An automatic generation system and method for residential building planning and design schemes

Abstract

This invention discloses an automatic generation system for residential building planning and design schemes, relating to the field of residential building design technology. The invention includes: hardware comprising a high-performance computer and virtual simulation equipment; and a software system designed according to functional requirements and structure, including a low-level logic module, input / output function modules, unit type allocation modules, forced arrangement algorithm modules, optimal selection algorithm modules, scheme adjustment modules, sunlight analysis modules, gridding modules, archiving and access authentication modules, etc. The advantages of this invention are: Based on Unreal Engine-based industrial design software, this invention takes a designer's perspective, using artificial intelligence as the main body, carrying structured data, and realizing interactive models. Visual (virtual reality) experiences and a new architectural language incorporating logical application genes are integrated into various interfaces, thereby achieving one-click automatic batch generation of residential building planning and design schemes that simultaneously meet high-quality design and economic and technical indicators.

Description

Technical Field

[0001] This invention relates to the field of architectural design technology, specifically to an automatic generation system and method for residential building planning and design schemes. Background Technology

[0002] In the traditional residential building planning and design industry, the process involves a large amount of calculation and solar radiation analysis verification. Currently, in the traditional residential building planning and design process, designers mainly treat each building as a planning unit, performing design calculations building by building. This involves a large amount of repetitive manual work to complete the design calculations and economic indicator measurements for the residential building planning scheme. This results in high time investment, low calculation speed, and repeated verifications due to overlooking details, leading to high error rates and slow output, making it difficult to improve work efficiency. The burgeoning development of artificial intelligence offers new possibilities for solving traditional residential building planning and design problems using AI. Summary of the Invention

[0003] This invention aims to at least solve one of the technical problems existing in the prior art. Therefore, one objective of this invention is to propose an automatic generation system and method for residential building planning and design schemes. This system utilizes a combination of computer algorithms and virtual reality technology to automatically generate 3D effects of residential building planning and design schemes within the Unreal Engine 4, significantly reducing the computational workload for architects, facilitating verification and correction by designers, and improving work efficiency.

[0004] To address the aforementioned technical problems, this invention proposes an automatic generation system for residential building planning and design schemes, comprising:

[0005] The hardware includes a computer and a virtual simulation device. The computer is signal-connected to the virtual simulation device. The virtual simulation device includes a junction box, a handle, a head-mounted display, and a vertical sensor. The handle, head-mounted display, and vertical sensor are electrically connected to the junction box, respectively.

[0006] The software includes a low-level logic module, a UI function module, a gridding module, a model data module, a unit type ratio module, an automatic generation module, a scheme optimization module, a sunlight analysis module, a scheme adjustment module, an output scheme module, an archiving module, a network linking module, and an authorization module configured in the computer.

[0007] The virtual simulation device is used in conjunction with the software.

[0008] As an improvement, the UI function module includes a UI operation interface unit, a 3D UI display and interaction unit, a module and interface binding unit, a multiple interface switching unit, a user input unit, and a solution output unit.

[0009] As an improvement, the grid module includes an input parameter acquisition unit, a grid coordinate generation unit, and an unavailable grid exclusion unit, which are used to obtain output parameters, generate grid coordinates, and exclude unavailable grids.

[0010] As an improvement, the automatic generation module includes an available grid coordinate acquisition unit, a random generation unit according to specifications, and a visualization output unit, which are used to obtain available grid coordinates, generate building bodies according to algorithm logic, and output the generated building bodies.

[0011] As an improvement, the scheme adjustment module includes an adjusted grid coordinate data acquisition unit, an unavailable unit re-exclusion unit, a re-generated scheme and output unit, which are used to obtain adjusted grid coordinate data, exclude unavailable grids, re-generate a scheme and output it.

[0012] As an improvement, the underlying logic module includes a model unit, a View unit, and a visualization output unit.

[0013] A method for automatically generating a residential building planning and design scheme, the method comprising the following steps:

[0014] S1. Set up the UI interface

[0015] S1-1. After entering the setting interface, add the required number of plots, and input the economic and technical indicators of different plots respectively;

[0016] S1-2. Input the coordinates of the building control line to determine the outline formed by the building control line;

[0017] S1-3. Bind the UI interface to the background for data transmission;

[0018] S2. Terrain drawing: Read the coordinates of the building control lines of different plots and draw them in the Unreal Engine UE4 through a drawing tool;

[0019] S3. Camera reset: Place the camera at a suitable position through reset, and obtain the X max and Y max as the coordinates of the camera in the X and Y directions. If the height of the Z axis is not equal to 100m, the height of Z is jointly determined by X and Y. If X>Y, then Z = X / 2. If X<Y, then Z = Y / 2. If the height of the Z axis is equal to 100m, the height of the Z axis is set to 100m;

[0020] S4. Select the model to be used for this plot in the editing interface: Select any n models in the model library. The models in the model library are imported through the datasmith plug-in. The imported models are required to be able to input different types of styles according to requirements at the same time. Register the relevant resources of the model through the json tool, and classify and name the models of the affiliated floors;

[0021] S5. Unit Type Ratio: Obtain the building model selected in the editing interface, read the square meterage of the unit types involved in the model through a JSON tool, and arrange the unit types in ascending order of square meterage. Adjust the proportion of each unit type according to the requirements. This proportion will be used as the indicator for forced arrangement generation. Each time a building is placed at a random coordinate point, the type of building placed and the proportion of the current type are counted at the same time. Compare with the target proportion. When the target proportion and the plot ratio meet the requirements at the same time, exit the random coordinate point and automatically generate the model.

[0022] S6. Generate design schemes using the forced ranking algorithm.

[0023] S6-1: Read the economic and technical indicators of different plots of land entered in the S1 settings UI interface, the unit type ratio in S5, and the model selected in S4. Divide the different plots of land into 1×1 grids, clear the existing buildings in the space, set the status of each point to unused, and ensure that all points in the plot are in an unused state.

[0024] S6-2. The number of units for buildings of different heights needs to be set according to the actual situation, with priority given to buildings with more units, and then decreasing in that order.

[0025] S6-3. Set the building's coordinates to the coordinates of its top-left corner. Use the satisfaction of different constraints in the economic and technical indicators as the condition for iterating through random coordinates. The iterative process is as follows: a. Randomize the building's coordinates; b. Determine the number of building units, model, and area coefficient based on the positional relationship of the coordinates; c. Determine if the fire separation distance and building spacing between the building at this coordinate point and other buildings meet the requirements, and whether the building is within the current plot; d. If all of the above are satisfied, the building can be placed at this point; e. Set the points within the plot occupied by the building as used; f. Calculate the economic and technical indicators after placing the building. If the economic and technical indicators have been exceeded, the building should be destroyed, and the entire design scheme ends; otherwise, continue the loop. Then, calculate the path marker of the corresponding building in the JSON based on the length, width, and height of the FVector, and output the corresponding result, which is the generated architectural design scheme.

[0026] S7, Solution Optimization

[0027] S7-1, Terrain Extraction via CV

[0028] S7-1-1 Edge Extraction

[0029] S7-1-1-1 Gaussian Blur: Gaussian filtering is a linear smoothing filter, mainly used to eliminate Gaussian noise. It is a process of weighted averaging of the entire image. Based on a two-dimensional Gaussian function, a weight matrix and a Gaussian kernel are constructed to filter each pixel.

[0030] S7-1-1-2, Grayscale Conversion: Add a conversion format to the Gaussian blurred image to convert the BGR format to a grayscale image;

[0031] S7-1-1-3, Gradient Calculation: Obtain the grayscale converted image and calculate the Sobel operators in the X and Y axis directions respectively;

[0032] S7-1-1-4, Non-maximum signal suppression: Traverse all pixels. If a point is the maximum value on the same gradient as its surrounding pixels, retain the point; otherwise, suppress the point.

[0033] S7-1-1-5, High and low threshold output binary graphics: After the above operations are completed, the resulting virtual edges are processed and the binary graphics are output;

[0034] S7-1-2, Angle Extraction

[0035] S7-1-2-1, Hough Transform Data: Call the Hough Transform function to extract the set of line segments;

[0036] S7-1-2-2, Two points determine a straight line: Traverse the data after the Hough transform, and determine a straight line for every two points;

[0037] S7-1-2-3, Floating-point numbers, Slope of a line: Convert the data type after Hough transform to a floating-point number and calculate the slope of the line;

[0038] S7-1-2-4, Arctangent and Radius to Degree: Convert radians to degrees based on the arctangent value of the slope;

[0039] S7-1-2-5, Update Degrees: Update the tilt angles of all straight lines to degrees;

[0040] S7-2. Determine if the terrain is tilted: Based on the data obtained in step one, determine whether the tilt is horizontal or vertical;

[0041] S7-3, OpenCV for terrain shape recognition

[0042] S7-3-1. After drawing the terrain, use the Unreal Engine 4's photo function to convert the terrain into JPG format;

[0043] S7-3-2. Use Imread() to load images captured by S7-3-1;

[0044] S7-3-3: Use the S7-1-1 edge extraction algorithm to obtain the edge contours of the image;

[0045] S7-3-4. Hough transform is used to detect whether an image contains curve segments. If so, it is marked as 0. If the image contains only straight line segments, the shape of the image is determined by combining different feature factors, including the polygonal approximation of the contour, length, area, and aspect ratio, and is marked as 1.

[0046] S7-4. Adjust the layout of buildings in the existing scheme based on terrain shape recognition using OpenCV.

[0047] S7-4-1 When the terrain recognition is 0, the buildings will be arranged horizontally;

[0048] S7-4-2 When the terrain identification is 1, the buildings will be adjusted according to the terrain trend;

[0049] S7-4-3 If the terrain has parts of 0 and parts of 1, the terrain will be divided and the layout of different areas will be adjusted accordingly.

[0050] S7-5, Bp training model: Reading images with high color contrast and high clarity, as follows:

[0051] S7-5-1. Read key feature data;

[0052] S7-5-2, Set up training targets and prediction data;

[0053] S7-5-3. Normalize the data in the training objects;

[0054] S7-5-4, Constructing a Bp neural network;

[0055] S7-5-5, Set the training parameters for the grid, including the number of training iterations, learning efficiency, and error of the training target;

[0056] S7-5-6, Bp training;

[0057] S7-5-7, Normalization of test sample data;

[0058] S7-5-8, Result Prediction;

[0059] S7-5-9, Comparison of actual and predicted errors;

[0060] S7-5-10. Comparison of actual and predicted results;

[0061] S7-5-11. If the comparison between the above two is poor, then adjust the Bp training model.

[0062] S7-6. If the space cannot be placed during the adjustment process, the solution adjustment module should be used for reasonable adjustment.

[0063] S8, Sunlight Analysis

[0064] S8-1. Calculate the declination angle: Determine the local latitude and longitude and convert the local time to the corresponding angle of the Earth's rotation at that time;

[0065] S8-2. Calculate solar time, solar altitude angle, and solar azimuth angle: Determine the angular position of the sun by using the solar altitude angle and solar azimuth angle;

[0066] S8-3. Conduct solar radiation analysis on the east, west, and south parts;

[0067] S9. Output of solar radiation analysis results and design scheme output

[0068] S9-1, Output of Sunlight Analysis Results: Divide the floor plan of the east, west and south sides of the building into a 1×1 grid, and label the sunlight time of each grid in hours. Different sunlight times are represented by different colors. Display the specific sunlight time of each grid in the grid, accurate to one decimal place.

[0069] S9-2, Design Scheme Output: The output scheme is a top view of the design scheme with data annotations, including building number, building height, number of floors, economic and technical indicators of the current plot, building dimensions, building spacing between buildings, between buildings and boundary lines, land boundary lines, building control lines. The building dimensions are the dimensions of the smallest bounding rectangle, and the distance between buildings and boundary lines is the distance between the nearest points parallel to the boundary.

[0070] S10. Archiving: The first part involves re-arranging the design schemes generated in S6 using the S7 scheme optimization algorithm and ensuring the S8 sunlight analysis is satisfactory. This process involves obtaining the current site's economic and technical indicators (total land area, net land area, total building area, residential area, commercial area, other area, building footprint, plot ratio, and building density), the coordinates of the site's building control lines, the coordinates of the buildings, the building models, the total number and percentage of each unit type, the total number of each floor, the sunlight analysis results, and the scheme design drawings, all saved using a JSON tool. The second part involves using the archived records to select a previously saved scheme to restore the previously saved design.

[0071] S11 Immersive VR Experience: After generating the design scheme using the strong ranking algorithm in S6, you can set up the scene of the generated scheme and experience the actual situation after the scheme is implemented. For example, you can simulate the sunshine conditions on the winter solstice and truly feel the actual changes in sunshine.

[0072] As an improvement, the model naming principle is floor number + building + model number, where the format for a separate name is floor number - number of unit types in the model H - which model number.

[0073] As an improvement, the solar radiation analysis for the east, west, and south parts in S8-3 is as follows:

[0074] East: Divide the east wall into a 1m×1m grid, and calculate the solar radiation impact of each building on each grid in a loop. The calculation time is the winter solstice or the coldest day. The eight hours of solar radiation are counted as 0.1 hours of solar radiation impact. The cumulative result is the solar radiation data of the grid.

[0075] West: Divide the east wall into 1m×1m grids and calculate the solar radiation impact of each building on each grid in a loop. The calculation time is the winter solstice or the coldest day. Eight hours of solar radiation is counted as 0.1 hours of solar radiation impact. The cumulative result is the solar radiation data of the grid. Each grid must meet the minimum requirement of two hours of solar radiation. If it meets the requirement, it is qualified; otherwise, it is unqualified.

[0076] South: Divide the south wall into 1m×1m grids. Calculate the solar radiation impact of each building on each grid in a loop. Calculate the impact from 8:00 AM to 4:00 PM on the winter solstice or the coldest day of winter. The solar radiation impact is statistically calculated in 0.1-hour increments. Every 0.1 hours, calculate whether each grid is blocked by other buildings, i.e., calculate the shadow range of each building every 0.1 hours. Determine whether each grid is within the shadow range. If it is not within the shadow range, i.e., it is not blocked, and add 0.1 hours. The result of the eight-hour solar radiation calculation is the solar radiation data for each 1m×1m grid.

[0077] As an improvement, the data annotation in S9-2 also includes the distance of each building to the surrounding buildings, including straight-line distance and diagonal distance, with each building as the center. If the building is close to the building control line, the distance of the building to the building control line is marked. The building number, number of floors, and minimum rectangular range that the building can accommodate are also marked. The computer interface displays the following: (1) total building area, plot ratio, and building density; (2) main technical and economic indicators table; (3) after clicking on a block, all related attributes are displayed; (4) the top of the block is marked with building number, building height, and number of floors.

[0078] The advantages of this invention compared to the prior art are:

[0079] This invention's system plans the terrain from an overall layout and algorithmic perspective, generating multiple different schemes at once, allowing for selective adjustments, and batch production. The computation process is handled by the computer, ensuring high efficiency and accuracy. Based on the Unreal Engine's functional applications, it can easily display 3D effects of the planning schemes, providing technical support for computationally intensive architectural design using advanced computer technology. Furthermore, it is highly adaptable, supporting multiple system operations. On the other hand, it has the advantage of openness, with powerful built-in functions. With the expansion of functional modules, it can be applied to various scenarios such as residential building planning and design, road generation, and landscape design. Customization is convenient; new functions can be added according to different project conditions. It is simple and flexible to use, allowing for real-time adjustment of model parameters and regeneration of schemes. Users can freely select multiple schemes for observation and freely adjust the viewing angle of the scheme model.

[0080] The above overview is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the invention will become readily apparent from the accompanying drawings and the following detailed description. Attached Figure Description

[0081] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0082] Figure 1 This is a system block diagram of the present invention.

[0083] Figure 2 This is a schematic diagram of the process for generating the design scheme of this invention.

[0084] Figure 3 This is a schematic diagram of the economic and technical indicators of the land parcel of this invention.

[0085] Figure 4 This is a schematic diagram of the building control lines drawn according to the present invention.

[0086] Figure 5 This is a model diagram of the apartment type selection for this invention.

[0087] Figure 6 This is a schematic diagram of the apartment layout of this invention.

[0088] Figure 7 This is a schematic diagram of the results of the forced sorting algorithm of this invention.

[0089] Figure 8This is a schematic diagram of the optimal algorithm result of the present invention.

[0090] Figure 9 This is a schematic diagram of the solar radiation analysis of the present invention.

[0091] Figure 10 This is a schematic diagram of the sunlight situation at eleven o'clock in the sunlight analysis module of the present invention.

[0092] Figure 11 This is a schematic diagram of a partial output of the solar radiation analysis results of the present invention.

[0093] Figure 12 This is a partial schematic diagram of the present invention displaying the specific sunshine hours for each grid.

[0094] Figure 13 This is a schematic diagram of the design scheme of the present invention. Figure 1 .

[0095] Figure 14 This is a schematic diagram of the design scheme of the present invention. Figure 2 .

[0096] Figure 15 This is a schematic diagram of the archiving interface of this invention. Detailed Implementation

[0097] The embodiments of this application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0098] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "joining" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0099] The present invention will now be described in further detail with reference to the accompanying drawings.

[0100] Combined with appendix Figures 1 to 14 An automatic generation system for residential building planning and design schemes, comprising:

[0101] The hardware includes a computer and a virtual simulation device. The computer is connected to the virtual simulation device via a signal connection. The virtual simulation device includes a junction box, a handle, a head-mounted display, and a vertical sensor. The handle, head-mounted display, and vertical sensor are electrically connected to the junction box, respectively.

[0102] The software includes a low-level logic module, a UI function module, a gridding module, a model data module, a unit type ratio module, an automatic generation module, a scheme optimization module, a sunlight analysis module, a scheme adjustment module, an output scheme module, an archiving module, an internet connection module, and an authorization and authentication module, etc., configured in the computer.

[0103] The software uses Unreal Engine 4 as its running and display platform. All source code is independently and controllably programmed in C++. Each functional module can realize an immersive virtual reality experience in the Unreal Engine 4 environment, achieving a one-click generation mode after data input. The completed planning schemes all comply with the requirements of laws, regulations, rules and technical specifications such as the "Urban and Rural Planning Law of the People's Republic of China", the "Land Administration Law of the People's Republic of China", and the "Opinions of the CPC Central Committee and the State Council on Establishing a Territorial Spatial Planning System and Supervising its Implementation".

[0104] The underlying logic module is used for the underlying architecture. This system uses the Model-View-Controller (MVC) pattern, dividing the entire interactive application into three parts: 1. Model: Contains the core functionality and the model itself; 2. View: Displays information to the user, i.e., the UI; 3. Controller: Processes user input data. The virtual simulation device works in conjunction with the software. The UI functional module includes UI operation interface units and 3D... The UI display and interaction unit, module and interface binding unit, multiple interface switching unit, user input unit, and scheme output unit are all included. The gridding module includes an input parameter acquisition unit, a grid coordinate generation unit, and an unusable grid exclusion unit, used to acquire output parameters, generate grid coordinates, and exclude unusable grids. The model data module includes a unit for converting unit outlines, a unit for storing unit input data, and a model import unit. The unit proportioning module includes a unit for acquiring unit types and a unit for inputting different unit proportions, used to generate requirements based on user-defined proportions of each unit type. The automatic generation module includes a unit for acquiring available grid coordinates, a unit for random generation according to specifications, and a visualization output unit, used to acquire available grid coordinates, generate building structures according to algorithmic logic, and output the generated building structures. The scheme optimization module includes an edge extraction unit, an angle extraction unit, a tilt adjustment unit, and a Bp training unit, used for machine learning from large amounts of model data to optimize the current design scheme. The sunlight analysis module includes a unit for acquiring local latitude and longitude and analysis date, a unit for determining sunlight range, and a unit for calculating the east and west sides of the building. The system includes several modules: a south-facing and south-facing solar radiation module for analyzing whether the automatically generated design meets national basic solar radiation requirements; a design adjustment module for obtaining generated grid coordinate data, eliminating unusable grids, regenerating designs, adjusting angles, and outputting data; an output design module for automatically generating site plan and building solar radiation output units for outputting site plan and building economic parameter charts, as well as outputting solar radiation information for the east, west, and south-facing sides of each building in the design, including solar radiation duration and color swatch mapping; an archiving module for obtaining the current design design and economic and technical indicator tables, storing the design design, and restoring the saved design design; an internet connection module for connecting the system to the internet for data transmission and other related operations, facilitating local area network or remote network use; and an authentication module for authorizing users to use the system resources, using HTC VIVE as the virtual simulation equipment.

[0105] Module Function Implementation

[0106] 1. UI Function Module

[0107] 1.1 User Input

[0108] Referring to the automatically generated technical solution: 1. Input conditions;

[0109] 1.2 Generation Scheme Demonstration

[0110] Referring to the automatically generated technical solution: II. Output conditions;

[0111] 1.3 Scheme Output

[0112] After selecting the generated display scheme, output it as a single file.

[0113] 2. Mesh module

[0114] 2.1 Obtaining Output Parameters

[0115] Store user input obtained from the UI interface in a structure for easy access at any time;

[0116] 2.2 Gridded Coordinate Generation

[0117] Call the terrain's length and width, generate a grid coordinate system based on its values, and use a boolean to determine whether a single grid is available;

[0118] 2.3 Exclusion of Unavailable Grids

[0119] Call the occupied grid coordinates and set its boolean value to true.

[0120] 3. Automatic generation module

[0121] 3.1 Obtaining Available Grid Coordinates

[0122] Re-establish the grid coordinate system and filter out grid coordinates with a bool value of false, indicating that they are usable;

[0123] 3.2 Randomly generate the scheme according to the specifications

[0124] On the available grid, generate building structures according to building codes and algorithmic logic;

[0125] 3.3 Visualization Output

[0126] Output the grid coordinates and grid boolean values ​​occupied by the generated building, and update the UI.

[0127] 4. Scheme Adjustment Module

[0128] By reusing the output parameters, unusable grid exclusion, and automatic generation modules, even if a building can still be placed on a local plot after the forced or optimized layout is completed, but it exceeds the building control line, it is only necessary to adjust the angle of the residential building to ensure that the residential building is within the building control line.

[0129] 5. Authentication Module

[0130] To ensure the rights and interests of developers, the software is bound with a unique code for each device, so that the software cannot be used on unauthorized computers.

[0131] A method for automatically generating a residential building planning and design scheme, the method comprising the following steps:

[0132] S1. Set up the UI interface

[0133] S1-1. After entering the setting interface, add the number of plots required, and input the economic and technical indicators of different plots, including the total land area, net land area, commercial area, other area, plot ratio, and building density. For example, Figure 3 as shown, the total land area is set to 72686 m 2 , the net land area is set to 72686 m 2 , the commercial area is set to 16055 m 2 , the other area is set to 0 m 2 , the plot ratio is set to 1.19, and the building density is set to 0.21;

[0134] S1-2. Input the coordinates of the building control lines of each plot to determine the outline formed by the building control lines;

[0135] S1-3. Bind the UI interface to the background for data transmission;

[0136] S2: Terrain drawing: Read the coordinates of the building control lines of different plots and draw them in the Unreal Engine UE4 through a drawing tool, as Figure 4 shown;

[0137] S3: Camera reset: Place the camera in a suitable position through reset, and obtain the X max and Y max coordinates of the camera in the X and Y directions. If the Z-axis height is not equal to 100 m, the height of Z is jointly determined by X and Y. If X > Y, then Z = X / 2. If X < Y, then Z = Y / 2. If the Z-axis height is equal to 100 m, the Z-axis height is set to 100 m;

[0138] S4: Select the models to be used for this plot in the editing interface: Select any n models from the model library, such as Figure 5As shown, the model design includes four different apartment types: floors 01-06, 07-11, 12-18, and 19-33. Models in the model library are imported through the Datasmith plugin. Importing models allows for the input of different styles as needed, just as a piece of land can simultaneously build buildings of different apartment types and heights. The relevant resources of the model are registered using a JSON tool, and models belonging to the floor number are classified and named. The model naming principle is: floor number + building + model number, such as 6building1, which means the first model of a building with a total of 6 floors. The format of the name that needs to be named separately is: floor number - number of apartment types in the model H - which model, such as 6-2H-001, which means the first model of a building with two apartment types with a total of 6 floors. Then, the minimum and maximum heights, total square meters, length, width, apartment type composition, total number of apartment types, variations and square meters of each apartment type, outline coordinates and other key parameters of the model are confirmed.

[0139] S5: Unit Type Ratio: Obtain the building model selected in the editing interface, read the square meterage of each unit type involved in the model using a JSON tool, and arrange the unit types in ascending order of square meterage. Without manual modification of the unit type ratio, the ratio of all unit types is uniform, with a total ratio of 100%. The ratio of each unit type can be adjusted according to needs. This ratio will be used as an indicator for forced layout generation. Each time a building is placed at a random coordinate point, the type of building and the current ratio of that type are simultaneously calculated and compared with the target ratio. Once both the target ratio and the floor area ratio meet the requirements, the random coordinate point is exited, and automatic model generation begins. Figure 6 As shown, the apartment size is 120m². 2 This unit type accounts for 34% of the total; the unit area is 116m². 2 This unit type accounts for 33% of the total; the unit area is 100m². 2 This unit type accounts for 33% of the total.

[0140] S6: Generating design schemes using a forced ranking algorithm

[0141] S6-1: Read the economic and technical indicators of different plots of land entered in the S1 setting UI interface, the unit type ratio in S5, and the model selected in S4. Divide the different plots of land into 1×1 grids, clear the existing buildings in the space, set the status of each point to unused, and ensure that all points in the plot are in an unused state.

[0142] S6-2. The number of units for buildings of different heights needs to be set according to the actual situation, with priority given to buildings with more units, and then decreasing in that order.

[0143] S6-3. Set the building's coordinates to the coordinates of its top-left corner. Use the satisfaction of different constraints in the economic and technical indicators as the condition for iterating through random coordinate points. The following steps are executed in the loop: a) Randomize the building's coordinates; b) Determine the number of building units, model, and area coefficient based on the positional relationship of the coordinates; c) Determine if the fire separation distance and building spacing between the building at that coordinate point and other buildings are satisfied, and whether the building is within the current plot; d) If all of the above are satisfied, the building can be placed at that point; e) Set the points within the plot occupied by the building at that point as used; f) Calculate the economic and technical indicators after placing the building. If the economic and technical indicators are exceeded, the building should be destroyed, and the entire design scheme ends; otherwise, continue the loop. Then, calculate the path marker of the corresponding building in the JSON based on the length, width, and height corresponding to FVector, and output the corresponding result, which is the generated architectural design scheme. Figure 7 As shown, the area is 120m² 2 There are a total of 288 units, accounting for 33.30% of the total; the area is 116 square meters. 2 There are a total of 226 units, accounting for 26.20% of the total; the area is 100 square meters. 2 There are a total of 350 units of this type, accounting for 40.50%;

[0144] S7, Solution Optimization

[0145] S7-1, Terrain Extraction via CV

[0146] S7-1-1 Edge Extraction

[0147] S7-1-1-1 Gaussian Blur: Gaussian filtering is a linear smoothing filter, mainly used to eliminate Gaussian noise. It is a process of weighted averaging of the entire image. Based on a two-dimensional Gaussian function, a weight matrix and a Gaussian kernel are constructed to filter each pixel.

[0148] S7-1-1-2, Grayscale Conversion: Add a conversion format to the Gaussian blurred image to convert the BGR format to a grayscale image;

[0149] S7-1-1-3, Gradient Calculation: Obtain the grayscale converted image and calculate the Sobel operators in the X and Y axis directions respectively;

[0150] S7-1-1-4, Non-maximum signal suppression: Traverse all pixels. If a point is the maximum value on the same gradient as its surrounding pixels, retain the point; otherwise, suppress the point.

[0151] S7-1-1-5, High and Low Threshold Output Binary Graphics: After completing the above operations, process the resulting virtual edges and output a binary graphic, as follows:

[0152] a. If the gradient value of the current edge pixel is greater than or equal to the high threshold, the pixel is marked as a strong edge;

[0153] b. If the gradient value of the current edge pixel is between the high and low thresholds, the pixel is marked as a virtual edge;

[0154] c. If the gradient value of the current edge pixel is less than the low threshold, the pixel is suppressed;

[0155] d. For the obtained virtual edges, if they are connected to strong edges, they are processed as edges; otherwise, they are suppressed.

[0156] S7-1-2, Angle Extraction

[0157] S7-1-2-1, Hough Transform Data: Call the Hough Transform function to extract the set of line segments;

[0158] S7-1-2-2, Two points determine a straight line: Traverse the data after the Hough transform, and determine a straight line for every two points;

[0159] S7-1-2-3, Floating-point numbers, Slope of a line: Convert the data type after Hough transform to a floating-point number and calculate the slope of the line;

[0160] S7-1-2-4, Arctangent and Radius to Degree: Convert radians to degrees based on the arctangent value of the slope;

[0161] S7-1-2-5, Update Degrees: Update the tilt angles of all straight lines to degrees;

[0162] S7-2. Determine if the terrain is tilted: Based on the data obtained in the first step, determine whether the tilt is horizontal or vertical. If it is horizontal, first summarize the horizontal tilt cases, then determine the optimal angle range for the horizontal tilt, and obtain the optimal tilt angle range. If it is vertical, first summarize the vertical tilt cases, then determine the optimal angle range for the vertical tilt, and obtain the optimal tilt angle range.

[0163] S7-3, OpenCV for terrain shape recognition

[0164] S7-3-1. After drawing the terrain, use the Unreal Engine 4's photo function to convert the terrain into JPG format;

[0165] S7-3-2. Use Imread() to load the image taken by S7-3-1;

[0166] S7-3-3: Use the S7-1-1 edge extraction algorithm to obtain the edge contours of the image;

[0167] S7-3-4. Hough transform is used to detect whether an image contains curve segments. If so, it is marked as 0. If the image contains only straight line segments, the shape of the image is determined by combining different feature factors, such as the polygonal approximation of the contour, length, area and aspect ratio, and marked as 1.

[0168] S7-4. Adjust the layout of buildings in the existing scheme based on terrain shape recognition using OpenCV.

[0169] S7-4-1 When the terrain recognition is 0, the buildings will be arranged horizontally;

[0170] S7-4-2 When the terrain identification is 1, the buildings will be adjusted according to the terrain trend;

[0171] S7-4-3 If the terrain has parts of 0 and parts of 1, the terrain will be divided and the layout of different areas will be adjusted accordingly.

[0172] S7-5, Bp training model: Reading images with high color contrast and high clarity, as follows:

[0173] S7-5-1. Read key feature data;

[0174] S7-5-2, Set up training targets and prediction data;

[0175] S7-5-3. Normalize the data in the training objects;

[0176] S7-5-4, Constructing a Bp neural network;

[0177] S7-5-5, Set the training parameters for the grid, including the number of training iterations, learning efficiency, and error of the training target;

[0178] S7-5-6, Bp training;

[0179] S7-5-7, Normalization of test sample data;

[0180] S7-5-8, Result Prediction, such as Figure 8 As shown;

[0181] S7-5-9, Comparison of actual and predicted errors;

[0182] S7-5-10. Comparison of actual and predicted results;

[0183] S7-5-11. If the comparison between the above two is poor, then adjust the Bp training model.

[0184] S7-6. If the space cannot be placed during the adjustment process, it needs to be adjusted appropriately.

[0185] S8, Sunlight Analysis

[0186] S8-1. Calculate the declination angle, determine the local latitude and longitude, and convert the local time to the corresponding angle of the Earth's rotation at that time;

[0187] S8-2. Calculate solar time, solar altitude angle, and solar azimuth angle, and determine the angular position of the sun using the solar altitude angle and solar azimuth angle;

[0188] S8-3, such as Figures 9-12 As shown, the solar radiation analysis for the east, west, and south sections is as follows:

[0189] East: Divide the east wall into a 1m×1m grid, and calculate the solar radiation impact of each building on each grid in a loop. The calculation time is the winter solstice or the coldest day. The eight hours of solar radiation are counted as 0.1 hours of solar radiation impact. The cumulative result is the solar radiation data of the grid.

[0190] West: Divide the east wall into 1m×1m grids and calculate the solar radiation impact of each building on each grid in a loop. The calculation time is the winter solstice or the coldest day. Eight hours of solar radiation is counted as 0.1 hours of solar radiation impact. The cumulative result is the solar radiation data of the grid. Each grid must meet the minimum requirement of two hours of solar radiation. If it meets the requirement, it is qualified; otherwise, it is unqualified.

[0191] South: Divide the south wall into 1m×1m grids. Calculate the solar radiation impact of each building on each grid in a loop. Calculate the solar radiation impact from 8:00 AM to 4:00 PM on the winter solstice or the coldest day of winter. The solar radiation impact is counted in 0.1-hour increments. Every 0.1 hours, calculate whether each grid is blocked by other buildings, i.e., calculate the shadow range of each building every 0.1 hours. Determine whether each grid is within the shadow range. If it is not within the shadow range, i.e., it is not blocked, and add 0.1 hours. The result of the eight-hour solar radiation calculation is the solar radiation data for each 1m×1m grid.

[0192] S9. Output of solar radiation analysis results and design scheme output

[0193] S9-1, Output of Sunlight Analysis Results: The floor plan of the building's east, west, and south sides is divided into a 1×1 grid. The sunlight duration of each grid is labeled in hours, and different sunlight durations are represented by different colors, such as... Figure 12 As shown, the grid displays the specific sunshine hours for each grid cell, accurate to one decimal place. The accuracy can also be adjusted according to actual conditions.

[0194] S9-2, Design Scheme Output: (e.g.) Figures 13-14As shown, the output scheme is a top view of the design scheme with data annotation, including building number, building height, number of floors, output of the current plot's economic and technical indicators, building size, building spacing between buildings, between buildings and boundary lines, land boundary lines, building control lines, building size is the size of the smallest bounding rectangle, building and boundary line is the distance between the nearest point parallel to the boundary, data annotation also includes the distance of the surrounding buildings from each building as the center, including straight distance and diagonal distance, if the building is close to the building control line, the distance of the building from the building control line is marked, the building number, number of floors of each building, and the minimum rectangular range that the building can accommodate, the computer interface displays the following content: (1) total building area, plot ratio, building density; (2) main technical and economic indicators table; (3) after clicking a square, all related attributes are displayed; (4) the top of the square is marked with building number, building height, and number of floors;

[0195] S10, Save File: (e.g.) Figure 15 As shown, the first part involves obtaining the current site's economic and technical indicators (total land area, net land area, total building area, residential area, commercial area, other area, building footprint, plot ratio, and building density), coordinates of the site's building control lines, coordinates of the buildings, building models, total number and percentage of each unit type, total number of floors, sunlight analysis results, and design drawings using a JSON tool after the S7 scheme optimization algorithm rearranges the design scheme generated in S6 and the S8 sunlight analysis is qualified. The second part involves restoring the previously saved design by selecting a previously saved scheme from the archived records.

[0196] S11, Immersive VR Experience: After generating the design scheme using the strong ranking algorithm in S6, the generated scheme can be used to set up a scene and experience the actual situation after the scheme is implemented. For example, the simulation of the sunshine on the winter solstice can be used to truly feel the actual changes in sunshine.

[0197] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for automatically generating residential building planning and design schemes, characterized in that: The method includes the following steps: S1. Set the UI interface S1-1. After entering the settings interface, add the required number of land parcels and enter the economic and technical indicators for different land parcels respectively. S1-2. Input the coordinates of the building control lines to determine the outline formed by the building control lines; S1-3. Bind the UI interface and the backend to perform data transmission; S2. Terrain rendering: Read the coordinates of the building control lines of different plots and render them in Unreal Engine 4 using drawing tools; S3. Lens reset: Place the lens in the appropriate position through reset, and obtain X in the current plot max and Y max are the coordinates of the lens in the X and Y directions. When the Z-axis height is not equal to 100m, the height of Z is jointly determined by X and Y. If X > Y, then Z = X / 2; if X < Y, then Z = Y / 2. If the Z-axis height is equal to 100m, the height of Z is set to 100m; S4. In the editing interface, select the model to be used for this plot: Select any n models from the model library. The models in the model library are imported through the Datasmith plugin. Importing models requires the ability to input different types of styles at the same time as needed. Register the relevant resources of the model through the JSON tool, and classify and name the models of the corresponding layers. S5. Unit Type Ratio: Obtain the building model selected in the editing interface, read the square meterage of the unit types involved in the model through a JSON tool, and arrange the unit types in ascending order of square meterage. Adjust the proportion of each unit type according to the requirements. This proportion will be used as the indicator for forced arrangement generation. Each time a building is placed at a random coordinate point, the type of building placed and the proportion of the current type are counted and compared with the target proportion. When the target proportion and the plot ratio meet the requirements at the same time, exit the random coordinate point and automatically generate the model. S6. Generate design schemes using the forced ranking algorithm. S6-1: Read the economic and technical indicators of different plots of land entered in the S1 settings UI interface, the unit type ratio in S5, and the model selected in S4. Divide the different plots of land into 1×1 grids, clear the existing buildings in the space, set the status of each point to unused, and ensure that all points in the plot are in an unused state. S6-2. The number of units for buildings of different heights needs to be set according to the actual situation, with priority given to buildings with more units, and then decreasing in that order. S6-3. Set the coordinates of the building to the coordinates of its upper left corner. Use the satisfaction of different constraints in the economic and technical indicators as the condition for cyclically selecting random coordinates. The cyclical execution is carried out according to the following steps: a. Coordinates of random buildings; b. Determine the number of building units, model, and area coefficient based on the positional relationship of the coordinate points; c. Determine whether the fire separation distance and building spacing between the building at this coordinate point and other buildings meet the requirements, and whether the building is within the current plot; d. If all of the above conditions are met, then place the building at that point; e. Set the points within the area occupied by the building to be used; f. Calculate the economic and technical indicators after placing the building. If the economic and technical indicators have been exceeded, the building should be destroyed and the entire design scheme ends. Otherwise, continue the loop and calculate the path marker of the corresponding building in the JSON according to the length, width and height of FVector. Output the corresponding result, which is the generated architectural design scheme. S7, Solution Optimization S7-1, Terrain Extraction via CV S7-1-1 Edge Extraction S7-1-1-1 Gaussian Blur: Based on a two-dimensional Gaussian function, a weight matrix and a Gaussian kernel are constructed to perform filtering on each pixel. S7-1-1-2, Grayscale Conversion: Add a conversion format to the Gaussian blurred image to convert the BGR format to a grayscale image; S7-1-1-3, Gradient Calculation: Obtain the grayscale converted image and calculate the Sobel operators in the X and Y axis directions respectively; S7-1-1-4, Non-maximum signal suppression: Traverse all pixels. If a point is the maximum value on the same gradient as its surrounding pixels, retain the point; otherwise, suppress the point. S7-1-1-5, High and low threshold output binary graphics: After the above operations are completed, the resulting virtual edges are processed and the binary graphics are output; S7-1-2, Angle Extraction S7-1-2-1, Hough Transform Data: Call the Hough Transform function to extract the set of line segments; S7-1-2-2, Two points determine a straight line: Traverse the data after the Hough transform, and determine a straight line for every two points; S7-1-2-3, Floating-point numbers, Slope of a line: Convert the data type after Hough transform to a floating-point number and calculate the slope of the line; S7-1-2-4, Arctangent and Radius to Degree: Convert radians to degrees based on the arctangent value of the slope; S7-1-2-5, Update Degrees: Update the tilt angles of all straight lines to degrees; S7-2. Determine if the terrain is tilted: Based on the data obtained in step one, determine whether it is tilted, horizontal, or vertical. S7-3, OpenCV for terrain shape recognition S7-3-1. After drawing the terrain, use the Unreal Engine 4's photo function to convert the terrain into JPG format; S7-3-2. Use Imread() to load images captured by S7-3-1; S7-3-3: Use the S7-1-1 edge extraction algorithm to obtain the edge contours of the image; S7-3-4. Hough transform is used to detect whether an image contains curve segments. If so, it is marked as 0. If the image contains only straight line segments, the shape of the image is determined by combining different feature factors, including the polygonal approximation of the contour, length, area, and aspect ratio, and is marked as 1. S7-4. Adjust the layout of buildings in the existing scheme based on terrain shape recognition using OpenCV. S7-4-1 When the terrain recognition is 0, the buildings will be arranged horizontally; S7-4-2 When the terrain identification is 1, the buildings will be adjusted according to the terrain trend; S7-4-3 If the terrain has parts of 0 and parts of 1, the terrain will be divided and the layout of different areas will be adjusted accordingly. S7-5, Bp training model: Reading images with high color contrast and high clarity, as follows: S7-5-1. Read key feature data; S7-5-2. Set up training targets and prediction data; S7-5-3. Normalize the data in the training objects; S7-5-4, Constructing a Bp neural network; S7-5-5, Set the training parameters for the grid, including the number of training iterations, learning efficiency, and error of the training target; S7-5-6, Bp training; S7-5-7, Normalization of test sample data; S7-5-8, Result Prediction; S7-5-9, Comparison of actual and predicted errors; S7-5-10. Comparison of actual and predicted results; S7-5-11. If the comparison between the above two is poor, then adjust the Bp training model. S7-6. If the space cannot be placed during the adjustment process, the solution adjustment module should be used for reasonable adjustment. S8, Sunlight Analysis S8-1. Calculate the declination angle: Determine the local latitude and longitude and convert the local time to the corresponding angle of the Earth's rotation at that time; S8-2. Calculate solar time, solar altitude angle, and solar azimuth angle: Determine the angular position of the sun by using the solar altitude angle and solar azimuth angle; S8-3. Conduct solar radiation analysis on the east, west, and south parts; S9. Output of solar radiation analysis results and design scheme; S9-1, Output of Sunlight Analysis Results: Divide the floor plan of the east, west and south sides of the building into a 1×1 grid, and label the sunlight time of each grid in hours. Different sunlight times are represented by different colors. Display the specific sunlight time of each grid in the grid, accurate to one decimal place. S9-2, Design Scheme Output: The output scheme is a top view of the design scheme with data annotations, including building number, building height, number of floors, economic and technical indicators of the current plot, building dimensions, and building spacing. The building spacing is the distance between buildings and between buildings and the boundary line. The building dimensions are the dimensions of the smallest bounding rectangle. The distance between buildings and the boundary line is the distance between the nearest point parallel to the boundary. S10, Archiving: The first part involves re-arranging the design schemes generated in S6 using the S7 scheme optimization algorithm and ensuring that the S8 sunlight analysis is qualified. This process involves obtaining the current site's economic and technical indicators, the coordinates of the site's building control lines, the coordinates of the buildings, the building models, the total number and proportion of each unit type, the total number of each floor, the sunlight analysis results, and the scheme design drawings, and saving them using a JSON tool. The second part involves using the archived records to select a previously saved scheme to restore the previously saved design. Economic and technical indicators include total land area, net land area, total building area, residential area, commercial area, other area, building footprint, plot ratio, and building density. S11, Immersive VR Experience: After generating the design scheme using the strong ranking algorithm in S6, the generated scheme is used to set up a scene and experience the actual situation after the scheme is implemented. For example, the simulation of the sunlight on the winter solstice allows you to truly feel the actual changes in sunlight.

2. The method for automatically generating residential building planning and design schemes according to claim 1, characterized in that: The solar radiation analysis for the east, west, and south parts in S8-3 is as follows: East: Divide the east wall into a 1m×1m grid, and calculate the solar radiation impact of each building on each grid in a loop. The calculation time is the winter solstice or the coldest day. The eight hours of solar radiation are counted as 0.1 hours of solar radiation impact. The cumulative result is the solar radiation data of the grid. West: Divide the west wall into 1m×1m grids and calculate the solar radiation impact of each building on each grid in a loop. The calculation time is the winter solstice or the coldest day. Eight hours of solar radiation is counted as 0.1 hours of solar radiation impact. The cumulative result is the solar radiation data of the grid. Each grid must meet the minimum requirement of two hours of solar radiation. If it meets the requirement, it is qualified; otherwise, it is unqualified. South: Divide the south wall into 1m×1m grids. Calculate the solar radiation impact of each building on each grid in a loop. Calculate the impact from 8:00 AM to 4:00 PM on the winter solstice or the coldest day of winter. The solar radiation impact is statistically calculated in 0.1-hour increments. Every 0.1 hours, calculate whether each grid is blocked by other buildings, i.e., calculate the shadow range of each building every 0.1 hours. Determine whether each grid is within the shadow range. If it is not within the shadow range, i.e., it is not blocked, and add 0.1 hours. The result of the eight-hour solar radiation calculation is the solar radiation data for each 1m×1m grid.

3. The method for automatically generating residential building planning and design schemes according to claim 1, characterized in that: The data annotation in S9-2 also includes the distance between each building and the surrounding buildings, including straight-line distance and diagonal distance, with each building as the center. If the building is close to the building control line, the distance between the building and the building control line is marked. The building number, number of floors, and minimum rectangular range that the building can accommodate are also marked. The computer interface displays the following: (1) total building area, plot ratio, and building density; (2) main technical and economic indicators table; (3) after clicking on a block, all related attributes are displayed; (4) the top of the block is marked with building number, building height, and number of floors.

4. A residential building planning and design scheme automatic generation system applied to the residential building planning and design scheme automatic generation method described in claim 1, characterized in that: The system includes: The hardware includes a computer and a virtual simulation device. The computer is signal-connected to the virtual simulation device. The virtual simulation device includes a junction box, a handle, a head-mounted display, and a vertical sensor. The handle, head-mounted display, and vertical sensor are electrically connected to the junction box, respectively. The software includes a low-level logic module, a UI function module, a gridding module, a model data module, a unit type ratio module, an automatic generation module, a scheme optimization module, a sunlight analysis module, a scheme adjustment module, an output scheme module, an archiving module, and an internet connection and authorization module configured in the computer. The virtual simulation device is used in conjunction with the software.

5. The automatic generation system for residential building planning and design schemes according to claim 4, characterized in that: The UI functional module includes a UI operation interface unit, a 3D UI display and interaction unit, a module and interface binding unit, a multiple interface switching unit, a user input unit, and a solution output unit.

6. The automatic generation system for residential building planning and design schemes according to claim 4, characterized in that: The meshing module includes an input parameter acquisition unit, a mesh coordinate generation unit, and an unusable mesh exclusion unit, used to acquire output parameters, generate mesh coordinates, and exclude unusable meshes.

7. The automatic generation system for residential building planning and design schemes according to claim 4, characterized in that: The automatic generation module includes a unit for obtaining available grid coordinates, a unit for random generation according to specifications, and a visualization output unit, used to obtain available grid coordinates, generate building structures according to algorithm logic, and output the generated building structures.

8. The automatic generation system for residential building planning and design schemes according to claim 4, characterized in that: The scheme adjustment module includes a unit for obtaining adjusted grid coordinate data, a unit for excluding unusable grids again, and a unit for regenerating and outputting the scheme. It is used to obtain adjusted grid coordinate data, exclude unusable grids, regenerate the scheme, and output it.

9. The automatic generation system for residential building planning and design schemes according to claim 4, characterized in that: The underlying logic module includes a model unit, a view unit, and a visualization output unit.

Images