Dissertation > Agricultural Sciences > Agriculture as the foundation of science > Soil > Soil physics > Soil moisture

A Preliminary Study on Bare Soil Extraction and Its Moisture Retrieval from High Resolution SAR Data

Author JinXi
Tutor ShiZhou
School Zhejiang University
Course Agricultural Remote Sensing and IT
Keywords ALOS AVNIR-2 PRISM PALSAR Textual analysis PCA Land use classification Soil moisture retrieval An experimental model
CLC S152.7
Type Master's thesis
Year 2011
Downloads 34
Quotes 0
Download Dissertation

Being an important parameter in the interaction of land-atmosphere, soil moisture is critical in meteorology, soil science, hydrology and ecology, as well as a significant part of regulating the hydrological circulation and the climate change. To spatially investigate soil moisture in a really wide range, can not only supply basic information for aridity or flood predict, but also provide elementary input data for hydrology, climate and ecology models. While traditional investigations have their limits in timely delay and spatially unrepresentativeness, and Visible Near-Infrared Remote Sensing, Thermal Infrared Remote Sensing and Passive Microwave Remote Sensing have much too coarse resolution for the application on the study area, Active Microwave imagery ALOS/PALSAR was selected as data source. After summarizing the research history and achievements of SAR applications on soil moisture and found no proper model for the selected data, a new empirical model for soil moisture inversion was planned to retrieve surface soil moisture of bare soil of our study area.The first move of our study was to make analysis of all ALOS sensors’potential on land use classification, especially on extracting seasonal bare soils, which is base work for the next step. Thus, three kinds of data, namely, only AVNIR-2, fusion image of AVNIR-2 and PRISM, as well as textural filtered and principal analyzed PALSAR HH/HV dual-polarized data, was processed to classify land use types. It is showed that optical imagery can separate 11 types of secondary land use types well; in the meantime, PALSAR image can separate no more than 5 main types. Therefore, PALSAR can be chose as substitute data source while no optical data can be obtained in rainy, misty, cloudy and snowy days. Based on the last step, bare soil on the PALSAR data on May 21,2010 was extracted for the next move.Thereafter, an empirical model was built for bare soil surface moisture retrieval based on the analysis of backscattering of PALSAR data on Nov.21,2010 and in-situ soil moisture measurements of the selected northern field. Afterwards, the first accuracy assessment was analyzed on the southern field, in which we did experiments on the same day; and a promising correlation between inversed results and the measurements was found. To more explore its applicability, the model was applied to make inversion on the PALSAR data acquired on May 21,2010, which was later compared with near in-situ ground truth data. However, due to a lot of uncertainties, such as the ground truth was not true, and it is possible that the field we did experiments on May 17, 2010 was covered by plastic film on the date of May 21,2010, we do not have a concise conclusion on the model’s applicability. And yet another in-situ experiment is needed to confirm the accuracy and applicability of the empirical model.Finally, based on some situations we met through the experiments and data processing, some advice was listed in the last chapter, which aims at getting better experimental results and easier data processing.

Related Dissertations
More Dissertations