A brown land landscape evaluation and design method based on multi-source data

Abstract

This invention relates to the field of landscape evaluation and design technology, and particularly to a brownfield landscape evaluation and design method based on multi-source data. The method includes: collecting multi-source data of the brownfield site; dividing the site into several spatial units; constructing a spatial correlation analysis model to identify the influence domain of brownfield problems and calculate a comprehensive problem index; evaluating the brownfield landscape and delineating problem areas; establishing a historical case database; retrieving historical cases similar to the current site; extracting design strategies to form a candidate strategy set; performing compatibility checks based on the strategy synergy coefficient matrix and determining the strategy combination; establishing a three-dimensional parametric model based on the strategy combination; establishing a constraint hierarchy system; optimizing the model parameters; generating a three-dimensional scene and visually annotating the constraint satisfaction status; and outputting a brownfield landscape design scheme. This invention constructs a complete technical chain from problem evaluation to three-dimensional parametric design, realizing data-driven decision-making in brownfield landscape design.

Description

Technical Field

[0001] This invention relates to the field of landscape evaluation and design technology, and in particular to a brownfield landscape evaluation and design method based on multi-source data. Background Technology

[0002] With the acceleration of urbanization and industrial restructuring, large amounts of abandoned or inefficiently used land, polluted and ecologically damaged by industrial and mining activities, have become a significant challenge for urban renewal. While environmental science-based ecological baseline assessment methods possess objectivity and quantifiability in pollutant concentration monitoring and risk assessment, their evaluation index systems are entirely focused on environmental safety and ecological resilience, lacking assessment of the post-restored site's human-centered usability, social appeal, and psychological comfort. This often results in restored brownfield sites meeting ecological standards but lacking user appeal. Human-centered experience assessment methods based on design and sociology directly address user feelings through post-use evaluation questionnaires or behavioral observation. However, their data sources heavily rely on users' subjective recollections and expressive abilities, and these assessments often occur after project completion, making it difficult to effectively predict and optimize project plans still in the design phase.

[0003] Existing brownfield assessment methods primarily employ the analytic hierarchy process (AHP) or fuzzy comprehensive evaluation, integrating multi-source information such as soil monitoring data, remote sensing imagery, and social survey data through weighted summation. These methods have significant limitations: First, they only evaluate the severity of problems within a single spatial unit, neglecting the spatial diffusion characteristics of the problems, such as the impact of pollution sources on surrounding areas; second, they simply add up or weighted average multiple problems, ignoring the synergistic effects between problems, such as the mutual exacerbation of pollution and ecological degradation when they coexist. In the design phase, the industry commonly relies on renderings, physical models, or 3D animations for visual representation. While virtual reality technology has been introduced in recent years to provide immersive experiences, the core functions of these technologies remain at the level of presentation and communication. Designers and decision-makers evaluate the merits of solutions based on personal experience and aesthetic judgment, lacking objective data feedback mechanisms. This results in a lack of precise basis for design optimization, leading to significant uncertainty and trial-and-error costs in design decisions. In terms of 3D parametric design and constraint solving, existing software supports parametric modeling and geometric constraint solving, but it is mainly designed for architectural design scenarios and lacks the ability to express the environmental and ecological constraints unique to brownfields. Furthermore, the constraint solving algorithm adopts a local optimization strategy and does not consider the spatial transmission effect of constraints when adjusting individual model parameters, which leads to cyclical conflicts during the iteration process.

[0004] Chinese patent CN110796375A discloses a method for screening and arranging ecological technologies in urban relocation site landscaping. This method achieves ecological technology screening and layout through steps such as site data import, design goal setting, site analysis, and export. It collects historical and current site data, establishes an intelligent analysis database to construct a database corresponding to keywords and pollution sources, and a database corresponding to classified production enterprises and pollution sources. It assesses pollution sources through a multi-dimensional environmental pollution assessment database and establishes an environmental pollution and remediation scheme database for screening and evaluating remediation schemes. This method can screen and evaluate ecological technologies and layout schemes according to design objectives. However, this method focuses on pollution source identification and remediation technology matching, and its assessment of brownfield landscapes is not comprehensive enough. Furthermore, it lacks effective technical support in transforming the assessment results into specific landscape design schemes. Summary of the Invention

[0005] In view of this, the present invention provides a brownfield landscape evaluation and design method based on multi-source data to solve the technical problems in the prior art, such as the lack of quantitative evaluation ability of spatial correlation and synergistic effects of brownfield landscape assessment, the reliance on experience in design strategy selection and the lack of systematic matching methods, and the lack of effective technical support means from site assessment to design scheme generation.

[0006] The technical solution of this invention is implemented as follows: On the one hand, this invention provides a brownfield landscape evaluation and design method based on multi-source data, including: S1. Collect multi-source data of the brownfield site, divide the site into several spatial units and establish a unit attribute database; S2. Based on spatial units and multi-source data, a spatial correlation analysis model is constructed to identify the impact domain of brownfield problems. A collaborative reinforcement evaluation model is used to calculate the comprehensive problem index, evaluate the brownfield landscape, and delineate the problem area. S3. Establish a historical case database, use case reasoning methods to retrieve historical cases similar to the current site, extract design strategies to form a candidate strategy set, and conduct compatibility tests based on the strategy synergy coefficient matrix to determine the strategy combination. S4. Establish a three-dimensional parametric model based on the strategy combination, establish a constraint hierarchy, use the constraint transmission method to dynamically adjust the constraint requirements according to the problem's influence domain, and optimize the model parameters through an iterative adjustment algorithm; S5. Generate a 3D scene and visualize the constraint satisfaction status, then output the brownfield landscape design scheme.

[0007] Based on the above technical solutions, preferably, in step S1, the multi-source data includes soil pollution monitoring data, remote sensing image data, geographic information data, field survey data, and social survey data. The soil pollution monitoring data includes the coordinates of monitoring points, pollutant types, and concentration values. The remote sensing image data includes multispectral images and normalized difference vegetation index (NDVI). The geographic information data includes road networks, plot boundaries, and surrounding land use properties. The field survey data includes the locations of abandoned buildings and damaged areas of the ground. The social survey data includes heat maps of surrounding residents' needs and pedestrian activity.

[0008] Based on the above technical solutions, preferably, step S1, which involves dividing the site into several spatial units and establishing a unit attribute database, specifically includes: dividing the site into irregular polygonal units using natural or artificial boundaries as dividing lines. The natural boundaries include rivers and mountains, and the artificial boundaries include roads and walls. For large areas without clear boundaries, regular grids are used for division, and the grid size is determined based on the density of pollution monitoring points. Each spatial unit is assigned a unique number, and a spatial unit attribute table is established. The spatial unit attribute table includes the unit area, center point coordinates, and adjacency matrix with adjacent units.

[0009] Based on the above technical solutions, preferably, step S2 specifically includes: S21. For five types of problems in brownfields, namely pollution remediation, visual obstruction, ecological fragmentation, insufficient accessibility and safety hazards, the spatial impact domain of each type of problem is calculated separately. A diffusion model is used to calculate the impact intensity of the problem source on the surrounding spatial units. The cumulative impact intensity of a unit is the sum of all problem sources within the site. S22. For a spatial unit with multiple problems, establish a coupled evaluation model based on the synergistic reinforcement effect to calculate the comprehensive problem index of the unit. The synergistic reinforcement effect takes into account the mutual aggravation effect between different problem types. S23. Set a threshold based on the comprehensive problem index, mark the spatial units that exceed the threshold as problem areas, calculate the contribution rate of each problem type, select the problem type with the largest contribution rate as the dominant problem type, and record other problem types whose contribution rates exceed the set proportion as secondary problem types.

[0010] Based on the above technical solutions, the preferred coupling evaluation model in step S22 based on the synergistic reinforcement effect is: ; in, It is a comprehensive problem index for spatial units. For the number of question types, For the first The intensity of the impact of this type of problem It is a non-linear exponent. For the synergy coefficient, For question type With type The correlation coefficient, This is the saturation rate parameter.

[0011] Based on the above technical solutions, preferably, step S3 specifically includes: S31. Establish a historical case library, with each case containing a problem feature vector, a strategy feature vector, and an evaluation of the implementation effect; S32. Extract the problem feature vector of the current site, use the weighted nearest neighbor algorithm to calculate the similarity with the cases in the case library, automatically increase the weight of the comprehensive problem index feature for sites with particularly high comprehensive problem index, select the cases with the highest similarity, and extract the design strategies used in the cases to form a candidate strategy set. S33. Establish a strategy synergy coefficient matrix, calculate the implementation strength of each strategy based on the priority weight and synergy relationship of the strategy, and perform a compatibility test on the strategy combination. When there is serious incompatibility, adjust the strategy combination. S34. Based on the determined strategy combination and its implementation strength, call the corresponding 3D model and perform initial layout in the 3D scene.

[0012] Based on the above technical solutions, preferably, in step S31, the problem feature vector includes the dominant problem type, comprehensive problem index, problem area area, surrounding land use nature and terrain slope, and the strategy feature vector includes the design strategy type, the implementation intensity of each strategy and the spatial layout pattern of the strategy.

[0013] Based on the above technical solutions, preferably, step S4 specifically includes: S41. Establish a constraint hierarchy system, dividing constraints into three levels: hard constraints, soft constraints, and optimization constraints. Assign priority weights to each constraint. Hard constraints include pollution isolation constraints, safety distance constraints, and mandatory regulatory requirements. Soft constraints include plant spacing constraints, walkway width constraints, and structure size constraints. Optimization constraints include maximizing green visibility, minimizing restoration costs, and maintaining landscape aesthetic symmetry. S42. Establish a spatial transmission model of constraints, and dynamically adjust the constraint requirements at different locations based on the distribution of the problem's influence intensity. S43. Perform constraint verification on the current parameters of each model. When a constraint is violated, treat each constraint as an elastic tension applied to the model parameters. Take into account the priority weights and stiffness coefficients of multiple constraints and calculate the parameter adjustment amount. S44. Use an iterative algorithm to adjust parameters in a coordinated manner. During the iteration process, consider the impact of changes in model parameters on surrounding models. When constraint conflicts still exist after the iteration ends, relax soft constraints and optimization constraints according to priority.

[0014] Based on the above technical solution, preferably, in step S42, the spatial transmission model of the constraint is as follows: for the coordinates... The facilities at that location must meet a safe distance based on the coordinates. The intensity of pollution impact on a location is dynamically determined, with coordinates... The required safe distance is as follows: ; in, coordinates Safety distance requirements at the location, The preset minimum safe distance, The pollution sensitivity coefficient is the number of facilities. coordinates The intensity of the cumulative impact of pollution on the corresponding spatial unit.

[0015] Based on the above technical solution, preferably, in step S44, the parameter linkage iterative adjustment algorithm specifically includes: initializing the current parameters of all models as initial parameters, setting the maximum number of iterations and the convergence threshold; in each iteration, sequentially calculating the current values ​​of all constraints on each model, determining whether there is a constraint violation, if so, using the tension balance model to calculate the parameter adjustment amount and update the parameters, calculating the spatial unit corresponding to the adjusted model position and recalculating the constraint transmission value, checking whether the parameter change of the model affects other models, marking the models within the influence domain as pending adjustment; calculating the global parameter change amount, and terminating the iteration when it is less than the convergence threshold or reaches the maximum number of iterations.

[0016] The present invention has the following advantages over the prior art: This invention constructs a complete technical chain from problem evaluation to 3D parametric design. Through the synergistic effect of three core technologies—spatial correlation analysis, case reasoning, and constraint transmission—data-driven decision-making in brownfield landscape design is achieved. Compared to existing design methods that rely on experience-based judgment, this invention introduces a spatial diffusion model in the problem identification stage to quantify the impact of problem sources on surrounding areas, establishes a compatibility verification mechanism in the strategy matching stage to avoid strategy conflicts, and achieves coordinated solutions under multiple constraints through a constraint transmission model in the parameter optimization stage. This ensures both ecological restoration effectiveness and functional and safety requirements are met, thereby improving the scientific rigor and feasibility of the design scheme.

[0017] This invention employs a problem influence domain identification method based on a diffusion model, overcoming the limitation of existing evaluation methods that only focus on the problems of a single spatial unit. By establishing a collaborative reinforcement evaluation model, it quantitatively calculates the interaction effects between multiple problem types, solving the technical problem that simple weighted summation methods cannot reflect the collaborative escalation of problems. This method can accurately delineate high-risk areas and identify dominant problem types, providing a reliable basis for the targeted configuration of subsequent design strategies and avoiding design failures caused by incomplete problem identification.

[0018] This invention establishes a dynamic matching mechanism for design strategies based on a historical case database. It achieves rapid retrieval of similar sites through a weighted nearest neighbor algorithm and automatically adjusts feature weights according to the problem severity of the current site, improving the accuracy of strategy recommendations. By introducing a strategy synergy coefficient matrix and a compatibility testing algorithm, it can quantitatively analyze the synergistic and competitive relationships between multiple strategies when they are used concurrently. Through iterative solving, it determines the optimal implementation strength of each strategy, effectively addressing the technical shortcomings of existing methods that rely on experience for strategy selection and lack quantitative basis for resource allocation.

[0019] This invention establishes a constraint hierarchy and spatial transmission model, enabling dynamic adjustment of constraint requirements based on the intensity of the problem's influence, thus overcoming the limitations of existing parametric design tools that use fixed constraint values. Employing a tension balance-based parameter linkage adjustment algorithm, it considers the impact of a model parameter change on surrounding models during iteration, avoiding constraint conflict loops caused by local optimization and improving the convergence of solutions under multiple constraints. Through hierarchical processing and compromise strategies for constraints of different priorities, it maximizes the satisfaction of soft and optimization constraints while meeting hard constraints, enhancing the overall coordination of the design scheme. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart of the brownfield landscape evaluation and design method of the present invention; Figure 2 This is a structural diagram of the collaborative reinforcement evaluation model of the present invention; Figure 3 This is a flowchart of the case reasoning and strategy matching process of the present invention; Figure 4 This is a schematic diagram of the constraint hierarchy and transmission optimization of the present invention. Detailed Implementation

[0022] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0023] like Figure 1 As shown, this invention provides a brownfield landscape evaluation and design method based on multi-source data, including: S1. Collect multi-source data of the brownfield site, divide the site into several spatial units and establish a unit attribute database; S2. Based on spatial units and multi-source data, a spatial correlation analysis model is constructed to identify the impact domain of brownfield problems. A collaborative reinforcement evaluation model is used to calculate the comprehensive problem index, evaluate the brownfield landscape, and delineate the problem area. S3. Establish a historical case database, use case reasoning methods to retrieve historical cases similar to the current site, extract design strategies to form a candidate strategy set, and conduct compatibility tests based on the strategy synergy coefficient matrix to determine the strategy combination. S4. Establish a three-dimensional parametric model based on the strategy combination, establish a constraint hierarchy, use the constraint transmission method to dynamically adjust the constraint requirements according to the problem's influence domain, and optimize the model parameters through an iterative adjustment algorithm; S5. Generate a 3D scene and visualize the constraint satisfaction status, then output the brownfield landscape design scheme.

[0024] In one embodiment of the present invention, step S1 specifically includes: after inputting the brownfield site boundary vector data, collecting multi-source data of the site, including soil pollution monitoring data, remote sensing image data, geographic information data, field survey data, and social survey data. Soil pollution monitoring data includes the coordinates of monitoring points, pollutant types, and concentration values; the data source is a site environmental investigation report or actual sampling analysis. Remote sensing image data includes multispectral images and the Normalized Difference Vegetation Index (NDVI), acquired through satellite remote sensing or UAV aerial photography, with an image resolution of no less than 2 meters, used to identify vegetation cover and surface changes. Geographic information data includes road networks, plot boundaries, and surrounding land use properties, obtained in vector format from urban planning departments. Field survey data includes the locations of abandoned buildings and damaged areas of the ground; building outlines and damaged areas are recorded using GPS positioning. Social survey data includes heat maps of surrounding residents' needs and pedestrian activity, obtained through questionnaire surveys and mobile signal data analysis.

[0025] The specific method for dividing the site into several spatial units is as follows: using natural or artificial boundaries as dividing lines, the site is divided into irregular polygonal units. Natural boundaries include rivers and mountains, while artificial boundaries include roads and walls. For large areas without clear boundaries, regular grid division is used. The grid size is determined based on the density of pollution monitoring points. When the monitoring point density is higher than 4 points per hectare, the grid size is 20m × 20m; when the monitoring point density is 2 to 4 points per hectare, the grid size is 30m × 30m; and when the monitoring point density is lower than 2 points per hectare, the grid size is 50m × 50m. When dividing the grid, if the boundary of a unit happens to pass through a monitoring point, that monitoring point is assigned to the unit with the larger area.

[0026] Assign a unique number to each spatial unit. ,in M represents the total number of units. Establish a spatial unit attribute table, including unit area. Center point coordinates Adjacency matrix with neighboring cells The unit area is obtained through the geometric calculation function in GIS software. The center point coordinates are taken as the geometric center of the unit polygon. The adjacency matrix is ​​defined as follows: if unit m and unit n have a common boundary, then... ,otherwise Units that are only connected at their vertices are not considered adjacent. The attribute table is stored in a relational database, supporting subsequent spatial queries and association analysis. Step S1 outputs the set of partitioned spatial units. and its attribute table.

[0027] In one embodiment of the present invention, such as Figure 2 As shown, the specific process of step S2 includes: based on the spatial unit set output in step S1... The measured values ​​of each evaluation index in each unit of the multi-source data.

[0028] S21. For five types of issues in brownfields, namely pollution remediation, visual obstruction, ecological fracture, insufficient accessibility and safety hazards, calculate the spatial impact domain of each type of issue, and use a diffusion model to calculate the impact intensity of the issue source on the surrounding spatial units. The cumulative impact intensity of a unit is the sum of all issue sources within the site.

[0029] Taking pollution remediation as an example, for spatial units If the concentration of a certain pollutant in its soil is Exceeding the background value The pollution intensity of this unit is defined as follows: ; in The risk screening value for this pollutant was determined with reference to the screening values ​​for Class I or Class II land use in the "Soil Environmental Quality Standard for Construction Land Soil Pollution Risk Control (Trial)" GB 36600-2018. For brownfield reuse scenarios, the Class I land use screening value is preferred to ensure the safety of subsequent land use. Background value This data was obtained from soil monitoring data of uncontaminated areas surrounding the site.

[0030] The impact intensity of the pollution source on surrounding units was calculated using a diffusion model. For units located a certain distance from the pollution source... The center distance is unit The intensity of the impact it receives is: ; in This is the spatial attenuation coefficient of pollutants, set according to the type of pollutant: heavy metal pollutants diffuse slowly into space. Organic pollutants are highly volatile. . This is a correction factor for groundwater flow direction, determined through site hydrogeological investigation. If the unit... lie in Downstream direction of groundwater This reflects the aggravating effect of pollutants migrating with groundwater; otherwise... .unit The cumulative impact intensity of pollution is the sum of all pollution source units within the site: .

[0031] To address the issue of visual occlusion, a view domain analysis algorithm is employed to calculate the impact intensity. Abandoned buildings or structures within the site are identified as occlusion sources. A building outline model is constructed in the 3D scene, with the viewing height set at 1.6 meters (human eye height). Rays are projected from the center point of each spatial unit outwards to the surrounding areas, and the range of views obstructed by the building is calculated. The proportion of the obstructed area is used as the visual occlusion impact intensity for that unit. The value ranges from 0 to 1.

[0032] To address the issue of ecological fragmentation, the landscape connectivity index is used for evaluation. Vegetation patches within the site are identified, using areas with an NDVI greater than 0.3 in remote sensing imagery as vegetation patches, and the shortest distance between any two patches is calculated. For spatial units located between patches, the intensity of their ecological fracture impact is: The greater the distance, the worse the connectivity, and the stronger the impact.

[0033] To address the accessibility issue, a network analysis algorithm is used to calculate the shortest path length to each spatial unit, starting from the site entrances and exits. The intensity of the impact is 500 meters is the acceptable walking distance threshold. For safety hazards, areas of surface damage, steep slopes (greater than 25°), and waterlogged areas are identified, with the impact intensity of these areas set as... The surrounding area is calculated based on distance attenuation.

[0034] S22. For spatial units with multiple problems simultaneously, establish a coupled evaluation model based on synergistic reinforcement effects to calculate the comprehensive problem index of the unit. Let the spatial unit... Simultaneously affected The impact of the problem type, and the intensity of the impact of each problem are as follows: Then the coupled evaluation model based on the synergistic reinforcement effect is: ; in, L represents the comprehensive problem index of a spatial unit, where L is the number of problem types. For the first The intensity of the impact of this type of problem It is a non-linear exponent, with a value ranging from 1.2 to 1.8, which is preferred in this invention. This demonstrates the amplifying effect when a problem becomes severe. The synergy coefficient ranges from 0.2 to 0.5, and is preferably defined in this invention. , For question type With type The correlation coefficient is set based on the interaction mechanism of the brownfield problem. The hyperbolic tangent function reflects the saturation effect of correlation. The saturation rate parameter takes a value from 1.5 to 3.0, which is preferred in this invention. .

[0035] Correlation coefficient The coefficient is set based on the physical or ecological correlation mechanisms between different problem types, determined through expert evaluation and statistical analysis of historical cases. When there is a strong interaction between two types of problems, the correlation coefficient is set to a higher value. For example, the correlation coefficient between pollution and ecological fracture is high because the lack of vegetation leads to exposed soil, which accelerates the diffusion of pollutants; the correlation coefficient between pollution and obstruction is second, as building obstruction hinders air circulation and impedes the natural volatilization of volatile pollutants; the correlation coefficient between ecological fracture and inadequate accessibility is moderate, as the lack of trails prevents connectivity between vegetation patches; the correlation coefficient between inadequate accessibility and safety hazards is high, as obstructed paths make it difficult for people to evacuate dangerous areas in a timely manner. The specific values ​​are determined in practical applications based on site characteristics and problem types using expert scoring or the Delphi method. The first term in the formula reflects the independent cumulative effect of each problem, while the second term reflects the synergistic reinforcement effect between problems through geometric mean and hyperbolic tangent function. When the influence intensity of both types of problems is high and the correlation is strong, the contribution of the synergistic term to the comprehensive index increases significantly.

[0036] A threshold is set based on a comprehensive problem index, and spatial units exceeding the threshold are marked as problem areas. Threshold Determined based on the actual site conditions, generally taken as... Add a standard deviation to the mean of the site's overall problem index, or set it to 1.5 times the maximum impact intensity of any single problem, according to project management requirements. That is, when the overall problem index exceeds 1.5 times the maximum impact intensity of any single problem, it is identified as a problem area requiring focused attention. Calculate the contribution rate of each problem type: ; Select the question type with the highest contribution rate as the dominant question type, denoted as . Other question types with a contribution rate exceeding 0.2 are recorded as a secondary question type set. This is used to guide subsequent multi-strategy combination design. Step S2 outputs the comprehensive problem index for each spatial unit. Dominant problem types Secondary problem types set and the distribution of influence intensity .

[0037] The improvement of this evaluation model compared to the traditional weighted summation method is that it uses a diffusion model to characterize the spatial correlation of the problem, so that the pollution source not only affects its own unit, but also has a decaying effect on the surrounding units; and it uses a synergistic reinforcement model to quantify the mutual aggravation effect when multiple problems coexist, avoiding the underestimation caused by simple addition.

[0038] In one embodiment of the present invention, such as Figure 3 As shown, the specific process of step S3 includes: S31. Establish a historical case database. Each case includes a problem feature vector, a strategy feature vector, and an evaluation of implementation effectiveness. The problem feature vector includes the dominant problem type, comprehensive problem index, problem area area, surrounding land use, and terrain slope. The dominant problem type is a categorical variable, taking values ​​such as pollution remediation, visual obstruction, ecological fracture, insufficient accessibility, or safety hazard. The comprehensive problem index and problem area area are numerical variables, recording the case site's... The average value and total area of ​​the problem area are included. Surrounding land use is a categorical variable, including residential areas, commercial areas, industrial areas, parks, and green spaces. Topographic slope is a numerical variable, recording the average slope of the site. The strategy feature vector includes the design strategy type, the implementation intensity of each strategy, and the spatial layout pattern of the strategies. Design strategy types include vegetation restoration, engineering barriers, activity areas, ecological corridors, and terrain modification. The implementation intensity of each strategy is expressed as a percentage of occupied area or investment. The spatial layout pattern of the strategies is divided into centralized and decentralized. The evaluation of implementation effectiveness includes a comprehensive score of indicators such as restoration compliance rate, improvement in ecological connectivity, and public satisfaction, serving as a reference for case quality. Each case in the case library is denoted as […]. ,in This represents the total number of cases. Case data comes from completed brownfield landscape projects, collected through literature review and field visits.

[0039] S32. Extract the problem feature vector of the current site. The weighted nearest neighbor algorithm is used to calculate the similarity with cases in the case library. For sites with particularly high comprehensive problem index, the weight of the comprehensive problem index feature is automatically increased. Several cases with the highest similarity are selected, and the design strategies used in the cases are extracted to form a candidate strategy set.

[0040] The specific steps for similarity calculation are as follows: For the first feature vector in the problem... Each feature component defines local similarity. For categorical variables (such as dominant problem type, surrounding land use), if the target site is exactly the same as the case, then... If they belong to the same broad category (e.g., pollution remediation and safety hazards both fall under the category of environmental pollution issues), then ,otherwise For numerical variables (such as the comprehensive problem index, problem area, and terrain slope), normalized difference calculation is used: in Let q be the feature value of the target site. Let c be the feature value of the q-th problem in the current case, and c in the denominator is calculated by iterating through all cases in the case library. ), and These are the maximum and minimum values ​​of the q-th feature in the case library, respectively. The difference between them is the global range of that feature, used to eliminate the influence of dimensions. Case The overall similarity with the target site is: ; Where Q is the dimension of the problem feature vector. Let be the weight of the q-th feature. Let be the local similarity of the q-th feature. The weights are determined using the analytic hierarchy process (AHP): the weight of the dominant problem type is... The highest weighting value is 0.35, because the problem type directly determines the applicability of the design strategy; the comprehensive problem index is the next highest, taking a weighting value of 0.35. =0.25, reflecting the severity of the problem; the area weight of the problem region is... =0.2; the weight of surrounding land use is =0.12; terrain slope weight is =0.08. The weights satisfy the normalization condition. .

[0041] For venues with a particularly high comprehensive problem index, i.e. At that time, the weight of the comprehensive problem index features is automatically increased. Several cases with the highest similarity are selected; this invention selects the top 5 cases (i.e., Extract the design strategies used in these cases to form a candidate strategy set. , where G is the number of candidate strategies. If a strategy appears repeatedly in the first 5 cases, the frequency of that strategy is recorded as a reference for the initial allocation strength.

[0042] S33. Establish a strategy synergy coefficient matrix, calculate the implementation strength of each strategy based on the priority weight and synergy relationship of the strategy, and perform a compatibility test on the strategy combination. When there is serious incompatibility, adjust the strategy combination.

[0043] Establish a strategy synergy coefficient matrix Its elements Representation Strategy With strategy The synergistic relationship between strategies is determined through expert evaluation and statistical analysis of historical case effects. When two strategies work together to significantly improve remediation effectiveness or resource utilization efficiency, the synergistic coefficient is positive, reflecting the strength of the synergy. When two strategies compete for the same space or resource, the synergistic coefficient is negative. When two strategies are independent, the synergistic coefficient is zero. For example, phytoremediation combined with ecological corridors can enhance pollutant absorption and ecological connectivity, showing a positive synergistic relationship. Engineering barriers combined with terrain modification can form physical isolation zones, also showing a positive synergistic relationship. Engineering barriers occupy space and reduce activity area, indicating competition, resulting in a negative synergistic coefficient. Large-area phytoremediation and high-density activity areas compete for space, resulting in a negative synergistic coefficient. Trail paving and lighting installation do not interfere with each other, resulting in a zero synergistic coefficient. The synergistic coefficient matrix is ​​a symmetric matrix. diagonal elements .

[0044] The implementation strength of each strategy is calculated based on its priority weight and collaborative relationships. Let the strategy be... The initial allocation strength is Its value is determined based on two factors: first, the severity of the problem; second, a higher initial strength is assigned to the strategy corresponding to the dominant problem. For example, if the dominant strategy for pollution remediation is phytoremediation, then... Second, the average implementation intensity of the strategy in the case library is calculated by averaging the implementation intensity of the strategy across the five retrieved cases. The optimized formula for calculating implementation intensity is: ; in For strategy Optimized implementation intensity For strategy The initial allocation strength, For strategy The priority weights are G, where G is the number of candidate strategies. For strategy index ( ), The competition intensity index ranges from 1.1 to 1.5, and is preferably defined in this invention. This reflects a high-priority strategy that prioritizes resource acquisition. The coefficient of sensitivity is a co-sensitivity factor, with a value ranging from 0.3 to 0.7. Preferably, this is the case in this invention. , For strategy With strategy The synergy coefficient (defined in the strategy synergy coefficient matrix). For strategy Optimized implementation intensity To find the normalized maximum value, take The strength formula is an implicit equation because... When both sides of the equals sign appear simultaneously, an iterative method is used to solve the problem. Initialization Substitute into the right side of the above equation to calculate. Repeat the iteration until ,in This serves as the convergence criterion. After iterative convergence, for... Perform normalization processing to make This ensures that the sum of the implementation strengths of all strategies is 1.

[0045] The compatibility of strategy combinations is then tested, and the compatibility index between any two strategies is calculated. If it exists If the inconsistencies are not met, the strategy combination is deemed severely incompatible and requires adjustment. The adjustment rule is as follows: retain the higher-priority strategies and replace the lower-priority strategies with the strategies based on the second-highest similarity cases in the case retrieval, or reduce their implementation intensity. Re-perform compatibility checks until all strategies meet the requirements. .

[0046] S34. Based on the determined strategy combination and its implementation intensity, call the corresponding 3D model and perform initial placement in the 3D scene. The 3D model comes from a pre-established parametric model library, which is organized by function, including a vegetation restoration model library (categorized by pollution tolerance level, canopy width, and height), a structure model library (pergolas, fences, pavilions, etc.), a public facility model library (walkways, seating, lighting, etc.), and a terrain modification model library (slopes, terraces, rain gardens, etc.). Each parametric model encapsulates adjustable parameters, including geometric parameters (length, width, height, radius) and positional parameters (insertion point coordinates, rotation angle), as well as constraint interface parameters (minimum spacing, distance from the boundary, prohibited placement area range).

[0047] For strategy According to its implementation intensity Calculate the number of models or the area required for instantiation. For example, the implementation intensity of a phytoremediation strategy is... Then, within the problem area, according to The area is planted with pollution-tolerant vegetation, among which This represents the total area of ​​the problem region. The number of model instances is determined based on vegetation density standards: 3 to 5 trees per 100 square meters, 8 to 12 shrubs per 100 square meters, and ground cover plants are configured for 100% coverage. For engineering barrier strategies, the length and width of the barrier strips are determined based on the implementation intensity. For activity area strategies, the number of point facilities such as seats and playground equipment is allocated based on the implementation intensity.

[0048] The initial layout method in the 3D scene is as follows: For planar features (such as green spaces and paving), a space-filling algorithm is used to distribute them evenly within the problem area to ensure coverage of the specified area; for linear features (such as walkways and ecological corridors), the skeleton lines of the problem area are extracted, and the layout is carried out along the skeleton lines, which are extracted using Voronoi diagrams or median transformation algorithms; for point features (such as structures and facilities), a mesh sampling method is used to distribute them evenly within the problem area, with the mesh spacing calculated based on the number of models and the area, or a random point-scattering method is used, followed by Poisson disk sampling to remove points that are too close together. Step S3 outputs the determined set of design strategies. and its implementation intensity A collection of parameterized models instantiated in a 3D scene (in ) and its initial parameters .

[0049] This strategy matching method avoids the limitations of a fixed rule base through case-based reasoning, enabling dynamic retrieval of applicable strategies based on site characteristics; it quantifies the synergistic and competitive relationships between strategies through a strategy synergy coefficient matrix, achieving optimized configuration of multiple strategies; and it avoids strategy conflicts through compatibility testing, improving the feasibility of the solution.

[0050] In one embodiment of the present invention, such as Figure 4 As shown, the specific process of step S4 includes: S41. Establish a constraint hierarchy system, dividing constraints into three levels: hard constraints, soft constraints, and optimization constraints. Assign priority weights to each constraint. Hard constraints include pollution isolation constraints, safety distance constraints, and mandatory regulatory requirements. Soft constraints include plant spacing constraints, walkway width constraints, and structure size constraints. Optimization constraints include maximizing green visibility, minimizing restoration costs, and maintaining landscape aesthetic symmetry.

[0051] Hard constraints are constraints that must be satisfied. Pollution isolation constraints are expressed as: for any active facility model... (Such as seats, amusement facilities), the coordinates of its center point With any pollution source unit distance Must meet ,in This refers to the minimum safe distance required for this type of facility. Safety distance constraints require that walkways and activity areas be at least 5 meters away from steep slopes and areas prone to flooding. Mandatory regulatory requirements include building setback distances and fire lane widths, which are determined according to relevant standards and regulations.

[0052] Soft constraints are constraints that allow for appropriate deviations. The plant spacing constraint is expressed as: for any two plant models... and The distance between their center points Should meet ,in , These are the crown radii of the two plants, with an allowable deviation range of [value missing]. ,Right now The width of the trails must be at least 2.5 meters for main trails and at least 1.5 meters for secondary trails, with allowable deviations. Meters. The dimensions of the structure are constrained by ergonomics, such as a seat height of 0.40 to 0.45 meters and a pergola clear height of not less than 2.2 meters.

[0053] The optimization constraints are those that should be satisfied as much as possible. Maximizing green view rate requires calculating the proportion of green vegetation occupying the field of view at key viewpoints, with a target value of no less than 30%. Minimizing restoration costs involves establishing an objective function based on the unit area cost of each strategy, reducing total investment while meeting restoration requirements. Landscape aesthetic symmetry requires that landscape elements in important node spaces exhibit axial or central symmetry.

[0054] For each constraint (in , Define priority weights for the total number of constraints. Hard constraints soft constraints The value ranges from 0.3 to 0.6. A value of 0.6 is used for soft constraints with strong functional requirements (such as walkway width), while a value of 0.3 is used for soft constraints related to aesthetics (such as structure color). Optimization constraints... The value ranges from 0.1 to 0.2.

[0055] S42. Establish a spatial transmission model for constraints, and dynamically adjust the constraint requirements at different locations based on the distribution of the problem's influence intensity. Specifically, the spatial transmission model for constraints is as follows: for constraints located at coordinates... The facilities at that location must meet a safe distance based on the coordinates. The intensity of pollution impact at a location is dynamically determined. Let coordinates be used. The corresponding spatial unit is The intensity of its cumulative pollution impact is (Calculated from step S2), then the coordinates The required safe distance is as follows: ; in The preset minimum safety distance is determined according to relevant regulations and actual site conditions, and generally ranges from 15 to 30 meters. The pollution sensitivity coefficient is denoted by facility type. Children's activity facilities (such as swings and slides) are highly sensitive to pollution, with larger sensitivity coefficients; adult rest facilities (such as seats and pergolas) are moderately sensitive, with medium-sized sensitivity coefficients; and service facilities (such as trash cans and signs) are low sensitive, with smaller sensitivity coefficients. This model reflects the spatial transmission effect of constraints: areas with higher pollution intensity require stricter safety distance requirements; areas far from pollution sources can have more relaxed constraints. For example, when the cumulative pollution impact is high, the safety distance requirement for children's activity facilities increases accordingly; when the pollution impact is low, the safety distance requirement is relatively lenient.

[0056] For ecological connectivity constraints, a minimum cost path model is adopted. It requires that there be a distance of at least [width value missing] between any two green patches. The connecting corridors must not pass through high-pollution areas, meaning the pollution impact intensity of all units along the path must meet the requirements. The paths connecting the corridors are planned using the A* algorithm, with the corridor length and the pollution intensity of the traversed units weighted as the cost function.

[0057] S43. Perform constraint verification on the current parameters of each model. When a constraint is violated, treat each constraint as an elastic tension applied to the model parameters. Consider the priority weights and stiffness coefficients of multiple constraints to calculate the parameter adjustment. In the 3D scene, for each model instance... Current parameters Perform constraint verification. Suppose model i is subject to constraints. The effect of a constraint is denoted as... (in For each constraint Define its target value (Expected value), Current value (Actual value calculated based on current parameters), tolerance (Allowable deviation range). If a constraint violation occurs, i.e. Then the parameter adjustment algorithm will be activated.

[0058] Treating each constraint as an elastic tension applied to the model parameters, a parameter adjustment model based on tension balance is established. The parameter adjustment amount for model i is... The calculation formula is ; in Let k and j be the number of constraints on model i, and k and j be the constraint indices. ), To constrain The priority weights inherit the weight values ​​defined in the aforementioned constraint hierarchy, and the denominator... The sum of all constraint weights for model i is used for normalization. To constrain Stiffness coefficient, hard constraint soft constraints The value can be between 2 and 5. To constrain The target value, To constrain The current value (calculated based on the current parameters). To constrain The required parameter adjustment direction is a unit vector. For example, spacing constraints require adjusting the position coordinates along the line connecting the two models, while size constraints require adjusting the radius or height parameters. To constrain The tolerance (allowable deviation range) is defined by the exponential term within square brackets, which is the saturation factor. When the deviation is significantly greater than the tolerance, it approaches 1 (full adjustment); when it approaches the target value, it approaches 0 (stop adjustment to avoid oscillations). In this formula, each constraint applies a tension force to the model. The magnitude of the tension force is proportional to the degree of constraint violation, priority weight, and stiffness coefficient. The direction of the tension force is the shortest path direction that satisfies the constraint. The tension forces of multiple constraints are normalized according to their weights and then superimposed to form a resultant force. The model parameters are adjusted along the direction of this resultant force.

[0059] S44. Use an iterative algorithm to adjust parameters in a coordinated manner. During the iteration process, consider the impact of changes in model parameters on surrounding models. When constraint conflicts still exist after the iteration ends, relax soft constraints and optimization constraints according to priority.

[0060] The parameter-linked iterative adjustment algorithm specifically includes: initializing the current parameters of all models as initial parameters, i.e. Set the maximum number of iterations. The present invention takes Set convergence threshold The present invention takes In each iteration, each model i (i ranging from 1 to ...) is processed sequentially. ): Calculate the current values ​​of all constraints imposed on model i. Determine if there is a constraint violation; if so, calculate the parameter adjustment amount using the aforementioned tension balance model. And update parameters Calculate the spatial elements corresponding to the adjusted model positions and recalculate the constraint transmission values. For example, for models involving pollution isolation constraints, recalculate... To check whether changes in the parameters of this model affect other models, the influence region is defined as a region centered on model i with a radius of [missing information]. For all models j within the range, this invention takes... The model within the influence domain is marked as pending adjustment.

[0061] Calculate the change in global parameters When it is less than the convergence threshold Or reach the maximum number of iterations Terminate the iteration when the time is right, otherwise let Return to the iteration process and continue adjusting.

[0062] If constraint conflicts still exist after the iteration, soft constraints and optimization constraints are relaxed according to priority. If unsatisfactory constraint conflicts still exist after the iteration, a compromise strategy is initiated. Identify conflicting constraint pairs: If model i is adjusted to satisfy the constraints... This leads to constraints. Violation, then Conflict pairs are formed. Constraints within conflict pairs are prioritized, ensuring that high-priority constraints (hard constraints) are always satisfied, while lower-priority constraints (soft constraints, optimization constraints) are relaxed. The relaxation of soft constraints uses a proportional compromise approach: for an original objective value of... Tolerance is If the soft constraint cannot be met, the tolerance will be increased to... ,in To relax the proportions, for functional soft constraints For aesthetic soft constraints Simultaneously, the relaxed constraints and their degree of relaxation are recorded as annotation information for the design scheme, prompting designers to focus on them during subsequent manual intervention. Step S4 outputs the optimized model parameter set. Constraint satisfaction report (listing satisfied constraints and constraints that need to be relaxed), and 3D scene model file.

[0063] In this invention, a constraint transmission model is used to establish a spatial correlation between constraint requirements and the distribution of site problems, enabling dynamic adjustment of constraints; a tension balance model and parameter linkage algorithm are used to achieve global optimization under multiple constraints, avoiding cyclical conflicts caused by local adjustments; and a constraint layering and compromise strategy are used to balance soft constraints and optimization constraints while satisfying hard constraints, thereby improving the feasibility of the solution.

[0064] In one embodiment of the present invention, step S5 specifically includes: based on the optimized 3D scene model and constraint satisfaction report output in step S4, loading the optimized parametric model into the 3D viewport to support multi-view roaming. The 3D viewport provides a bird's-eye view (viewing from a 45-degree angle above the site), a human-view view (a first-person view at a height of 1.6 meters), and a section view (displaying site cross-sections along specified cutting lines). Designers can control view switching and roaming paths using a mouse or keyboard.

[0065] Generate a 3D scene and visualize the constraint satisfaction status. Areas that meet hard constraints are shown with a green semi-transparent overlay, indicating that the design of that area meets mandatory requirements; areas requiring relaxation of soft constraints are marked in yellow, showing the relaxation percentage; areas with conflicts are marked in red with a warning and the type of conflicting constraint indicated. The visual annotations are overlaid on the 3D scene as layers, and designers can selectively show or hide them.

[0066] The system provides a manual intervention interface, allowing designers to directly drag and drop models in the 3D scene to adjust their position, or modify geometric parameters (such as plant canopy width and structure height) through the parameter panel. After each manual adjustment, the system automatically triggers constraint verification, providing real-time feedback on constraint compliance. If the adjustment results in a new constraint violation, the violated constraint is highlighted in the viewport along with suggested modifications. If designers deem certain constraints too strict, they can temporarily relax them through the constraint management panel, for example, by reducing the tolerance for plant spacing from... Adjusted to The system then re-optimized the parameters and updated the 3D scene.

[0067] Generate design deliverables, exporting 2D drawings (plans, sections, DWG or PDF format), 3D model files (Sketchup or Revit format), a bill of quantities (automatically calculating plant quantity, paving area, structure specifications, and material usage), and an improvement effect evaluation report (comparing pollution coverage, greening rate, accessibility indicators, etc. before and after the design). Output editable 3D design schemes, 2D drawings, bill of quantities, and evaluation reports for construction and acceptance use.

[0068] In this invention, problem identification based on spatial correlation is achieved through steps S1 and S2, and the output influence intensity distribution provides a data foundation for constraint transmission modeling in step S4; step S3 matches design strategies based on case reasoning and calls the parameterized model, and its output strategy implementation strength and synergy coefficient matrix are used to guide priority determination when there is constraint conflict in step S4; step S4 achieves parameter linkage adjustment based on constraint transmission and tension balance, dynamically adjusts constraint requirements using the problem influence domain calculated in step S2, and outputs a three-dimensional design scheme that satisfies multiple constraints; step S5 provides visualization and interactive interfaces to support designers to perform manual optimization based on algorithm assistance, realizing human-computer collaborative design.

[0069] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A brownfield landscape evaluation and design method based on multi-source data, characterized in that, include: S1. Collect multi-source data of the brownfield site, divide the site into several spatial units and establish a unit attribute database; S2. Based on spatial units and multi-source data, a spatial correlation analysis model is constructed to identify the impact domain of brownfield problems. A collaborative reinforcement evaluation model is used to calculate the comprehensive problem index, evaluate the brownfield landscape, and delineate the problem area. S3. Establish a historical case database, use case reasoning methods to retrieve historical cases similar to the current site, extract design strategies to form a candidate strategy set, and conduct compatibility tests based on the strategy synergy coefficient matrix to determine the strategy combination. S4. Establish a three-dimensional parametric model based on the strategy combination, establish a constraint hierarchy, use the constraint transmission method to dynamically adjust the constraint requirements according to the problem's influence domain, and optimize the model parameters through an iterative adjustment algorithm; S5. Generate a 3D scene and visualize the constraint satisfaction status, then output the brownfield landscape design scheme.

2. The brownfield landscape evaluation and design method based on multi-source data as described in claim 1, characterized in that, In step S1, the multi-source data includes soil pollution monitoring data, remote sensing image data, geographic information data, field survey data, and social survey data. The soil pollution monitoring data includes the coordinates of monitoring points, pollutant types, and concentration values. The remote sensing image data includes multispectral images and normalized difference vegetation index (NDVI). The geographic information data includes road networks, plot boundaries, and surrounding land use properties. The field survey data includes the locations of abandoned buildings and damaged areas of the ground. The social survey data includes heat maps of surrounding residents' needs and pedestrian activity.

3. The brownfield landscape evaluation and design method based on multi-source data as described in claim 1, characterized in that, Step S1, which involves dividing the site into several spatial units and establishing a unit attribute database, specifically includes: dividing the site into irregular polygonal units using natural or artificial boundaries as dividing lines. The natural boundaries include rivers and mountains, and the artificial boundaries include roads and walls. For large areas without clear boundaries, regular grids are used for division, and the grid size is determined based on the density of pollution monitoring points. Each spatial unit is assigned a unique number, and a spatial unit attribute table is established. The spatial unit attribute table includes the unit area, center point coordinates, and adjacency matrix with adjacent units.

4. The brownfield landscape evaluation and design method based on multi-source data as described in claim 1, characterized in that, Step S2 specifically includes: S21. For five types of problems in brownfields, namely pollution remediation, visual obstruction, ecological fragmentation, insufficient accessibility and safety hazards, the spatial impact domain of each type of problem is calculated separately. A diffusion model is used to calculate the impact intensity of the problem source on the surrounding spatial units. The cumulative impact intensity of a unit is the sum of all problem sources within the site. S22. For a spatial unit with multiple problems, establish a coupled evaluation model based on the synergistic reinforcement effect to calculate the comprehensive problem index of the unit. The synergistic reinforcement effect takes into account the mutual aggravation effect between different problem types. S23. Set a threshold based on the comprehensive problem index, mark the spatial units that exceed the threshold as problem areas, calculate the contribution rate of each problem type, select the problem type with the largest contribution rate as the dominant problem type, and record other problem types whose contribution rates exceed the set proportion as secondary problem types.

5. The brownfield landscape evaluation and design method based on multi-source data as described in claim 4, characterized in that, The coupled evaluation model based on the synergistic reinforcement effect in step S22 is as follows: ； in, It is a comprehensive problem index for spatial units. For the number of question types, For the first The intensity of the impact of this type of problem It is a non-linear exponent. For the synergy coefficient, For question type With type The correlation coefficient, This is the saturation rate parameter.

6. The brownfield landscape evaluation and design method based on multi-source data as described in claim 1, characterized in that, Step S3 specifically includes: S31. Establish a historical case library, with each case containing a problem feature vector, a strategy feature vector, and an evaluation of the implementation effect; S32. Extract the problem feature vector of the current site, use the weighted nearest neighbor algorithm to calculate the similarity with the cases in the case library, automatically increase the weight of the comprehensive problem index feature for sites with particularly high comprehensive problem index, select the cases with the highest similarity, and extract the design strategies used in the cases to form a candidate strategy set. S33. Establish a strategy synergy coefficient matrix, calculate the implementation strength of each strategy based on the priority weight and synergy relationship of the strategy, and perform a compatibility test on the strategy combination. When there is serious incompatibility, adjust the strategy combination. S34. Based on the determined strategy combination and its implementation strength, call the corresponding 3D model and perform initial layout in the 3D scene.

7. The brownfield landscape evaluation and design method based on multi-source data as described in claim 6, characterized in that, In step S31, the problem feature vector includes the dominant problem type, comprehensive problem index, problem area area, surrounding land use nature and terrain slope, and the strategy feature vector includes the design strategy type, the implementation intensity of each strategy and the spatial layout pattern of the strategy.

8. The brownfield landscape evaluation and design method based on multi-source data as described in claim 1, characterized in that, Step S4 specifically includes: S41. Establish a constraint hierarchy system, dividing constraints into three levels: hard constraints, soft constraints, and optimization constraints. Assign priority weights to each constraint. Hard constraints include pollution isolation constraints, safety distance constraints, and mandatory regulatory requirements. Soft constraints include plant spacing constraints, walkway width constraints, and structure size constraints. Optimization constraints include maximizing green visibility, minimizing restoration costs, and maintaining landscape aesthetic symmetry. S42. Establish a spatial transmission model of constraints, and dynamically adjust the constraint requirements at different locations based on the distribution of the problem's influence intensity. S43. Perform constraint verification on the current parameters of each model. When a constraint is violated, treat each constraint as an elastic tension applied to the model parameters. Take into account the priority weights and stiffness coefficients of multiple constraints and calculate the parameter adjustment amount. S44. Use an iterative algorithm to adjust parameters in a coordinated manner. During the iteration process, consider the impact of changes in model parameters on surrounding models. When constraint conflicts still exist after the iteration ends, relax soft constraints and optimization constraints according to priority.

9. A brownfield landscape evaluation and design method based on multi-source data as described in claim 8, characterized in that, In step S42, the spatial transmission model of the constraint is specifically as follows: for the coordinates located at... The facilities at that location must meet a safe distance based on the coordinates. The intensity of pollution impact on a location is dynamically determined, with coordinates... The required safe distance is as follows: ； in, coordinates Safety distance requirements at the location, The preset minimum safe distance, The pollution sensitivity coefficient is the number of facilities. coordinates The intensity of the cumulative impact of pollution on the corresponding spatial unit.

10. The brownfield landscape evaluation and design method based on multi-source data as described in claim 8, characterized in that, In step S44, the parameter linkage iterative adjustment algorithm specifically includes: initializing the current parameters of all models as initial parameters, setting the maximum number of iterations and the convergence threshold; in each iteration, calculating the current value of all constraints on each model in turn, determining whether there is a constraint violation, and if so, using the tension balance model to calculate the parameter adjustment amount and update the parameters, calculating the spatial unit corresponding to the adjusted model position and recalculating the constraint transmission value, checking whether the parameter change of this model affects other models, and marking the models within the influence domain as pending adjustment; calculating the global parameter change amount, and terminating the iteration when it is less than the convergence threshold or the maximum number of iterations is reached.

Images