CN116167661A - Land utilization change simulation credibility assessment method based on space dislocation - Google Patents

Abstract

The invention relates to a land utilization change simulation credibility assessment method based on space dislocation, which comprises the following steps: collecting a land utilization classification map and a driving factor map for constructing a CA model; acquiring a land transition probability map based on an intelligent optimization algorithm, and establishing a CA model; simulating the city pattern by using the CA model and outputting and storing a simulation result; performing single numerical evaluation on the simulation result of the CA model by adopting a pixel-by-pixel comparison method; respectively generating a spatial distribution map of a simulation result state-static index and a simulation result state-change index by adopting a spatial dislocation-moving window analysis method; and outputting and storing the simulation result and the spatial distribution diagram as a credibility evaluation result. Compared with the prior art, the invention has the advantages that an evaluation chart for reporting the simulation precision/error of each pixel can be generated, the spatial distribution of the simulation precision and the error is displayed, the spatial heterogeneity of the precision can be captured, the reliability of the CA model and the remote sensing classifier can be detected, and the like.

Description

A Credibility Evaluation Method for Land Use Change Simulation Based on Spatial Misalignment

technical field

The invention relates to the field of land use change simulation, in particular to a method for assessing the credibility of land use change simulation based on spatial dislocation.

Background technique

The cellular automata (CA) model is the most commonly used spatial simulation method for simulating and predicting the dynamic growth of cities. The reliability of the CA model is the basis for its simulation of historical patterns and prediction of future scenarios. In CA-based urban growth simulations, model performance evaluation is usually affected by the accuracy of data sources and driving factors, the choice of CA parameters, and the effectiveness of evaluation methods. Therefore, evaluating the simulation accuracy is very important for judging whether the CA model can satisfy urban planning and decision-making. Evaluation methods for model simulation results usually include visual discrimination, pixel-by-pixel comparison, spatial structure evaluation, and landscape index calculation.

For each simulation result map, when using pixel-by-pixel comparison to generate state-static and state-change indicators, each evaluation indicator corresponds to a numerical value. The values of these indicators are closely related to the modeling method and the study area, and often cannot accurately estimate the accuracy of the simulation. Since the simulation results often produce spatial heterogeneity, the obtained accuracy results will also produce similar spatial heterogeneity. For CA-based simulations, an evaluation metric value that is globally high may lead to low simulation accuracy in local regions, while an evaluation metric value that is globally low may lead to high simulation accuracy in local regions. Therefore, a single numerical evaluation cannot reflect the simulation accuracy across space, resulting in the inability to evaluate the simulation performance in local regions. Therefore, the key problem that needs to be solved at present is: how to evaluate the simulation results of the local area.

Contents of the invention

The purpose of the present invention is to provide a method for evaluating the credibility of land use change simulation based on spatial dislocation, which can not only report the quantitative evaluation index of each pixel, but also capture the spatial heterogeneity of accuracy, which is helpful for the detection of CA Trustworthiness of models and remote sensing classifiers to support land resource management and decision-making.

The purpose of the present invention can be achieved through the following technical solutions:

A method for assessing the credibility of land use change simulation based on spatial dislocation, comprising the following steps:

Step 1) collect the land use classification map and driving factor map for building the CA model;

Step 2) Obtain the land conversion probability map based on an intelligent optimization algorithm, and establish a CA model;

Step 3) Utilize the CA model to simulate the urban pattern and output and save the simulation results;

Step 4) Carry out a single numerical evaluation to the simulation results of the CA model by using the pixel-by-pixel comparison method;

Step 5) adopt the spatial dislocation-moving window analysis method to generate the simulation result state-spatial distribution diagram of static index;

Step 6) adopt the spatial dislocation-moving window analysis method to generate the spatial distribution diagram of the simulation result state-change index;

Step 7) outputting and saving the simulation result and the spatial distribution map as the reliability evaluation result.

Described step 1) comprises the following steps:

Step 1-1) using Landsat remote sensing images, using a supervised classification method to obtain land use classification maps;

Steps 1-2) Using vector datasets and raster images, extract a map of driving factors affecting urban sprawl.

Described step 2) comprises the following steps:

Step 2-1) Systematically sample the urban land use classification map and driving factor map, and provide training samples for constructing CA conversion rules, where the function of the conversion rules is expressed as:

P·g～UrbanTranRule(State _cu ,P _tr ,NE,Res,Sto)

In the formula, Pg is the global transition probability, State _cu is the current cell state, P _tr is the transition probability defined by the driving factor, NE is the neighborhood effect, Res is the global and local constraints, and Sto is the random factor;

Step 2-2) Based on the cross-entropy optimization algorithm, CEO searches for CA parameters, and establishes a CA model.

Described step 2-2) comprises the following steps:

Step 2-2-1) Construct an objective function linking the CA model to actual urban growth;

Step 2-2-2) Construct a group of random sequence samples according to the probability distribution;

Step 2-2-3) Find a better solution by updating the probability distribution parameters and the objective function, wherein, in order to find the most suitable probability P _tr (a), the automatically updated probability parameters P={P ₁ ,P ₂ ,..., P _k } is expressed as:

为指示函数，γ为预定义的小残差，a_ij为第i个驱动因子的第j个参数，k为CEO的种群规模；In the formula,

is an indicator function, γ is a predefined small residual, a _ij is the jth parameter of the i-th driving factor, and k is the population size of the CEO;

Step 2-2-4) When the CEO iteration ends, the optimal solution a, namely the CA parameters, is obtained, and the CA model is determined by the CA parameters.

Described step 3) comprises the following steps:

Step 3-1) Calibrate the constructed CA model: in the GIS modeling and simulation environment, select the urban distribution pattern of a certain year as the initial state of the correction period, run the constructed model A times, and obtain the urban distribution pattern Simulation results, and output the simulation results, where A is the year difference between the beginning and end of the model calibration period;

Step 3-2) Verify the constructed CA model: in the GIS modeling and simulation environment, select the urban distribution pattern of a certain year as the initial state of the verification period, run the constructed model B times, and obtain the urban distribution pattern Simulation results, and output the simulation results, where B is the difference between the beginning and end of the model verification period.

Described step 4) comprises the following steps:

Step 4-1) using the contingency matrix in the pixel-by-pixel comparison method to evaluate the state-stationary index of the simulation results;

Step 4-2) Using the real-simulated overlay in the pixel-by-pixel comparison method to generate four indicators: hits, false negatives, false positives and correct rejections, and calculate the status for evaluating the dynamic growth of the city based on the four indicators - Change indicators.

The state-stationary indicators include overall accuracy, user accuracy, and producer accuracy.

The state-change indicators include quality factor, precision rate, and recall rate.

Described step 5) comprises the following steps:

Step 5-1) Generate the basic evaluation indicator layer:

Layer a. Both the real model and the simulation results are of the city category Cor_Ur;

Layer b. The real model is the urban category, and the simulation result is the non-urban category Fal_Ur;

Layer c. The real mode is the non-urban category, and the simulation result is the urban category Fal_NUr;

Layer d. The real model and simulation results are both non-urban categories Cor_NUr;

Step 5-2) Using spatial dislocation-moving window analysis: using the focus statistics tool in ArcGIS to calculate the focus level index;

Step 5-3) Generate precision and error graphs: Generate state-stationary index focus level evaluation graphs by the following formula:

Among them, Map _mw () represents the layer, SPOA represents the spatialization result of overall precision, SPPA represents the spatialization result of producer precision, and SPUA represents the spatialization result of user precision.

Described step 6) comprises the following steps:

Step 6-1) Generate the basic evaluation index layer:

Layer a. The simulation results and the actual model are urban URHit;

Layer b. The simulation result is urban, and the actual mode is non-urban URFalse;

Layer c. The simulation result is non-urban, and the actual model is urban URMiss;

Layer d. The simulation results and the actual model are both non-urban CR;

Step 6-2) using spatial dislocation-moving window analysis: using the focus statistics tool in ArcGIS to calculate the focus level index;

Step 6-3) Generate precision and error graphs: generate evaluation graphs of state-change indicator focus levels by the following formula:

Among them, Map _mw () represents the layer, SPFOM represents the spatialization result of quality factor, SPPRE represents the spatialization result of accuracy, and SPRE represents the spatialization result of recall rate.

Compared with the prior art, the present invention has the following beneficial effects:

The method of the present invention can identify the quantitative accuracy of each pixel under the condition of considering the influence of the neighborhood, instead of only providing a single value to summarize the simulation quality and model performance, which better reflects the CA model in simulating cities in different regions The advantages of the growth effect are: 1) It can generate an evaluation map reporting the simulation accuracy/error of each pixel, showing the spatial distribution of simulation accuracy and error; 2) It can quantify the relationship between the driving factor and the evaluation index. This spatial evaluation method can be used not only to evaluate the simulation results of the CA model, but also to evaluate remote sensing image classifiers.

Description of drawings

Fig. 1 is a schematic flow chart of the method of the present invention;

Fig. 2 is the historical land use distribution diagram of the case region of the embodiment of the present invention;

Figure 3 shows the driving factors of land use change from 1995 to 2015;

Figure 4 shows the simulation results and evaluation charts in 2005 and 2015;

Fig. 5 is the state-stationary index spatial model diagram of the simulation results in 2005;

Figure 6 is the state-change index spatial model diagram of the simulation results in 2005;

Fig. 7 is the state-stationary index spatial model diagram of the simulation results in 2015;

Figure 8 is the state-change indicator spatial model diagram of the simulation results in 2015.

Detailed ways

The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is carried out on the premise of the technical solution of the present invention, and detailed implementation and specific operation process are given, but the protection scope of the present invention is not limited to the following embodiments.

Unlike global and local levels, moving window analysis is an efficient way to evaluate simulation results at the focal level. Some GIS scholars have applied this method to detect local differences or similarities between simulation results and actual patterns, and pointed out that moving window analysis can be used to assess the similarity of land use categories in composition and spatial structure. The method is characterized by fuzzy evaluation using a single numerical evaluation metric, such as Kappa. Compared with the precise comparison of pixel-by-pixel comparison, the land use change simulation credibility assessment based on spatial dislocation overcomes the shortcomings of previous assessment methods that may underestimate the simulation accuracy. Therefore, the present invention provides a land use change simulation credibility evaluation method based on spatial dislocation, as shown in Figure 1, including the following steps:

Step 1) Collect land use classification maps and driving factor maps that affect urban expansion covering the initial and end years of the study for building the CA model.

Step 1-1) Using Landsat remote sensing images, use the supervised classification method in the ENVI software to obtain the land use classification maps of the initial and end years that need to be input during the calibration and verification of the CA model.

Steps 1-2) Using vector datasets and raster images, extract a map of driving factors affecting urban sprawl.

Specifically, comprehensive topographic maps, population distribution maps, administrative division maps, road traffic maps, POI and other data, use the Euclidean distance method in the GIS environment to obtain key points such as city centers, district and county centers, main roads, and infrastructure. The distance of the elements is used to extract the driving factors affecting urban expansion.

Step 2) Obtain the land conversion probability map based on the intelligent optimization algorithm, and establish the CA model.

Step 2-1) Systematically sample urban land use classification maps and driving factor maps to provide training samples for building CA conversion rules.

Cellular Automata (CA) is a general term for a class of self-organizing models that adopt bottom-up methods for automatic evolution and consist of rules for model construction. The global transition probability P g determined by the current cell state (State _cu ), the transition probability based on the driving factor (P _tr ), the neighborhood effect (NE), the global and local constraints (Res), and the stochastic factor (Sto) And the transformation rule UrbanTranRule can be expressed as:

P·g～UrbanTranRule(State _cu ,P _tr ,NE,Res,Sto) (1)

where Pg is the global transition probability, State _cu is the current cell state, P _tr is the transition probability defined by the driving factor, NE is the neighborhood effect, Res is the global and local constraints, and Sto is the random factor.

Realize the construction of CA model in UrbanCA software, the global conversion probability P _all can be calculated by the following formula:

In the formula, TIP represents the scaling parameter for adjusting the probability attenuation effect, the range is 0.0-0.1, and the larger the value, the stronger the scaling effect; while LAP represents the scaling parameter for adjusting the neighborhood effect, the range is 0.5-1.0, and the smaller the value represents the scaling The effect is weaker.

The transition probability P _tr determined by the driving factor is the core part of the CA transition rule, which expresses the impact of the driving factor on land use change and urban development, and affects the cell state at the next moment through probability.

Step 2-2) Based on the cross-entropy optimization algorithm, CEO searches for CA parameters, and establishes a CA model.

Cross-entropy Optimization Algorithm (CEO) CEO is a heuristic algorithm that is widely used to search for the difference between the target value and its predicted value. In CA modeling, the objective function can be used to guide the model to achieve automatic parameterization. Specifically, the following steps are included:

Step 2-2-1) Construct an objective function linking the CA model to actual urban growth;

Step 2-2-2) Construct a group of random sequence samples according to the probability distribution;

Step 2-2-3) Find a better solution (ie, CA parameters) by updating the probability distribution parameters and the objective function. In order to find the most suitable probability P _tr (a), the automatically updated probability parameter P={P ₁ ,P ₂ ,…,P _k } is expressed as:

为指示函数，γ为预定义的小残差，a_ij为第i个驱动因子的第j个参数，k为CEO的种群规模；In the formula,

is an indicator function, γ is a predefined small residual, a _ij is the jth parameter of the i-th driving factor, and k is the population size of the CEO;

Step 2-2-4) When the CEO iteration ends, the optimal solution a, namely the CA parameters, is obtained, and the CA model is determined by the CA parameters.

Step 3) Use the CA model to simulate the urban pattern and output and save the simulation results.

Step 3-1) Calibrate the constructed CA model: in the GIS modeling and simulation environment, select the urban distribution pattern of a certain year as the initial state of the correction period, run the constructed model A times, and obtain the urban distribution pattern The simulation results are output, where A is the year difference between the beginning and end of the model calibration period.

Step 3-2) Verify the constructed CA model: in the GIS modeling and simulation environment, select the urban distribution pattern of a certain year as the initial state of the verification period, run the constructed model B times, and obtain the urban distribution pattern Simulation results, and output the simulation results, where B is the difference between the beginning and end of the model verification period.

Step 4) Use the pixel-by-pixel comparison method to perform a single numerical evaluation on the simulation results of the CA model.

Step 4-1) Use the contingency matrix in the pixel-by-pixel comparison method to evaluate the state-stationary index of the simulation results. In this embodiment, the state-static index includes overall accuracy (OA), producer accuracy (PA), and user accuracy (UA).

Step 4-2) Using the real-simulated overlay in the pixel-by-pixel comparison method to generate four indicators: hit (URHit), false negative (URMiss), false negative (URFalse) and correct rejection (CR), and based on the four indicators The indicators are calculated to assess the dynamic growth of the city - a state-change indicator. In this embodiment, the state-change indicators include figure of merit (FOM), precision rate (PRE), and recall rate (RE).

Step 4) is specifically: use the pixel-by-pixel comparison method to measure whether the state of each pixel in the simulation results matches the actual urban pattern. Selected as a metric for the quantitative evaluation of the CA model. OA is a single numerical probability representing the agreement of the simulated results with the actual urban pattern, which can be expressed as:

In the formula, p _nn is the number of pixels correctly simulated, SUM is the number of all pixels, and t _cn is the number of urban pixels in the actual result.

The spatial superposition of the actual map of the start year, the actual map of the end year and the simulated map of the end year can reveal the differences in structural patterns. The real-simulated overlay generates four metrics, including hits (URHit), false negatives (URMiss), false positives (URFlase), and correct rejections (CR). Among them, URHit indicates that a cell is in the urban state in both the actual model and the simulation result, URMiss indicates that a cell is in the urban state in the actual model but is a non-urban state in the simulation result, and URFlase indicates that a cell is in the non-urban state in the actual model. The urban state is urban in the simulation results, and CR indicates that a cell is non-urban in both the actual model and the simulation results. The indicators generated by the above spatial superposition can be used to calculate the state-change indicators used to evaluate the dynamic growth of cities, such as FOM:

In the formula, FOM represents the ability of the CA model to capture new urbanization pixels.

Step 5) Using the spatial dislocation-moving window analysis method to generate the spatial distribution diagram of the state-stationary index of the simulation results.

The moving window analysis generally takes a certain pixel as the center, sets a window with a fixed size and shape and performs statistical calculations in the window, and then returns the calculation results to the center pixel of the window. Repeating the calculation process for each cell produces a new raster surface layer. If the moving window analysis is used for model evaluation, the spatial distribution results of precision and error can be obtained.

Based on the traditional pixel-by-pixel comparison method, the visualization of the state-stationary index space mainly includes the following steps:

Step 5-1) Based on the actual land use pattern and the simulated land use pattern at _T2 time, four basic evaluation index TIF layers are obtained:

Layer a. Both the real model and the simulation results are of the city category Cor_Ur;

Layer b. The real model is the urban category, and the simulation result is the non-urban category Fal_Ur;

Layer c. The real mode is the non-urban category, and the simulation result is the urban category Fal_NUr;

Layer d. Both ground truth and simulated results are for the non-urban category Cor_Nur.

Step 5-2) Use spatial dislocation-moving window analysis: use the focus statistics tool in ArcGIS, input the index layers extracted in step 5-1), set the shape and size of the moving window and determine the numerical statistics method to calculate the focus Level indicators, the generated results are used to calculate the spatial visualization results of each state-stationary indicator.

Step 5-3) Generate precision and error graphs: Generate state-stationary index focus level evaluation graphs by the following formula:

Among them, Map _mw () represents the layer, SPOA represents the spatialization result of overall precision, SPPA represents the spatialization result of producer precision, and SPUA represents the spatialization result of user precision.

Step 6) Using the spatial dislocation-moving window analysis method to generate a spatial distribution diagram of the state-change index of the simulation result.

Based on the traditional pixel-by-pixel comparison method, the visualization of the state-change indicator space mainly includes the following steps:

Step 6-1) Based on the actual land use pattern at _T1 time, the actual land use pattern at _T2 time and the simulated land use pattern, generate a basic evaluation index layer for obtaining the spatial visualization results of state-change indicators:

Layer a. The simulation results and the actual model are urban URHit;

Layer b. The simulation result is urban, and the actual mode is non-urban URFalse;

Layer c. The simulation result is non-urban, and the actual model is urban URMiss;

Layer d. Simulation results and actual model are non-urban CR.

Step 6-2) Use spatial dislocation-moving window analysis: use the focus statistics tool in ArcGIS, input the index layers extracted in step 6-1), set the shape and size of the moving window and determine the numerical statistics method, the generated The results are used to calculate spatial visualization results for each state-change indicator.

Step 6-3) Generate precision and error graphs: generate evaluation graphs of state-change indicator focus levels by the following formula:

Among them, Map _mw () represents the layer, SPFOM represents the spatialization result of quality factor, SPPRE represents the spatialization result of accuracy, and SPRE represents the spatialization result of recall rate.

Step 7) outputting and saving the simulation result and the spatial distribution map as the reliability evaluation result.

This example uses Landsat remote sensing images and vector datasets to obtain the urban land pattern and driving factors affecting urban expansion in Ningbo in 1995, 2005 and 2015. In the UrbanCA software, the CA conversion rules are constructed based on the CEO method, the land conversion probability map of Ningbo is generated, and the CEO-CA model is constructed to simulate the urban expansion from 1995 to 2015. After using the pixel-by-pixel comparison method to obtain a single numerical evaluation result, the proposed spatial evaluation method is used to evaluate the simulation results in the calibration and verification stages, and the SPOA, SPPA, SPUA, SPFOM, SPPRE of the simulation results in 2005 and 2015 are obtained and SPRE.

In step 1), Ningbo City was taken as the research area to collect Landsat remote sensing images in 1995, 2005 and 2015, as well as administrative division maps, traffic road network maps, population maps, topographic maps and POI data, as the data constructed in the study The basic data of the CA model.

In the ENVI software, the Landsat remote sensing image of Ningbo is classified by the supervised classification method to generate the real urban spatial pattern map of Ningbo in 1995, 2005 and 2015, as shown in Figure 2.

Based on the administrative division layer, road traffic layer and infrastructure layer, use the Euclidean distance in the spatial analysis tool to calculate the distance between each cell to the city center, district and county center, main roads, and educational institutions in the GIS environment The distance to medical facilities, etc., and the grid map of terrain and population are obtained to form a map of driving factors that affect the urban spatial expansion of Ningbo, as shown in Table 1 and Figure 3.

Table 1. Driving factors affecting Ningbo's urban development

name category source explain distance to city center center National Basic Geographic Data The Influence of Ningbo Administrative Center on Urban Development distance to town center center National Basic Geographic Data The Influence of Ningbo Town Center on Urban Development distance to main road network OpenStreetMap road data Influence of Main Roads on Urban Development in Ningbo population density WorldPOP Population density per cell terrain altitude Terrain Remote Sensing Products Elevation of each cell distance to hospital POIs Baidu map Influence of Ningbo City Hospital on Urban Development distance to school POIs Baidu map Influence of Schools on Urban Development in Ningbo City

In step 2)-step 4), take the real urban model of Ningbo in 1995 as the initial state, use the CEO-CA model to simulate the urban pattern in 2005; take the real urban model of Ningbo in 2005 as the initial state, use The CEO-CA model simulates the urban landscape in 2015. A single numerical evaluation is performed on the simulation results, and Table 2 and Figure 4 show the simulation results and accuracy.

Table 2. Accuracy assessment of simulation results during calibration and validation periods

In step 5) and step 6), for the simulation results of the CA model, the focus-level accuracy was visualized and its statistics (range, mean, and SD) were analyzed using a moving window analysis method. Table 3 and Table 4 are the statistical information of the state-stationary and state-change space evaluation indicators during CA model calibration and verification, respectively.

Table 3. Statistics of evaluation indicators for state-stationary and state-change spaces during the correction period

Table 4. Statistics of the state-stationary and state-change space evaluation indicators during the verification period

Figure 5 and Figure 6 are the spatial evaluation diagrams of the simulation results during the calibration period of the CA model, in which Figure 5 shows the SPOA, SPPA and SPUA of the simulation results in 2005, and Figure 6 shows the SPFOM, SPPRE and SPRE of the simulation results in 2005 .

Figure 7 and Figure 8 are the spatial evaluation diagrams of the simulation results during the CA model verification period, in which Figure 7 shows the SPOA, SPPA and SPUA of the simulation results in 2015, and Figure 8 shows the SPFOM, SPPRE and SPRE of the simulation results in 2015 .

The preferred specific embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make many modifications and changes according to the concept of the present invention without creative efforts. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experiments on the basis of the prior art shall be within the scope of protection defined in the claims.

Claims (10)

1. The land utilization change simulation credibility evaluation method based on space dislocation is characterized by comprising the following steps of:

step 1) collecting a land utilization classification map and a driving factor map for constructing a CA model;

step 2) acquiring a land transition probability map based on an intelligent optimization algorithm, and establishing a CA model;

step 3) simulating the city pattern by using the CA model and outputting and storing a simulation result;

step 4) performing single numerical evaluation on the simulation result of the CA model by adopting a pixel-by-pixel comparison method;

step 5) generating a spatial distribution diagram of a simulation result state-static index by adopting a spatial dislocation-moving window analysis method;

step 6) generating a spatial distribution diagram of the simulation result state-change index by adopting a spatial dislocation-moving window analysis method;

step 7) outputting and storing the simulation result and the spatial distribution diagram as a credibility evaluation result.

2. The land use variation simulation reliability assessment method based on spatial dislocation according to claim 1, wherein said step 1) comprises the steps of:

step 1-1), land utilization classification images are obtained by using a land satellite remote sensing image and a supervision classification method;

step 1-2) extracting a driving factor graph affecting urban expansion by using the vector data set and the raster image.

3. The land use variation simulation reliability assessment method based on spatial dislocation according to claim 1, wherein said step 2) comprises the steps of:

step 2-1) performing systematic sampling on the urban land utilization classification map and the driving factor map, and providing training samples for constructing CA conversion rules, wherein the function of the conversion rules is expressed as follows:

P·g～UrbanTranRule(State _cu ,P _tr ,NE,Res,Sto)

where Pg is the global transition probability, state _cu Is the current cell state, P _tr Is the transition probability defined by the driving factor, NE is the neighborhood effect, res is the global and local constraint, sto is the random factor;

step 2-2) searching CA parameters based on a cross entropy optimization algorithm CEO, and establishing a CA model.

4. A land use change simulation reliability assessment method based on spatial dislocation according to claim 3, wherein said step 2-2) comprises the steps of:

step 2-2-1) constructing an objective function which relates the CA model to the actual city growth;

step 2-2-2) constructing a group of random sequence samples according to probability distribution;

step 2-2-3) finding a better solution by updating the probability distribution parameters and the objective function, wherein in order to find the most suitable probability P _tr (a) Automatically updated probability parameter p= { P ₁ ,P ₂ ,…,P _k Expressed as:

in the method, in the process of the invention,

for the indication function, γ is a predefined small residual, a _ij The j-th parameter which is the i-th driving factor, k is the population size of CEO;

step 2-2-4) when the CEO iteration is finished, obtaining an optimal solution a, namely CA parameters, and determining a CA model by the CA parameters.

5. The land use variation simulation reliability assessment method based on spatial dislocation according to claim 1, wherein said step 3) comprises the steps of:

step 3-1) correcting the constructed CA model: under the GIS modeling and simulation environment, selecting an urban distribution pattern of a certain year as an initial state of a correction period, operating the constructed model for A times to obtain a simulation result of the urban distribution pattern, and outputting the simulation result, wherein A is the difference between the initial and final years of the model correction period;

step 3-2) verifying the constructed CA model: under the GIS modeling and simulation environment, selecting an urban distribution pattern of a certain year as an initial state of a verification period, operating the constructed model for B times to obtain a simulation result of the urban distribution pattern, and outputting the simulation result, wherein B is the difference between the initial and the final years of the model verification period.

6. The land use variation simulation reliability assessment method based on spatial dislocation according to claim 1, wherein said step 4) comprises the steps of:

step 4-1), performing state-static index evaluation on the simulation result by adopting a column-linked matrix in a pixel-by-pixel comparison method;

step 4-2) generating four indexes by adopting real-analog superposition in a pixel-by-pixel comparison method: hit, miss, false positive, and correct rejection, and calculate a state-change indicator for assessing dynamic growth of the city based on the four indicators.

7. The method for evaluating reliability of land use variation simulation based on spatial misalignment of claim 6, wherein the status-stationary index comprises overall accuracy, user accuracy, and producer accuracy.

8. The method for evaluating reliability of land use variation simulation based on spatial dislocation according to claim 6, wherein said state-variation index comprises quality factor, accuracy, recall.

9. The method for evaluating the reliability of land use change simulation based on spatial dislocation according to claim 7, wherein said step 5) comprises the steps of:

step 5-1) generating a basic evaluation index layer:

layer a. Both the real mode and the simulation result are city class Cor_Ur;

layer b. True mode is city class, simulation result is non-city class Fal _ur;

layer c. True mode is non-city class, simulation result is city class Fal _ NUr;

layer d. Real mode and simulation result are non-city class Cor_ NUr;

step 5-2) using spatial dislocation-moving window analysis: calculating a focus level index by adopting a focus statistics tool in ArcGIS;

step 5-3) generating an accuracy and error map: an evaluation chart of the state-stationary index focus level is generated by the following formula:

wherein, map _mw () The layer is shown, SPOA shows the result of the spatialization of the overall accuracy, SPPA shows the result of the spatialization of the producer accuracy, SPUA shows the result of the spatialization of the user accuracy.

10. The land use variation simulation reliability assessment method based on spatial dislocation according to claim 8, wherein said step 6) comprises the steps of:

step 6-1) generating a basic evaluation index layer:

layer a. Both simulation result and actual mode are city URHit;

layer b. Simulation results are urban and actual mode is non-urban URFalse;

layer c. Simulation result is non-city and actual mode is city URMiss;

layer d. Simulation result and actual mode are non-urban CR;

step 6-2) using spatial dislocation-moving window analysis: calculating a focus level index by adopting a focus statistics tool in ArcGIS;

step 6-3) generating an accuracy and error map: an evaluation chart of the state-change index focus level is generated by the following formula:

wherein, map _mw () The layer is shown, SPFOM shows the spatialization result of the quality factor, SPPRE shows the spatialization result of the accuracy, and SPRE shows the spatialization result of the recall.

Images