Analysis Of Wuhan Greenbelt Landscape Dynamics And Correlation Models Based On Remote Sensing

Exploring the dynamic association rules of urban green space landscape and predicting the evolution of urban green space landscape has always been an important goal of urban ecological planning.The urban green space landscape pattern is of great significance to the sustainable development of urban ecosystems.In this study,the green space landscape data set was studied to study the dynamic change process of green space landscape.On this basis,the correlation and importance of impact factors and green space landscape were analyzed,and the GWR model was improved.In the analysis of urban green space landscape association mode.The main contents of this study are as follows.In order to study the dynamic evolution of Wuhan green space landscape,first of all,based on the GEE platform,this study used the index method to calculate the NDVI of Wuhan Landsat remote sensing images from 2000 to 201 8,and mapped Wuhan green space.Then,based on the remote sensing images,the global landscape area percentage(PLAND)and aggregation degree(AI)of the green space landscape in Wuhan are calculated year by year.Finally,the method of Mann-Kendall combined with Sen's slope was used to analyze the time series of PLAND and AI.The experimental results show that in PLAND's MK curve,UF shows a downward trend,and ? is-0.190,indicating that the overall area of urban green space in Wuhan has decreased and changed significantly;in AI's MK curve,UF has a downward trend,and ? is-0.216,indicating The degree of polymerization of Wuhan urban green space is decreasing,which proves the conclusion of Wuhan urban green space fragmentation.Aiming at the problem of screening the influence factors of Wuhan greenbelt landscape correlation mode,this study uses the double coefficient correlation method and random forest classification model to select it.First,select 8 influencing factors from climatic factors,topographical factors and human factors,and use PLAND and AI as indicators of the urban green space landscape pattern in Wuhan;then,use Pearson and Spearman double coefficients for correlation analysis,from 8 influencing factors The relevant factors that have significant correlation with PLAND and AI are screened out.Finally,a random forest classification model is used to calculate the importance of the correlation factor on PLAND and AI.The experimental results show that DEM,slope,NDBI,precipitation,railway,highway and water network are all related to the urban green space landscape pattern in Wuhan,and the impact is significant.The importance of the impact on PLAND is 24%,21%,18%,10%,10%,9%and 8%,the importance of the impact on AI is 23%,20%,17%,13%,13%,8%and 6%.In view of the uncertainty of the influence factor value range in the correlation model,this study improves the GWR model to the PB-GWR model,and uses the greenland landscape system analysis method.First,compare the traditional OLS and GWR correlation models,and select the GWR model with higher R2 and lower AICc;then,integrate the P-Box model to construct the PB-GWR model and apply it to the green space landscape in Wuhan;finally,adopt Greenland landscape systematic analysis method,used to explain the related mode of Wuhan greenland landscape.The experimental results show that the mountains are positively correlated to PLAND and AI,DEM and slope form a weak synergistic effect in the green landscape system;in non-mountain areas,the three influencing factors of DEM,slope and water form in the green landscape system Strong synergy effect,showing a very high positive correlation;the strong trade-off effect formed by the three impact factors DEM,slope and NDBI is shown as a negative correlation;the trade-off effect of roads and water networks in the downtown area makes GWR The corresponding coefficient of the model approaches zero.