Wetland Landscape Pattern Change In The Yellow River Delta Based On Spatio-temporal Fusion Of Remote Sensing

Wetland systems are an important part of the global natural ecological environment.The Yellow River Delta is the youngest coastal wetland in China and an important coastal wetland conservation area.The Yellow River Delta is a habitat for many rare species.In recent years,the wetland has been increasingly damaged due to human activities in the study area,and the surface landscape structure has undergone complex and drastic changes,resulting in threats to the ecological environment,reduced land use efficiency,and increased fragility of the landscape pattern.Therefore,it is of great significance to study the spatio-temporal distribution and dynamic changes of the landscape pattern in the Yellow River Delta wetland for the protection of the ecological environment and the regional economic development.In this study,Dong ying City was selected as the study area,and Landsat(MSS/TM/ETM+/OLI)remote sensing images with a period of two years were used as the data source from 2000 to 2020.The 2000 image was polluted by clouds,causing poor data quality.The Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model(ESTARFM)algorithm was used to supplement the missing data.Combined with spectral indices,spectral characteristics,texture characteristics,etc.,the landscape type information in the study area was extracted based on the support vector machine algorithm.According to the single dynamism of the landscape type,the landscape index method of centroid migration,patch type level and landscape level was used to analyze the spatio-temporal changes of the landscape type in the study area from 2000 to 2020.The results show that:(1)The fusion data obtained using the ESTARFM algorithm is of practical value and provides base data for the classification of the Yellow River Delta wetlands.The ESTARFM algorithm has practical value.The spectral and spectral index comparison between the fusion data and the original data showed that the determination coefficient R2 was greater than 0.66;the spectral,texture,spectral index,etc.features of the fusion data were basically the same as those of the original image locally.According to the results of other researchers and the actual situation in the study area,the results showed that the fusion data had practical value.(2)Support Vector Machine classification algorithms have better performance in extracting landscape type information in the Yellow River Delta region.For the small sample of this research,the proposed support vector machine method has high classification accuracy.The overall accuracy of landscape type classification in 11 periods between 2000 and 2020 is higher than 84%,and the Kappa coefficient is higher than 0.82. The overall accuracy of classification in 2012 is the highest up to 90.78%.(3)Dynamic degree model,landscape index,center of gravity migration and other methods can be used to make reasonable analysis of landscape pattern change in the Yellow River Delta in the past 20 years.Finally,based on the analysis of the experiment results,reasonable suggestions on the protection of the Yellow River Delta wetlands are proposed from different angles,which is of great significance for Regional protection.