基于Landsat8 OLI遥感数据的辽河口湿地分类与动态监测研究

Classification and Dynamic Monitoring of Wetland at Liaohe Estuary Based on Landsat8 OLI

湿地复杂地类的时空分布特征提取和变化监测,对湿地资源的合理保护与社会可持续发展具有重要意义。以辽宁辽河口湿地作为研究区,结合生态环境野外数据采集,利用Landsat8 OLI遥感影像,开展2013,2015,2017,2020共4年的辽河口湿地信息提取、分类、变化研究。利用不同湿地类型的空间分布,采用监督分类方法,从Landsat8 OLI遥感影像中精准提取湿地信息,基于3种监督分类方法,针对不同地物的分类结果,选出每种地物分类精度最高的结果图像融合,并进行变化分析。结果表明:最高的总体分类精度方法为支持向量机,分类精度81%,Kappa系数为0.76。矢量融合后的分类图,平均分类精度为87.4%,Kappa系数为0.82,提高湿地分类精度效果明显。研究区2013~2020年天然湿地面积变化总体呈减少趋势,总面积减少12.43%,约113.46km^(2);人工湿地面积总体增加24.96%,约87.10km^(2);非湿地总体面积增加89.08%,约24.85km2;天然湿地地物中,碱蓬草湿地草地面积基本持平,总面积维持在26km^(2);芦苇湿地、滩涂面积总体呈减少趋势,分别减少126.26km^(2)和36.41km^(2)。研究结果为辽河口湿地分类以及动态监测提供方法和数据支持,对辽河口湿地生态保护具有重要的意义。

The extraction and monitoring of temporal and spatial distribution characteristics of wetland complex landscape is of great significance to the rational protection of wetland resources and the sustainable development of society.In this study,Liaohe estuary Wetland in Liaoning province was taken as the research area.Combined with the field data collection of ecological environments,the Landsat8 OLI remote sensing images were used to carry out the extraction,classification and change of liaohe estuary wetland information in 2013,2015,2017 and 2020.Based on the spatial distribution of different wetland types,the supervised classification method was adopted to accurately extract the wetland information from the Landsat8 OLI’s remote sensing images.Based on the three supervised classification methods,the result image fusion with the highest classification accuracy of each ground object was selected according to the classification results of different ground objects,and the change was analyzed.The results show that support vector machine (SVM) has the highest classification accuracy of 81%and Kappa coefficient of 0.76.The average classification accuracy of the vector fusion map is 87.4%,and the Kappa coefficient is 0.82,it obviously improves the classification accuracy of wetland.From2013 to 2020,the total area of natural wetland decreased by 12.43%,about 113.46km^(2).The total area of constructed wetland increased by 24.96%,about 87.10km^(2);The total area of non-wetland increased by 89.08%,about 24.85km^(2);among the natural wetland features,the area of suaeda wetland grassland was basically stable,with a total area of 26km^(2).The area of reed wetland and tidal flat decreased by 126.26km^(2) and 36.41km^(2) respectively.This study provides methods and data support for wetland classification and dynamic monitoring at Liaohe Estuary,which is of great significance for the wetland ecological protection.