Landscape design plant configuration image simulation and seasonal change visualization methods and systems

Abstract

This invention relates to the field of digital landscape design technology, and discloses a method and system for simulating landscape plant configuration images and visualizing seasonal changes. The method includes: acquiring site image data and performing environmental element identification and light condition analysis; retrieving plant model data and seasonal morphological parameters from a plant model library; fusing plant rendering images with site images using a light-adaptive image fusion algorithm; performing seasonal gradient rendering of plants based on HSV color space mapping to generate seasonal landscape effect images; and calculating long-term plant sizes based on a growth prediction model to generate long-term landscape effect images and visualization reports.

Description

Technical Field

[0001] This invention relates to the field of digital landscape design technology, and in particular to a method and system for simulating images of plant configurations and visualizing seasonal changes in landscape design. Background Technology

[0002] Plant selection is a core element of landscape design, directly impacting the aesthetic value and ecological function of the garden. In the traditional landscape design process, designers select plants and arrange spaces based on their professional experience and botanical knowledge, while clients understand the design scheme by reviewing drawings and renderings. However, because plants are living landscape elements, their forms change with the seasons and expand with age. This dynamic characteristic makes it difficult for static design drawings to fully represent the actual effect of plant arrangements, resulting in a lack of intuitive visual reference for clients when making decisions.

[0003] Existing digital design technologies for garden landscapes primarily focus on the construction and rendering of 3D models. Chinese patent CN119810313A discloses a method and system for intelligent digital design of garden landscapes based on 3D models. This technology acquires garden landscape data, preprocesses it to obtain standard garden landscape data, and then performs multi-dimensional feature recognition on the standard garden landscape data to obtain multi-dimensional feature data. Next, it extracts garden terrain features and determines spatial layout from the multi-dimensional feature data, and performs correlation analysis between the garden spatial layout data and plant landscape features to generate spatial correlation data. Then, it performs 3D data mapping on the spatial correlation data, and uses the 3D garden landscape data to construct a 3D model of the garden landscape based on the multi-dimensional feature data. Finally, it performs lighting simulation and plant seasonal change analysis on the 3D model of the garden landscape, and enhances intelligent digital visualization by adjusting the data based on landscape changes.

[0004] While the aforementioned existing technologies have achieved a three-dimensional digital presentation of garden landscapes, they still suffer from the following technical problems: First, the technology relies on a complete three-dimensional modeling process, requiring extensive data collection and processing of garden landscapes, which is insufficient for rapid previewing of effects during the early design phase; Second, the seasonal change analysis of plants in this technology is mainly achieved by replacing plant species and updating plant morphology data, lacking a continuous and smooth seasonal transition effect, making it difficult to realistically reproduce the natural transition process of plant changes throughout the four seasons; Third, the technology does not involve long-term plant growth prediction functions, failing to show clients the landscape effect of plants three, five, or even ten years later, limiting clients' comprehensive understanding of the spatiotemporal evolution of the landscape design; Fourth, the core of this technology is the construction of a three-dimensional model, lacking the ability to directly utilize the site photos provided by clients, and failing to achieve a natural integration of photos and plant models.

[0005] Therefore, it is necessary to develop a specialized technical solution for predicting the effects of plant configuration in landscape design. This solution can directly simulate plant configuration based on existing site photos, achieve natural integration of plants with the site background, provide continuous visualization of seasonal landscape changes, and support long-term growth prediction. This will help owners fully understand the spatiotemporal evolution characteristics of the landscape design scheme and assist them in making more accurate decisions. Summary of the Invention

[0006] Therefore, it is necessary to provide a method and system for simulating landscape plant configuration images and visualizing seasonal changes to solve the technical problems in existing technologies, such as relying on the designer's experience to predict the effect of plant configuration, difficulty in intuitively presenting seasonal landscape changes, and inability to preview long-term growth effects, which leads to difficulties in decision-making for owners.

[0007] To achieve the above objectives, a method for simulating landscape plant configuration images and visualizing seasonal changes is provided, the method comprising the following steps: Step S1: Obtain site image data of the site to be designed, perform semantic segmentation of environmental elements on the site image data, identify the terrain undulation area, existing vegetation area, structure area and configurable area in the site image data, and generate environmental element identification results; perform illumination condition analysis on the site image data, extract the light source direction information, light intensity distribution information and shadow distribution information of the site image data, and generate illumination condition parameters.

[0008] Step S2: Receive the plant species information specified in the design scheme, retrieve the corresponding target plant model data from the plant model library based on the plant species information; obtain the set of seasonal morphological parameters corresponding to the target plant model data; determine the plant configuration location based on the configurable area in the environmental element identification results and the design scheme, and generate configuration location coordinate information.

[0009] Step S3: Render the target plant model data to the corresponding location according to the configured location coordinate information to generate a plant rendering image. Perform illumination correction processing on the plant rendering image according to the illumination condition parameters. Calculate the illumination consistency weight value between the site image data and the plant rendering image. Use an image fusion algorithm based on the illumination consistency weight value to fuse the illumination-corrected plant rendering image with the site image data to generate fused image data.

[0010] Step S4: Based on the set of seasonal morphological parameters, render the seasonal changes of the plant areas in the fused image data. Use the color space mapping method to perform continuous gradient processing on the plant leaf color. Combine the plant crown shape change parameters to adjust the plant outline and generate landscape effect images of spring, summer, autumn and winter in sequence, forming a seasonal rendering image sequence.

[0011] Step S5: Obtain the growth prediction model parameters corresponding to the target plant model data, calculate the predicted crown width and height of the plant at the preset time node based on the growth prediction model parameters, perform scale transformation processing on the plants in the seasonal rendering image sequence based on the predicted values, and generate a long-term landscape effect image; integrate and output the seasonal rendering image sequence and the long-term landscape effect image to generate a visualization report file.

[0012] This invention acquires site image data and performs environmental element identification and lighting condition analysis in step S1, providing an accurate site information foundation for subsequent plant configuration and image fusion, allowing design simulation work to be carried out without complex 3D modeling. Step S2 obtains plant model data containing seasonal morphological parameters from a plant model library, providing complete parameter support for seasonal change rendering. Step S3 employs a lighting-adaptive image fusion algorithm, adjusting the fusion process by calculating lighting consistency weight values, achieving a natural transition between the rendered plant image and the site background, avoiding harsh boundaries and inconsistent lighting effects. Step S4 uses the HSV color space mapping method to perform continuous gradient processing on plant leaf color, achieving a smoother and more natural seasonal transition effect compared to the model replacement method in existing technologies. Step S5 introduces a parametric growth prediction model, which can calculate and visualize the predicted size values ​​of plants at different time points, filling the gap in existing technologies for previewing long-term growth effects. Therefore, this invention, through the synergistic cooperation of four core technologies—environmental element identification, light-adaptive image fusion, HSV color space seasonal phase gradation, and parametric growth prediction—forms a complete closed loop from site analysis to visualization report output, solving the technical problems of difficulty in predicting the effects of plant configuration and difficulty in intuitively presenting seasonal landscape changes.

[0013] This specification also provides a landscape design plant configuration image simulation and seasonal change visualization system for executing the above-mentioned methods. The system includes a site analysis module, a plant configuration module, an image fusion module, a seasonal simulation module, and a growth prediction and reporting module. The site analysis module acquires site image data and performs environmental element identification and light condition analysis; the plant configuration module retrieves plant model data from a plant model library and determines the configuration location; the image fusion module performs light-adaptive image fusion processing; the seasonal simulation module renders the seasonal landscape changes; and the growth prediction and reporting module performs long-term growth prediction and generates a visualization report. Through the collaborative work of its modules, this invention achieves complete functionality for plant configuration image simulation and seasonal change visualization. Each module corresponds one-to-one with the method steps, ensuring the integrity and consistency of the system implementation. Attached Figure Description

[0014] Figure 1This is a schematic diagram of the steps of the landscape design plant configuration image simulation and seasonal change visualization method of the present invention.

[0015] Figure 2 This is a schematic diagram of the architecture of a landscape design plant configuration image simulation and seasonal change visualization system according to the present invention. Detailed Implementation

[0016] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0017] refer to Figure 1 The diagram shown is a flowchart illustrating the steps of the landscape design plant configuration image simulation and seasonal change visualization method of the present invention. In this embodiment, the method includes the following steps: Step S1: Site image acquisition and environmental element identification.

[0018] Acquire site image data of the site to be designed. In one embodiment of the invention, the site image data can be photographs taken on-site by the designer or owner using a digital camera, smartphone, or other device. The image resolution is preferably between 1920×1080 pixels and 4096×2160 pixels to ensure sufficient detail for subsequent processing. The site image data can also be a view image rendered from a 3D model of the site, or a top-down view image obtained by drone aerial photography. In this embodiment, the color space of the site image data is the standard sRGB color space, and the bit depth is 8-bit or 16-bit.

[0019] Semantic segmentation of environmental elements is performed on site image data. This embodiment employs a pre-trained deep learning semantic segmentation network model. Preferably, this network model is based on the DeepLabV3+ architecture, using ResNet-101 as the backbone network, and is fine-tuned on a dataset containing landscape scenes. The input to the semantic segmentation network model is the site image data, and the output is a category label map with the same size as the input image. Each pixel in the category label map corresponds to a category label value. The categories defined in this embodiment include: undulating terrain areas (category label value 1), existing vegetation areas (category label value 2), structure areas (category label value 3), water areas (category label value 4), hard paved areas (category label value 5), sky areas (category label value 6), and configurable areas (category label value 7). Configurable areas refer to open areas in the site suitable for vegetation configuration, typically represented by bare soil, lawns, or simple greening.

[0020] Connectivity analysis is performed on the configurable region. This embodiment uses the eight-neighbor connectivity algorithm to label pixels with a label value of 7 in the category label image as connected components, and extracts the area parameters of each connected component. (Number of pixels), shape parameters (including aspect ratio) and roundness ) and position parameters (including centroid coordinates) (And bounding box parameters). Among the shape parameters, aspect ratio... Defined as the ratio of the longer side to the shorter side of the smallest bounding rectangle of a connected region, its value range is... ; Circularity Defined as ,in Let be the perimeter of the connected component, and its value range is . When the connected components are circular, the circularity is set to the maximum value of 1. The area, shape, and location parameters of each connected component are summarized to form configurable region feature data, which serves as a component of the environmental element identification results.

[0021] Illumination condition analysis is performed on the site image data. This embodiment first estimates the light source direction from the site image data. Preferably, a light source direction estimation method based on shadow analysis is used, specifically including: detecting shadow areas created by structures or existing vegetation in the site image data and extracting the direction information of the shadow edges; estimating the light source direction based on the shadow direction, the site's geographical location information, and the shooting time information (if available), where the light source direction is represented by an azimuth angle. and elevation angle Indicates azimuth angle The range of values ​​is The elevation angle is measured clockwise from true north (0°). The range of values ​​is , with the horizontal direction as 0° and measured upwards.

[0022] This embodiment further analyzes the light intensity distribution. The site image data is converted from the RGB color space to the Lab color space, and the L channel is extracted as brightness information. The L channel image is then partitioned statistically, dividing the image into [various sections]. One grid (in this embodiment) , ), calculate the average pixel brightness value within each grid. and standard deviation This generates a light intensity distribution map. Simultaneously, the global average brightness value of the site image data is extracted. and global brightness standard deviation .

[0023] This embodiment also extracts shadow distribution information. A threshold segmentation method is used to extract shadow regions in the L-channel image, with a threshold value... Set to 0.6 times the global average brightness value, that is Set the L channel value to less than The pixels are marked as shadow pixels, and a shadow mask image is generated. The light source direction information, light intensity distribution information, and shadow distribution information are summarized to form the lighting condition parameters.

[0024] Step S2: Plant model retrieval and configuration location.

[0025] The system receives plant species information specified in the design scheme. In one embodiment of the invention, the plant species information includes the plant's scientific name, Chinese name, plant type (tree, shrub, ground cover, flowering plant, etc.), and specifications (diameter at breast height, crown width, height, etc.). The design scheme can be input interactively through a user interface or imported from external design files.

[0026] The system retrieves the corresponding target plant model data from the plant model library based on the plant species information. In this embodiment, the plant model library is a pre-constructed database of 3D plant models and parameters, storing relevant data on commonly used landscape plants. The data structure of the plant model library includes: a plant basic information table (storing plant scientific name, Chinese name, family, genus, origin, suitable growing area, etc.), a plant 3D model table (storing the file paths of plant 3D geometric models, texture map files, model scale coefficients, etc.), a plant seasonal morphological parameter table (storing the set of seasonal morphological parameters), and a plant growth parameter table (storing growth prediction model parameters).

[0027] The target plant model data includes a 3D geometric model of the plant and texture maps. The 3D geometric model of the plant is stored in a polygonal mesh format, preferably OBJ or FBX format, and includes the geometric structure of the plant such as the trunk, branches, and leaves. The texture maps include diffuse maps, normal maps, and alpha maps, used to represent the color, detail, and translucency of the plant surface.

[0028] Obtain the set of seasonal morphological parameters corresponding to the target plant model data. In this embodiment, the data structure of the seasonal morphological parameter set is as follows: Spring budding morphological parameters include: leaf unfolding ratio coefficient The range of values ​​is This indicates the degree of leaf unfolding relative to its fully unfolded state; the hue value of the new leaf. The range of values ​​is For most deciduous plants, the hue value of new leaves is in to Between (yellowish-green to green); saturation value of new leaves The range of values ​​is , usually in to Between; Branch visibility coefficient The range of values ​​is This indicates the degree to which the branches are visible through the leaves.

[0029] Summer lush morphological parameters include: crown fullness coefficient The range of values ​​is This indicates the fullness of the summer crown shape compared to the standard crown shape; the hue value of mature leaves. The range of values ​​is For most plants, the hue value of mature leaves is... to Between (green); Saturation value of mature leaves The range of values ​​is , usually in to Between; Flowering period indicator A Boolean value indicating whether the plant flowers in summer.

[0030] Autumn color change morphology parameters include: initial hue value of color change. The range of values ​​is The hue value is usually the same as that of leaves that mature in summer; the hue value at the end of the color change is... The range of values ​​is For autumn foliage plants, usually in to Between (red to yellow); color change gradation period The unit is weeks, and the range of values ​​is... Initial proportion of fallen leaves The range of values ​​is This indicates the percentage of leaves that have already fallen when the leaves begin to fall at the end of autumn.

[0031] Winter leaf fall morphology parameters include: residual leaf ratio coefficient The range of values ​​is This indicates the proportion of leaves remaining on the tree during winter; branch texture markings. A string type pointing to the path of the branch texture map; evergreen indicator. A Boolean value; if true, the plant is an evergreen that retains its leaves during winter.

[0032] The plant placement locations are determined based on the configurable areas and design schemes identified in the environmental element identification results. In this embodiment, the plant placement points are mapped to the pixel coordinate system of the site image data according to the plant placement point information specified in the design scheme. If the design scheme uses relative coordinates (such as distance from the site boundary), coordinate transformation is performed based on the perspective relationship of the site image data. If a placement point falls outside the configurable area, the system issues a warning. Configuration location coordinate information is generated, including the pixel coordinates of each placement point. and estimated scene depth value The scene depth value is used to determine the rendering scale of the plant model.

[0033] Step S3: Illumination-adaptive image fusion.

[0034] The target plant model data is rendered to the corresponding location based on the configured location coordinates to generate a plant rendering image. This embodiment uses Physically Based Rendering (PBR) to render the plant's 3D geometric model. During rendering, the azimuth and elevation angles of the virtual light source are set according to the light source direction information in the lighting condition parameters to ensure that the rendered lighting matches the actual lighting direction in the site image data. The scene depth value is then used based on the configured location coordinates. Calculate the rendering scale factor of the plant model Ensure that the rendered plant size conforms to perspective. The plant rendering image is in RGBA format, containing RGB color channels and an alpha transparency channel. The alpha channel is used to represent the boundaries of the plant outline and the semi-transparent effect of the leaves.

[0035] The plant rendering image is subjected to illumination correction processing based on illumination condition parameters. This embodiment first extracts the average brightness value of the site from the site image data. and the variance of site brightness This was obtained through global statistical analysis of the L channel of the site image data. Then, the average brightness value of the plants in the rendered plant image was extracted. and plant brightness variance The data was obtained by converting the RGB channels of the plant rendering image to the L channels and then performing statistical analysis (only pixels with Alpha channel values ​​greater than 0 were counted).

[0036] The brightness correction factor is calculated based on the ratio of the average brightness value of the site to the average brightness value of the plants. : , in: This is the brightness correction factor, dimensionless, with a value range of [value range missing]. ; The average brightness value of the site, with a range of values ​​of [value missing]. (L channel range in Lab color space); This represents the average brightness value of the plant, with a range of values ​​ranging from [value missing]. .

[0037] The contrast correction factor is calculated based on the ratio of the site brightness variance to the plant brightness variance. : , in: This is the contrast correction factor, dimensionless, with a value range of [value range missing]. ; This represents the variance of site brightness. This represents the variance of plant brightness. The contrast correction factor is used to adjust the contrast of the rendered plant image to match the contrast characteristics of the site image.

[0038] The brightness correction factor and contrast correction factor are weighted and combined to generate the illumination consistency weight value. : , in: This is the dimensionless weighting value for illumination uniformity. The value of the brightness correction weighting coefficient is 0.6 in this embodiment; The contrast correction weighting coefficient is set to 0.4 in this embodiment; This is the upper limit of the brightness correction factor, which is set to 2.0 in this embodiment to prevent over-correction; This is the upper limit of the contrast correction factor; in this embodiment, it is set to 1.5. Illumination consistency weight value. Used to adjust the contribution ratio of the plant rendering image during subsequent image fusion.

[0039] Lighting correction is performed on the plant rendering image. The plant rendering image is converted from the RGB color space to the Lab color space, and the following correction formula is applied to the L channel: , in: These are the corrected pixel brightness values; This represents the original pixel brightness value. These are pixel coordinates. After correction, the image is converted back to the RGB color space to obtain a light-corrected plant rendering image.

[0040] An image fusion algorithm based on illumination consistency weights is used to fuse the illumination-corrected plant rendering image with the site image data. This embodiment employs an improved Poisson image fusion algorithm, and the specific steps are as follows: Construct gradient field data corresponding to the plant rendering image. Calculate the gradients in the horizontal and vertical directions of the illumination-corrected plant rendering image to obtain the gradient field. ,in and These represent the gradient operators in the horizontal and vertical directions, respectively.

[0041] The gradient field data is weighted and adjusted based on the illumination consistency weight value: , in: The adjusted gradient field; The gradient field of the fused region corresponding to the site image data; This is the weighting value for illumination consistency. This weighting adjustment makes it suitable for situations where illumination conditions vary significantly (...). (Smaller), the fusion result retains more of the gradient features of the site image, thus achieving a more natural fusion effect.

[0042] Set Dirichlet boundary conditions in the merging boundary region. Defined as the set of pixel locations in the plant rendering image where the alpha channel value changes from non-zero to zero, i.e., the boundary pixels of the plant outline. Boundary conditions are set as follows: , in: The resulting image after fusion; This is the site image data. This boundary condition ensures that the pixel values ​​at the boundaries of the fusion region are consistent with the site image, achieving a seamless transition.

[0043] Solve the Poisson equation with weighted gradient field constraints. In the fusion region. Solve the following Poisson equation within the set of pixels whose alpha channel value is greater than 0 in the plant rendering image: , in: For the Laplace operator; This is the divergence operator. In this embodiment, the Discrete Fourier Transform (DFT) method or the multigrid method is used to solve the Poisson equation to obtain the pixel value distribution of the fused region.

[0044] The merging boundary region is feathered. In this embodiment, the feathering width is adaptively determined based on the complexity of the plant outline. Plant outline complexity Defined as the ratio of the perimeter of the plant outline to the square root of the area of ​​the plant outline: , in: The perimeter of the plant outline (in pixels); This represents the area (in pixels) of the plant outline. The feathering width is adaptively determined based on the outline complexity. , in: The base feather width is set to 5 pixels in this embodiment; The complexity sensitivity coefficient is set to 0.5 in this embodiment; For reference complexity values, a value of 4.0 is used in this embodiment. For plants with high contour complexity (such as conifers), a larger feathering width is used to obtain a more natural edge transition effect.

[0045] Within the feathered region, the pixel values ​​of the fused image are calculated using the following formula: , in: The pixel values ​​obtained by solving the Poisson equation; The feathering weight is linearly transitioned from 1 (inner boundary of the fusion region) to 0 (outer boundary of the fusion region) within the feathering region.

[0046] After the above processing, fused image data is generated that seamlessly integrates the plants with the site background.

[0047] Step S4: Visualizing and rendering seasonal changes.

[0048] Seasonal changes are rendered in plant regions of fused image data based on a set of seasonal morphological parameters. This embodiment first defines seasonal time parameters. The value range is based on the month. These correspond to January through December, respectively. Based on time parameters... Determine the current seasonal phase: Spring corresponds to Summer correspondence Autumn corresponds Winter corresponds .

[0049] A color space mapping method is used to perform continuous gradient processing on plant leaf colors. This embodiment defines a seasonal gradient function. This is used to calculate the target hue and target saturation values ​​for the current month. For deciduous trees, the seasonal gradation function is defined as follows: Spring section ( Hue gradient function: , in: The target hue value for the current month, in degrees; This refers to the hue value of new leaves in the spring budding morphology parameters; This is the hue value of mature leaves in the summer foliage morphology parameters. This function gradually transitions the hue value from the yellowish-green of spring to the dark green of summer.

[0050] Summer section ( Hue preservation function: , Autumn Period ( Hue gradient function: , in: The initial hue value for the color change; The hue value at which the color change ends; The color-changing gradient cycle. When Exceed At that time, take This function causes the hue value to gradually transition from the green of summer to the red or yellow of autumn.

[0051] Winter zone ( Leaf treatment: If evergreen criteria If it is a pseudo-decaying plant, then the leaf transparency is determined based on the residual leaf ratio coefficient. Adjustments are made to make most of the leaves transparent, retaining only a few withered leaves; if the evergreen criterion is true (evergreen plant), then the leaves are preserved, and the hue value is maintained at [value missing]. nearby.

[0052] The definition of the saturation gradient function is similar to that of the hue gradient function, starting from a lower saturation level in the spring range (…). Gradually changing to summer saturation ( The value may increase or decrease in the autumn range depending on the plant characteristics (autumn foliage plants).

[0053] For evergreen trees, the seasonal gradation function maintains a relatively stable hue throughout the year, only applying a lighter adjustment to the new leaf area in spring. In this example, the geometric extent of the new leaf area is marked in the plant model data, and a light green adjustment is applied to the new leaf area during spring rendering. , in: The value is 15°, representing the hue shift of the new leaves, which makes the new leaves appear light green.

[0054] For flowering shrubs, a seasonal phase gradient function is overlaid with a flower color rendering layer during the flowering period. This embodiment uses flowering period information (starting month) stored in a plant model library. and the end month and flowering density parameters Flower color pixels are overlaid on the plant area within the flowering period. The distribution range of the flower color rendering layer is determined based on the flowering density parameter. A random sampling method is used to generate flower positions on the plant canopy surface, and the number of flowers is proportional to the flowering density parameter.

[0055] The pixels of the plant region in the fused image data are converted from the RGB color space to the HSV color space. For each plant pixel, the parameters of the current simulation month are used. Calculate the target hue value using the seasonal gradient function. and target saturation value Then perform the following color mapping: , , , in: The original pixel's HSV value; The mapped HSV value; and For reference hue and saturation values, the primary hue and primary saturation of the original texture of the plant model are usually taken; This is the hue mapping intensity coefficient, which is set to 0.8 in this embodiment to preserve some of the color change details of the original texture.

[0056] Convert the mapped HSV values ​​back to the RGB color space to complete the seasonal rendering for that month. Then proceed sequentially... Render the images, or select representative months (such as March, June, September, and December) to generate spring, summer, autumn, and winter landscape renderings, forming a seasonal rendering image sequence.

[0057] The plant outline is morphologically adjusted by incorporating parameters related to changes in plant crown shape. This embodiment uses the leaf unfolding ratio coefficient from a set of morphological parameters applicable to all four seasons. and crown fullness coefficient The plant's outline is scaled. In spring, the plant crown width is adjusted according to... The coefficients are scaled to simulate leaves that are not fully unfolded; in summer, the plant crown width is adjusted according to... The coefficients were adjusted to simulate the vigorous growth and full crown of summer plants. Morphological adjustments were made using a centroid-based proportional scaling method. , , in: These are the original pixel coordinates; These are the adjusted pixel coordinates; The coordinates of the centroid of the plant region; The scaling factor is taken as spring. Summer harvest .

[0058] Step S5: Generation of long-term growth forecast and visualization report.

[0059] Obtain the growth prediction model parameters corresponding to the target plant model data. In this embodiment, the growth prediction model parameters are stored in the plant growth parameter table of the plant model library, including: the asymptotic maximum value of the crown width. (Unit: m) Crown growth rate coefficient (Unit: 1 / year), Crown shape parameters (Dimensionless), height asymptotically maximum value (Unit: m) Height growth rate coefficient (Unit: 1 / year) and height shape parameters (Dimensionless). These parameters are obtained by fitting measured data from tree growth models, and different tree species have different parameter values.

[0060] The predicted crown width and height of the plant at preset time points are calculated based on the parameters of the growth prediction model. This embodiment uses the Chapman-Richards growth function as the parameterized growth prediction model, and the mathematical form of this function is as follows: Crown width growth function: , in: For the first time after planting Annual crown width forecast, in meters; The initial crown width at planting time, in meters, is obtained from the design plan or plant model data; The crown width is the asymptotic maximum value, expressed in meters, representing the maximum crown width that this tree species can achieve under ideal conditions; This is the crown growth rate coefficient, expressed in units of 1 / year, with a value range of [value missing]. A larger value indicates a faster growth rate; The crown shape parameter is dimensionless and its value range is [value range missing]. To control the shape of the growth curve; The time since planting is in years; is the base of the natural logarithm.

[0061] Height growth function: , in: For the first time after planting Annual height predictions, in meters; This represents the initial height at planting, in meters (m). This represents the asymptotic maximum height, expressed in meters (m). This is the height growth rate coefficient, expressed in units of 1 / year; This refers to the height and shape parameters.

[0062] In this embodiment, the preset time node includes the third year after the plant is planted ( ), the fifth year time node ( ) and the tenth year milestone ( For each time point, the corresponding predicted crown width value is calculated based on the growth function. and height prediction value .

[0063] Scale transformation is performed on plants in the seasonal rendering image sequence based on predicted crown width and height values. This embodiment calculates the scale transformation factor for each time point: Crown width scale transformation factor: , Height scale transformation factor: , For each preset time point, a non-uniform scaling transformation is performed on the plant regions in the seasonal rendering image sequence, with the horizontal direction (canopy direction) according to... Scaling is performed, vertically (height direction) according to... Scaling is performed. The scaling is anchored at the plant's root position, meaning the bottom of the plant remains unchanged while scaling upwards and to the sides.

[0064] To ensure image quality after scaling, this embodiment employs a bicubic interpolation algorithm for image resampling. For cases with large scale factors (such as...), A multi-level super-resolution method is used to first enhance the image of the plant region, and then scale transformation is performed to avoid image blurring.

[0065] Generate long-term landscape effect images corresponding to each preset time node. For each time node, output four seasonal effect images (spring, summer, autumn, and winter), generating a total of [number missing] images. Zhang Yuanqi landscape effect image (if three time points are selected).

[0066] The seasonal rendering image sequence and the long-term landscape effect image are integrated and output to generate a visualization report file. In this embodiment, the visualization report file is output in PDF or HTML format, and includes the following structure: The site status display section includes original photos of the site image data, labeled maps of environmental element identification results (with different colors used to mark terrain, vegetation, structures, configurable areas, etc.), and illustrated explanations of lighting condition parameters.

[0067] The plant configuration scheme display section includes: a map with coordinate information of the configuration location (marking the configuration location and number of each plant on the site image), a list of plant species (including botanical name, Chinese name, specifications, quantity, etc.), and a plant configuration effect diagram (integrating image data).

[0068] The seasonal landscape change display section includes a sequence of rendered images of the four seasons arranged in the order of spring, summer, autumn, and winter, with one representative image for each season, accompanied by seasonal labels and brief descriptions. Preferably, the continuous seasonal changes can be displayed using GIF animations or videos.

[0069] The long-term growth effect display section includes long-term landscape effect images corresponding to each preset time node (third year, fifth year, tenth year), plant size prediction data table (including crown width prediction value and height prediction value for each time node), and growth trend chart (showing the trend of plant size change over time in the form of a line graph).

[0070] The comparison display section includes before-and-after comparison images (left-right or overlay comparison of the original site photos and the configuration renderings), and growth comparison images at different time points (the renderings at the time of planting, the third year, the fifth year, and the tenth year are displayed side by side).

[0071] By following the steps above, a complete visual report is generated for owners and designers to review and make decisions.

[0072] refer to Figure 2 The diagram shown is a schematic representation of the architecture of the landscape design plant configuration image simulation and seasonal change visualization system of the present invention. The system provided in this embodiment is used to execute the landscape design plant configuration image simulation and seasonal change visualization method described in the foregoing method embodiments. The system includes a site analysis module, a plant configuration module, an image fusion module, a seasonal simulation module, and a growth prediction and reporting module.

[0073] The site analysis module is used to perform the functions of step S1, specifically including: site image acquisition unit, semantic segmentation unit, and illumination analysis unit.

[0074] The site image acquisition unit is used to acquire site image data of the site to be designed. This unit supports multiple image input methods, including importing image files from the local file system, directly acquiring images from a camera device, and downloading images from cloud storage. Preferably, this unit performs format checks and preprocessing on the input site image data, including color space conversion, size normalization, and noise reduction filtering, to ensure that subsequent processing modules receive input data of stable quality.

[0075] The semantic segmentation unit is used to perform semantic segmentation of environmental features on site image data. This unit incorporates a pre-trained deep learning semantic segmentation network model and supports GPU-accelerated inference. The semantic segmentation unit outputs environmental feature recognition results, including a category label map and configurable region feature data.

[0076] The illumination analysis unit is used to analyze illumination conditions in site image data. This unit includes a light source direction estimation submodule, a light intensity analysis submodule, and a shadow detection submodule. The illumination analysis unit outputs illumination condition parameters for use by the image fusion module.

[0077] The plant configuration module is used to perform the functions of step S2, specifically including: a design scheme receiving unit, a model retrieval unit, and a configuration positioning unit.

[0078] The design scheme receiving unit receives plant species information specified in the design scheme. This unit supports multiple input methods, including user interface interaction, design file import (supporting CAD format, XML format, etc.), and voice input. The design scheme receiving unit validates the input plant species information, including plant name matching checks and specification parameter rationality checks.

[0079] The model retrieval unit is used to retrieve corresponding target plant model data and seasonal morphological parameter sets from the plant model library based on plant species information. This unit connects to the plant model library database and supports queries by plant name, plant type, specifications, and other criteria. The plant model library can be deployed on local storage or a cloud server, and the model retrieval unit supports caching and incremental updates of model data.

[0080] The configuration and positioning unit determines the plant placement locations based on the configurable areas and design schemes identified in the environmental element recognition results. This unit receives configurable area feature data from the semantic segmentation unit and, combined with the placement point information from the design scheme, calculates and generates the coordinates of the placement locations. The configuration and positioning unit supports interactive adjustment of the placement locations; users can adjust the plant placement locations on the interface by dragging and dropping.

[0081] The image fusion module is used to perform the functions of step S3, specifically including: plant rendering unit, illumination correction unit and Poisson fusion unit.

[0082] The plant rendering unit is used to render the target plant model data to the corresponding location based on the configured location coordinates, generating a plant rendering image. This unit incorporates a physically-based rendering engine and supports GPU-accelerated rendering. The plant rendering unit sets virtual light source parameters according to lighting condition parameters to ensure that the rendered lighting matches the actual lighting on the site.

[0083] The illumination correction unit is used to perform illumination correction processing on the plant rendering image based on illumination condition parameters. This unit calculates the brightness correction factor and contrast correction factor, generates illumination consistency weight values, and performs illumination correction transformation on the plant rendering image.

[0084] The Poisson fusion unit employs an image fusion algorithm based on illumination consistency weights to fuse illumination-corrected plant rendering images with site image data. This unit handles gradient field construction, boundary condition setting, Poisson equation solving, and feathering, outputting fused image data. The Poisson fusion unit supports GPU parallel computing, enabling real-time or near-real-time fusion processing.

[0085] The seasonal phase simulation module is used to perform the functions of step S4, specifically including: seasonal phase parameter calculation unit, color mapping unit, and morphology adjustment unit.

[0086] The seasonal parameter calculation unit is used to calculate the target hue value, target saturation value, and other seasonal rendering parameters for the current month based on the parameters of the current simulated month and the set of seasonal morphological parameters. This unit has built-in multiple seasonal gradient functions, employing corresponding gradient rules for different plant types such as deciduous trees, evergreen trees, and flowering shrubs.

[0087] The Color Mapping Unit is used to perform continuous gradient processing of plant leaf colors using color space mapping methods. This unit performs operations such as RGB to HSV color space conversion, hue and saturation mapping transformation, and inverse HSV to RGB transformation. The Color Mapping Unit supports batch processing and can render multiple months simultaneously to generate a complete seasonal rendering image sequence.

[0088] The Morphology Adjustment unit is used to adjust the shape of plant outlines based on plant crown shape variation parameters. This unit performs a centroid-based scaling operation to remap the pixel coordinates of the plant region. The Morphology Adjustment unit supports real-time preview, allowing users to dynamically adjust seasonal parameters to view the changes in effect.

[0089] The growth prediction and reporting module is used to perform the functions of step S5, specifically including: a growth calculation unit, a scale transformation unit, and a report generation unit.

[0090] The growth calculation unit is used to calculate the predicted crown width and height of plants at preset time points based on the parameters of the growth prediction model. This unit incorporates the Chapman-Richards growth function and supports custom time point settings. The growth calculation unit can generate growth curve data, which can be used by the report generation unit to plot growth trend graphs.

[0091] The scaling unit performs scaling transformation on plants in the seasonal rendering image sequence based on predicted crown width and height values. This unit performs non-uniform scaling and image resampling operations to generate long-term landscape effect images for each preset time point. The scaling unit supports multi-level super-resolution enhancement to ensure image quality at large magnifications.

[0092] The report generation unit integrates seasonal rendered image sequences and long-term landscape effect images to generate a visual report file. This unit includes built-in report templates and supports PDF and HTML output formats. The report generation unit allows for customization of report content and layout, enabling users to selectively include or exclude specific sections.

[0093] This invention's system, through the collaborative work of its various modules, achieves complete functionality in plant configuration image simulation and seasonal change visualization. The output of the site analysis module (environmental element identification results and light condition parameters) serves as input to the plant configuration module and image fusion module. The output of the plant configuration module (target plant model data, four-season morphological parameter set, and configuration location coordinate information) serves as input to the image fusion module and seasonal change simulation module. The output of the image fusion module (fused image data) serves as input to the seasonal change simulation module. The output of the seasonal change simulation module (seasonal rendering image sequence) serves as input to the growth prediction and reporting module. The data flow between these modules forms a complete processing link, ensuring the integrity of the system's functions and the consistency of the data.

[0094] In one possible implementation, the system of the present invention is deployed on a cloud server and provides services to users through a web interface or mobile application client. Users upload site photos and design plans, the system performs calculations and processing in the cloud, and returns the generated visualization report to the user. This deployment method reduces the hardware requirements of the user end and supports concurrent access by multiple users.

[0095] In another possible implementation, the system of the present invention is deployed on the user's local computer, running as a plugin or standalone application for landscape design software. This deployment method ensures the privacy of user data and is suitable for users with high requirements for data security.

[0096] Compared with the prior art, the present invention has the following significant technical effects: First, this invention employs an image fusion technique based on site photographs, eliminating the need for complex 3D scene modeling and significantly reducing the workload of data preparation in the early stages of design. Designers only need to provide photographs of the current site conditions to simulate the effects of plant configurations, significantly improving design efficiency. Compared to existing methods that rely on building complete 3D models, this invention has lower input data acquisition costs and a simpler processing flow, making it more suitable for rapid iteration and comparison of landscape design schemes.

[0097] Secondly, this invention employs a lighting-adaptive image fusion algorithm. By calculating a lighting consistency weight value, it adjusts the fusion process between the rendered plant image and the site background, solving the technical problem of inconsistent lighting between the overlaid plant image and the background. Traditional image overlay methods easily produce obvious boundary marks and lighting inconsistencies, affecting the realism of the rendered image. This invention, through the combined action of brightness correction factors and contrast correction factors, ensures that the lighting characteristics of the rendered plant image are consistent with the site background, resulting in a more natural and realistic fusion effect.

[0098] Third, this invention employs the HSV color space mapping method to achieve continuous gradation of plant leaf color. Compared to existing technologies that use model replacement to achieve seasonal changes, this invention can generate a smooth transition of seasonal changes. In the HSV color space, the hue channel directly corresponds to the color category. By controlling the hue value gradient curve, the autumn color change process of plants from green to yellow and then to red can be accurately simulated. This continuous gradation processing method is more in line with the seasonal change patterns of plants in their natural environment, enhancing the realism and professionalism of the visualization effect.

[0099] Fourth, this invention introduces the Chapman-Richards growth prediction model, which can predict the size changes of plants at different future time points based on plant species and planting specifications, filling the gap in existing technology regarding the preview of long-term growth effects. When reviewing landscape design plans, owners can not only see the immediate effect at planting time, but also intuitively understand the growth status of plants three, five, and ten years later. This helps owners form accurate expectations of the long-term effects of the landscape design plan and assists them in making more rational decisions.

[0100] Fifth, this invention forms a complete closed loop from site analysis to visualized report output through the synergistic cooperation of four core technologies: environmental element identification, illumination-adaptive image fusion, HSV color space seasonal gradation, and parametric growth prediction. Each technical step is organically linked, with the output of one step serving as the input for the next, ensuring the data flow throughout the entire processing workflow and guaranteeing the integrity and consistency of the final output. Compared to the independent operation of each functional module, this closed-loop technical architecture offers higher system integration and a better user experience.

[0101] The technical solution of this invention can be applied to various landscape design-related application scenarios: In residential community landscape design, real estate developers need to showcase the future landscape of the community to homebuyers during the sales phase. Using the method of this invention, existing photos of the community can be integrated with planned plant arrangements to generate a visual report that includes seasonal changes and long-term growth effects. This helps homebuyers intuitively understand the landscape effects in different seasons after moving in, as well as the landscape state after the plants mature.

[0102] In park and green space design, municipal departments often need to solicit public opinions during park construction or renovation projects. Using the method of this invention, multiple plant configuration schemes can be quickly generated from on-site photos, showcasing the differences in landscape effects across spring, summer, autumn, and winter, providing an intuitive visual reference for public participation in decision-making.

[0103] In courtyard landscape design, private homeowners often have high expectations for the visualization of the design effect when commissioning landscape design. Using the method of this invention, designers can take photos of the current state of the courtyard, overlay the plant configuration from the design plan onto the photos, generate realistic renderings, and showcase the dynamic effects of plant changes with seasons and years, helping homeowners make design decisions.

[0104] In campus greening planning, schools need to showcase the effects of greening when implementing campus environment improvement projects. Using the method of this invention, planned street trees, landscape trees, flowering shrubs, etc., can be overlaid onto real-life campus photos to demonstrate the campus landscape effects in different seasons and the shady effects formed after the plants have grown, providing a reference for campus planning decisions.

[0105] In road landscape design, transportation departments need to evaluate the effects of plant configurations when designing road greening schemes. Using the method of this invention, the planned roadside trees can be overlaid onto real-world road photographs, showcasing the shading effect, seasonal changes, and long-term landscape effects of the trees, thus aiding in the comparison and decision-making process for road greening schemes.

[0106] To facilitate implementation of this invention by those skilled in the art, recommended configuration ranges for some key parameters are provided below: The recommended input image size for semantic segmentation networks is 512×512 pixels to 1024×1024 pixels. Too small an input size will reduce segmentation accuracy, while too large an input size will increase computation time. For 4K resolution site photos, it is recommended to first downsample to 1024×1024 pixels for semantic segmentation, and then upsample the segmentation results to the original resolution.

[0107] Luminance weighting coefficient in illumination consistency weighting value and contrast weighting coefficient Recommended value range is The sum of the two should equal 1. When the lighting conditions of the site photo are relatively uniform, the brightness weighting coefficient can be increased; when the site photo has strong light and dark contrast, the contrast weighting coefficient can be increased.

[0108] The recommended range for the feather width of Poisson fusion is: The specific value of the feather width is adaptively determined based on the complexity of the plant's outline. For plants with simple, round canopies, a smaller feather width can be used; for plants with complex, irregular canopies, a larger feather width should be used to achieve a smoother transition effect.

[0109] Hue mapping intensity coefficient in HSV color mapping Recommended value range is When the value is small, more detail in the color variations of the original texture is preserved, and the seasonal gradient effect is relatively soft; when the value is large, the seasonal gradient effect is more pronounced, but some texture detail may be lost. For high-definition plant models with rich textures, it is recommended to use a smaller value. For simplified plant models with simple textures, a larger value can be used. value.

[0110] The parameters of the Chapman-Richards growth function need to be calibrated for specific plant species. For common landscape plants, parameter fitting results from forestry research literature can be referenced. The plant model library should have default values ​​for the growth parameters of commonly used landscape plants, while also allowing users to adjust the parameters according to local growing conditions.

[0111] The embodiments of the present invention are not limited to the specific embodiments described above. Those skilled in the art can make various equivalent changes or substitutions based on the technical solutions of the present invention, and all such changes or substitutions should be included within the protection scope of the present invention.

Claims

1. A method for simulating plant configuration images and visualizing seasonal changes in landscape design, characterized in that, Includes the following steps: Step S1: Obtain site image data of the site to be designed, perform semantic segmentation of environmental elements on the site image data, identify configurable areas and generate environmental element identification results; Illumination condition analysis is performed on site image data to extract light source direction information and light intensity distribution information, and to generate illumination condition parameters; Step S2: Retrieve target plant model data from the plant model library based on the plant species information in the design scheme; obtain the set of seasonal morphological parameters corresponding to the target plant model data; determine the plant configuration location based on the configurable area in the environmental element identification results, and generate configuration location coordinate information; Step S3: Render the target plant model data to generate a plant rendering image based on the configured location coordinate information, perform lighting correction on the plant rendering image based on the lighting condition parameters, and calculate the lighting consistency weight value. An image fusion algorithm based on illumination consistency weight values ​​is used to fuse plant rendered images and site image data to generate fused image data. The image fusion algorithm includes: constructing gradient field data corresponding to the plant rendered image; weighting and adjusting the gradient field data according to the illumination consistency weight values ​​to obtain the adjusted gradient field; setting Dirichlet boundary conditions in the fusion boundary region formed by the intersection pixel set of the plant rendered image and the site image data; and solving the Poisson equation with the adjusted gradient field constraints to obtain the pixel value distribution of the fusion region. Step S4: Render seasonal changes in the plant region of the fused image data according to the four seasons morphological parameter set. Use a color space mapping method to perform continuous gradient processing on the plant leaf color, and combine this with crown shape change parameters for morphological adjustment to generate a seasonal rendering image sequence. The continuous gradient processing of plant leaf color using the color space mapping method includes: converting the pixels of the plant region in the fused image data from RGB color space to HSV color space; calculating the target hue value corresponding to the current simulated month using a seasonal gradient function based on the current simulated month parameters and the hue change parameters in the four seasons morphological parameter set, and calculating the target saturation value corresponding to the current simulated month based on the current simulated month parameters and the saturation change parameters in the four seasons morphological parameter set; adjusting the hue and saturation components of the plant region pixels to the target hue and saturation values ​​respectively, while keeping the lightness component unchanged, and then converting back to RGB color space. Step S5: Calculate the predicted crown width and height of the plants at the preset time nodes based on the growth prediction model parameters, perform scale transformation on the plants in the seasonal rendering image sequence to generate a long-term landscape effect image; integrate the seasonal rendering image sequence and the long-term landscape effect image and output them to generate a visualization report file. Step S3, calculating the illumination consistency weight value, includes: extracting the average brightness value and the variance value of the site image data, and extracting the average brightness value and the variance value of the plant rendering image; calculating a brightness correction factor based on the ratio of the average brightness value of the site to the average brightness value of the plant, and calculating a contrast correction factor based on the ratio of the variance value of the site brightness to the variance value of the plant brightness; and weighting the brightness correction factor and the contrast correction factor to generate the illumination consistency weight value, which is used to adjust the contribution ratio of the plant rendering image during the image fusion process.

2. The method for simulating landscape plant configuration images and visualizing seasonal changes according to claim 1, characterized in that, Step S1, which performs semantic segmentation of environmental elements on the site image data, includes: inputting the site image data into a pre-trained semantic segmentation network model, which outputs the category label corresponding to each pixel in the site image data; dividing the site image data into topographic relief areas, existing vegetation areas, structure areas, water areas, hard paving areas, and configurable areas based on the category labels; and performing connected component analysis on the configurable areas to extract the area parameters, shape parameters, and location parameters of each connected component as configurable area feature data.

3. The method for simulating landscape plant configuration images and visualizing seasonal changes according to claim 1, characterized in that, The data structure of the four seasons morphological parameter set in step S2 includes: spring budding morphological parameters including leaf unfolding ratio coefficient, new leaf hue value, new leaf saturation value, and branch visibility coefficient; summer lush morphological parameters including crown fullness coefficient, mature leaf hue value, mature leaf saturation value, and flowering period indicator; autumn color change morphological parameters including color change start hue value, color change end hue value, color change gradation cycle, and leaf fall start ratio; and winter leaf fall morphological parameters including residual leaf ratio coefficient, branch texture indicator, and evergreen determination indicator.

4. The method for simulating landscape plant configuration images and visualizing seasonal changes according to claim 1, characterized in that, The image fusion algorithm in step S3 further includes: feathering the fusion boundary region, with the feathering width adaptively determined based on the complexity of the plant outline.

5. The method for simulating landscape plant configuration images and visualizing seasonal changes according to claim 1, characterized in that, The seasonal gradient function described in step S4 employs different mapping rules for different plant types: For deciduous trees, the seasonal gradient function uses an increasing curve to simulate the leaf unfolding process in the spring interval, a hue shift curve to simulate the color change process in the autumn interval, and gradually changes the leaf transparency to a preset leaf fall threshold in the winter interval; For evergreen trees, the seasonal gradient function maintains a stable hue throughout the year, only adjusting the color of the new leaf area; For flowering shrubs, the seasonal gradient function overlays a flower color rendering layer during the flowering period, and the distribution range of the flower color rendering layer is determined according to the flowering density parameter.

6. The method for simulating landscape plant configuration images and visualizing seasonal changes according to claim 1, characterized in that, In step S5, the preset time nodes include the third year, the fifth year, and the tenth year after planting; the growth prediction model parameters include the asymptotic maximum value of crown width, the crown width growth rate coefficient, the crown width shape parameter, the asymptotic maximum value of height, the height growth rate coefficient, and the height shape parameter; based on the growth prediction model parameters, the parametric growth function is used to calculate the predicted crown width and height values ​​for each preset time node. The input of the parametric growth function is the planting time length, and the output is the predicted size value for the corresponding time.

7. The method for simulating landscape plant configuration images and visualizing seasonal changes according to claim 1, characterized in that, The visualization report includes: a site status display section, containing labeled maps of site image data and environmental element identification results; a plant configuration scheme display section, containing labeled maps of configuration location coordinates and a list of plant species; a seasonal landscape change display section, containing a sequence of seasonal rendering images arranged in the order of spring, summer, autumn, and winter; a long-term growth effect display section, containing long-term landscape effect images and plant size prediction data tables corresponding to each preset time node; and a comparison display section, containing before-and-after configuration comparison images and growth comparison images at different time nodes.

8. A landscape design plant configuration image simulation and seasonal change visualization system, characterized in that, For performing the landscape design plant configuration image simulation and seasonal change visualization method as described in claim 1, the system comprises: The site analysis module is used to acquire site image data of the site to be designed, perform semantic segmentation of environmental elements on the site image data, identify the terrain undulation area, existing vegetation area, structure area and configurable area in the site image data, and generate environmental element identification results; and perform illumination condition analysis on the site image data to extract light source direction information, light intensity distribution information and shadow distribution information, and generate illumination condition parameters. The plant configuration module is used to receive plant species information specified in the design scheme, retrieve the corresponding target plant model data from the plant model library based on the plant species information, obtain the set of seasonal morphological parameters corresponding to the target plant model data, determine the plant configuration location based on the configurable area and design scheme in the environmental element identification results, and generate configuration location coordinate information. The image fusion module is used to render the target plant model data to the corresponding location based on the configured location coordinates to generate a plant rendering image, perform illumination correction processing on the plant rendering image according to illumination condition parameters, and calculate the illumination consistency weight value; and use an image fusion algorithm based on the illumination consistency weight value to fuse the illumination-corrected plant rendering image with the site image data to generate fused image data. The image fusion algorithm includes: constructing gradient field data corresponding to the plant rendering image; weighting and adjusting the gradient field data according to the illumination consistency weight value; setting Dirichlet boundary conditions in the fusion boundary region formed by the intersection pixel set of the plant rendering image and the site image data; and solving the Poisson equation with the weighted gradient field constraints to obtain the pixel value distribution of the fusion region. The seasonal phase simulation module is used to render seasonal changes of plant areas in the fused image data according to the set of seasonal morphological parameters. It uses a color space mapping method to perform continuous gradient processing on the plant leaf color, and combines the plant crown shape change parameters to adjust the plant outline to generate a seasonal rendering image sequence. The continuous gradient processing of plant leaf color using the color space mapping method includes: converting the pixels of the plant area in the fused image data from RGB color space to HSV color space; calculating the target hue value and target saturation value according to the current simulation month parameters and the set of seasonal morphological parameters through the seasonal phase gradient function; adjusting the hue component and saturation component of the plant area pixels to the target hue value and target saturation value respectively while keeping the lightness component unchanged, and then converting back to RGB color space. The growth prediction and reporting module is used to calculate the predicted crown width and height of plants at preset time nodes based on the parameters of the growth prediction model. Based on the predicted crown width and height, the module performs scale transformation processing on the plants in the seasonal rendering image sequence to generate long-term landscape effect images. The module integrates the seasonal rendering image sequence and the long-term landscape effect images and outputs them to generate a visualization report file.

Images