Study On Method Of Data Quality Check For Remote Sensing Modeling In Inland Case ? Water

The use of remote sensing as a tool for the monitoring of inland Case ? Water include the constructing of remote sensing models,the estimation of water parameters,and the assessment of water quality.However,the storage of remote sensing data including the in situ data and satellites images is style different and is short of the uniform management and it also lacks uniform check rules of the control of the data quality.This problem limits the construction of best performance remote sensing models,the application of these models,and the share of remote sensing data.Therefore,only with high quality control,and then a scientific and normalized process of inputting to the database,the collected remote sensing data can express its value in the environment protection of inland waters.In this study we first collect mass in situ apparent and inherent optical data from eight inland waters in China,in situ data of Lake Erie,Lake Michigan in SeaBASS dataset,and three reservoirs of Indiana and mass long time series satellite images above these inland waters.Then we analyst the specialty of different type of remote sensing data and the modeling requirement of different water parameters.Based on this,we further develop the unified format of in situ data and satellites images before they inputting into the database and propose the rule of how to inputting these different data into the database.The primary goal of this study is to better serve the water environment monitoring.The main research contents and results are as follows:(1)Based on the analysis of methods for remote sensing modeling in inland case? water,we divided the methods into experience/semi-analytical model,bio-optical model and machine learning model.For each type of the model,we analysis and summarize the modeling data and verify data.And mainly focus on Remote Sensing Reflectance(Rrs),pigment particles absorption coefficient(ad),phytoplankton absorption coefficient(aph),CDOM absorption coefficient(ag),etc.And according to its optical properties,it can be divided into suspended sediment dominant water and phytoplankton dominant water.At the same time,Analyzing the data source and formation of different remote sensing data.(2)For the in suit data such as apparent optical property(AOP),inherent optical properties(IOP)and concentration of water color parameter,in this article we classify the data quality checking methods into three categories including numerical check,spectral shape check and the union check of concentration of water color parameter and spectral shape.The numerical check including domain check and influence coefficient method check.The spectral shape check including spectral angle matching(SAM)and specified band method.For the warehousing of remote sensing image data check,in this article we mainly study on cloud detection and the atmospheric correction results check method.(3)In this paper,combining the relevant standard of SeaWiFS Bio-optical Archive and Storage System(SeaBASS)and Watershed Allied Telemetry Experimental Research(WATER),we put forward a set of data reorganize standard for remote sensing modeling in inland case ? water.Studying on the meta data,in suit data and remote sensing image preprocessing to the database table structure design.