code. The memory 20 may also be used to temporarily store data that has been output or will be output. In one embodiment, the memory 20 stores a high-intensity area building form and three-dimensional visible green volume co-optimization program 40 based on multi-viewpoint window green view evaluation. This high-intensity area building form and three-dimensional visible green volume co-optimization program 40 based on multi-viewpoint window green view evaluation can be executed by the processor 10 to implement the steps of the high-intensity area building form and three-dimensional visible green volume co-optimization method based on multi-viewpoint window green view evaluation in this application.

[0077] In some embodiments, processor 10 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 20 or process data, such as executing a high-intensity area building form and three-dimensional visible green volume collaborative optimization program 40 based on multi-viewpoint window view green vision evaluation.

[0078] In some embodiments, display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. Display 30 is used to display information on the terminal and to display a visual user interface.

[0079] In one embodiment, when the processor 10 executes the high-intensity area building form and three-dimensional visible green volume collaborative optimization program 40 based on multi-viewpoint window view green vision evaluation in the memory 20, it implements the steps of the high-intensity area building form and three-dimensional visible green volume collaborative optimization method based on multi-viewpoint window view green vision evaluation as described above.

[0080] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a program for the collaborative optimization of high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window green view evaluation. When the program for the collaborative optimization of high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window green view evaluation is executed by a processor, it implements the steps of the above-described method for the collaborative optimization of high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window green view evaluation.

[0081] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Unless otherwise specified, an element defined by the phrase "comprising one" does not exclude the presence of other identical elements in the process, method, article, or terminal that includes that element.

[0082] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.). This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The computer-readable storage medium can be a memory, magnetic disk, optical disk, etc.

[0083] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A method for synergistic optimization of building form and three-dimensional visible green volume in high-intensity areas based on multi-viewpoint window view green vision evaluation, characterized in that, The method for synergistic optimization of high-intensity regional building form and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation includes: Acquire 3D scene data of the target area, perform semantic classification and 3D model standardization processing on the 3D scene data, and obtain a standardized 3D model with semantic category identifiers; Based on the standardized 3D model, effective building facades are selected and then gridded to obtain multiple window viewpoints with unique viewpoint numbers. Based on the window viewpoint, a parametric camera and pixel rays are constructed. Based on the parametric camera and pixel rays, ray detection and classification statistics are performed on the standardized 3D model to obtain the window view green view rate, near view ratio, mid view ratio, and far view ratio. The near view ratio, mid view ratio, and far view ratio are compared with effective ratio thresholds to obtain the window view distance type. Based on the window view green view rate and the appropriate green view rate range corresponding to the window view distance type, the window view green view evaluation result is calculated. Based on the window view green vision evaluation results, multiple optimization scenarios are generated. The multi-viewpoint window view green vision evaluation is re-executed for multiple optimization scenarios until the preset window view green vision target is met or the maximum number of iterations is reached, and the best collaborative optimization scheme is output.

2. The method for synergistic optimization of high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation as described in claim 1, is characterized in that, The process of acquiring 3D scene data of the target area, performing semantic classification and 3D model standardization on the 3D scene data to obtain a standardized 3D model with semantic category identifiers, specifically includes: Acquire 3D scene data of the target area, wherein the target area is a high-intensity area; The three-dimensional scene data is subjected to semantic classification and three-dimensional model standardization processing to obtain a standardized three-dimensional model with semantic category identifiers. The semantic category identifiers include buildings, plants, water bodies, ground, and sky. The buildings include the observation building and surrounding buildings.

3. The method for synergistic optimization of high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation as described in claim 2, is characterized in that... The process involves selecting effective building facades based on the standardized 3D model, then meshing these facades to obtain multiple window viewpoints with unique viewpoint numbers. Specifically, this includes: Based on the standardized three-dimensional model, multiple building surfaces of the observed building are obtained; Based on the normal direction, planar features, and geometric dimensions of the multiple building surfaces, the multiple building surfaces are screened, and the building surfaces that simultaneously meet the vertical sampling requirements, planar sampling requirements, and dimensional sampling requirements are selected as valid building facades; The effective building facade is gridded according to the set facade grid size to obtain multiple grids and multiple window viewpoints corresponding to the multiple grids. A unique viewpoint number is established for each window viewpoint. ； in, Indicates the viewpoint number. Indicates the building number. Indicates the facade number. Indicates the floor number. Indicates the grid number.

4. The method for synergistic optimization of high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation as described in claim 3, is characterized in that, The process involves constructing a parametric camera and pixel rays based on the window viewpoint, performing ray detection and classification statistics on the standardized 3D model based on the parametric camera and pixel rays to obtain the window view green view rate, near-field proportion, mid-field proportion, and far-field proportion. The near-field proportion, mid-field proportion, and far-field proportion are then compared with effective proportion thresholds to obtain the window view distance type. Based on the window view green view rate and the appropriate green view rate range corresponding to the window view distance type, the window view green view evaluation result is calculated, specifically including: Using the three-dimensional spatial coordinates of the window viewpoint as the camera position, a unified parametric camera is constructed. The parameters of the parametric camera include image resolution, image aspect ratio, focal length, and maximum viewing distance. Based on the parametric camera, a frustum target plane is set at the maximum viewing distance, and the frustum target plane is divided into multiple pixel grids according to the image resolution. The vectors pointing from the window viewpoint to the center of each pixel corresponding to each pixel grid constitute pixel rays. The mesh is split into triangular faces with semantic category identifiers. An axis-aligned bounding box is created for each triangular face. Based on the position and scale of the axis-aligned bounding box, a bounding box hierarchy is recursively constructed. The nodes of the bounding box hierarchy store the bounding box boundary, the left and right child node indices, or the triangular face index range contained in the leaf node. Based on the bounding box hierarchy, a parallel computing device is used to perform the nearest hit detection on the pixel ray in parallel to obtain the nearest hit result. The nearest hit result includes the hit category and the hit distance. The hit category is the semantic category identifier of the triangle face hit by the pixel ray. Pixel rays that do not hit any triangle face are treated as pixel rays with the hit category of sky; Based on the number of pixel rays that hit the category of building, the number of pixel rays that hit the category of plant, the number of pixel rays that hit the category of water, the number of pixel rays that hit the category of ground, the number of pixel rays that hit the category of sky, and the total number of pixel rays, the building ratio, plant ratio, water ratio, ground ratio, and sky ratio are calculated respectively, and the plant ratio is used as the window view green view rate. Based on the hit distance, near distance threshold and far distance threshold corresponding to each pixel ray, the viewing distance zone corresponding to each pixel ray is determined. The viewing distance zone includes near, medium and far. Pixel rays that do not hit any triangular face belong to the far. Based on the number of pixel rays corresponding to the foreground, the number of pixel rays corresponding to the midground, the number of pixel rays corresponding to the background, and the total number of pixel rays, the proportions of foreground, midground, and background are calculated respectively. The proportions of near view, medium view, and far view are compared with the effective proportion thresholds to obtain the window view distance type. Based on the green view rate of the window view and the suitable range of green view rate corresponding to the window view distance type, the green view quality value of the window view is calculated, and the green view level of the window view is obtained based on the green view quality value of the window view. The green view quality value and the green view level of the window view constitute the green view evaluation result of the window view.

5. The method for synergistic optimization of high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation as described in claim 4, is characterized in that... The step of calculating the green view quality value of the window view based on the green view rate of the window view and the suitable range of green view rate corresponding to the window view distance type specifically includes: Based on the green view rate of the window view and the suitable range of green view rate corresponding to the window view distance type, the green view quality value of the window view is calculated: ； ； ； in, Indicates the viewpoint of the window. The green view quality value of the window view. Indicates the viewpoint of the window. Green view rate of the window view Indicates the type of window view distance The corresponding suitable range of green visibility The lower limit, Indicates the type of window view distance The corresponding suitable range of green visibility The upper limit.

6. The method for synergistic optimization of high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation as described in claim 4, is characterized in that... The process involves generating multiple optimized scenarios based on the window view green view evaluation results, re-executing the multi-viewpoint window view green view evaluation for each of the optimized scenarios until a preset window view green view target is met or the maximum number of iterations is reached, and then outputting the best collaborative optimization scheme. Specifically, this includes: The window view green view evaluation results are aggregated by area, building, facade, floor and viewing distance type to identify priority improvement locations for low-quality window views; For the priority improvement locations, a set of candidate measures including building form adjustment and three-dimensional visible green volume configuration is established. Based on the set of candidate measures, multiple optimization scenarios are generated under planning control and limited green volume constraints. The multi-viewpoint window view green view evaluation was re-executed for multiple optimized scenarios to obtain multiple optimized window view green view evaluation results. If all the optimized window view green view evaluation results fail to meet the preset window view green view target and the maximum number of iterations is not reached, then the optimized scene is regenerated and the multi-viewpoint window view green view evaluation is re-executed. If the optimized window view green view evaluation result meets the preset window view green view target or reaches the maximum number of iterations, the iteration is terminated and the best collaborative optimization scheme is output.

7. The method for synergistic optimization of high-intensity area building form and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation as described in claim 4, is characterized in that, The method for synergistic optimization of high-intensity regional building form and three-dimensional visible green volume based on multi-viewpoint window view green vision evaluation also includes: Based on the standardized 3D model with semantic category identifiers, a semantically tinted window view image corresponding to the window viewpoint is generated; The pixel colors of the semantically tinted window image are converted from the RGB color space to the HSV color space to obtain the converted pixel colors; Based on the converted pixel color and the preset semantic colors of buildings, plants, water bodies and ground, determine the semantic category of the pixel corresponding to the converted pixel color, and classify the pixels that do not match any semantic category into the semantic category of sky; Based on the number of pixels corresponding to buildings, plants, water bodies, ground, sky, and the total number of pixels, the reference building proportion, reference plant proportion, reference water body proportion, reference ground proportion, and reference sky proportion are calculated respectively. The accuracy of the reference building proportion, reference plant proportion, reference water body proportion, reference ground proportion, and reference sky proportion is verified by performing a viewpoint-by-view alignment comparison with the building proportion, plant proportion, water body proportion, ground proportion, and sky proportion.

8. A high-intensity regional building form and three-dimensional visible green volume collaborative optimization system based on multi-viewpoint window view green vision evaluation, characterized in that, The high-intensity regional building form and three-dimensional visible green volume collaborative optimization system based on multi-viewpoint window view green vision evaluation includes: The preprocessing module is used to acquire the three-dimensional scene data of the target area, perform semantic classification and three-dimensional model standardization processing on the three-dimensional scene data, and obtain a standardized three-dimensional model with semantic category identifiers. The gridding module is used to filter out effective building facades based on the standardized 3D model, and to perform gridding processing on the effective building facades to obtain multiple window viewpoints with unique viewpoint numbers. The window view green vision evaluation module is used to construct a parametric camera and pixel rays based on the window viewpoint, and to perform ray detection and classification statistics on the standardized 3D model based on the parametric camera and pixel rays to obtain the window view green vision rate, near view ratio, mid view ratio and far view ratio. The near view ratio, mid view ratio and far view ratio are compared with the effective ratio threshold to obtain the window view distance type. The window view green vision evaluation result is calculated based on the window view green vision rate and the appropriate green vision rate range corresponding to the window view distance type. The collaborative optimization module is used to generate multiple optimization scenarios based on the window view green view evaluation results, re-execute the multi-viewpoint window view green view evaluation on the multiple optimization scenarios until the preset window view green view target is met or the maximum number of iterations is reached, and output the best collaborative optimization scheme.

9. A terminal, characterized in that, The terminal includes: a memory, a processor, and a program for co-optimizing high-intensity regional building form and three-dimensional visible green volume based on multi-viewpoint window green view evaluation, stored in the memory and executable on the processor. When the processor executes the program for co-optimizing high-intensity regional building form and three-dimensional visible green volume based on multi-viewpoint window green view evaluation, it implements the steps of the method for co-optimizing high-intensity regional building form and three-dimensional visible green volume based on multi-viewpoint window green view evaluation as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program for the collaborative optimization of high-intensity regional building form and three-dimensional visible green volume based on multi-viewpoint window green view evaluation. When the program is executed by a processor, it implements the steps of the method for the collaborative optimization of high-intensity regional building form and three-dimensional visible green volume based on multi-viewpoint window green view evaluation as described in any one of claims 1-7.

Images