A Geographic Information-Assisted Temporal Mixture Analysis for Addressing the Issue of Endmember Class and Endmember Spectra Variability.

Wenliang Li,Changshan Wu

doi:10.3390/s17030624

Abstract

Spectral mixture analysis (SMA) is a common approach for parameterizing biophysical fractions of urban environment and widely applied in many fields. For successful SMA, the selection of endmember class and corresponding spectra has been assumed as the most important step. Thanks to the spatial heterogeneity of natural and urban landscapes, the variability of endmember class and corresponding spectra has been widely considered as the profound error source in SMA. To address the challenging problems, we proposed a geographic information-assisted temporal mixture analysis (GATMA). Specifically, a logistic regression analysis was applied to analyze the relationship between land use/land covers and surrounding socio-economic factors, and a classification tree method was used to identify the present status of endmember classes throughout the whole study area. Furthermore, an ordinary kriging analysis was employed to generate a spatially varying endmember spectra at all pixels in the remote sensing image. As a consequence, a fully constrained temporal mixture analysis was conducted for examining the fractional land use land covers. Results show that the proposed GATMA achieved a promising accuracy with an RMSE of 6.81%, SE of 1.29% and MAE of 2.6%. In addition, comparative analysis result illustrates that a significant accuracy improvement has been found in the whole study area and both developed and less developed areas, and this demonstrates that the variability of endmember class and endmember spectra is essential for unmixing analysis.

Highlights

Spectral mixture analysis (SMA) is an inverse process that decomposes the spectra of a mixed pixel into areal fractions based on the pure spectra of its component land covers [1]
In order to assess the fitness of the regression model, the relative operating characteristic (ROC) was calculated and the ROC values for all models are over 0.65, indicating that the all chosen driving factors can be used to explain the spatial distribution of all land use types
With the knowledge of the spatial distribution probabilities of all endmember classes produced from the logistic regression model, the classification tree approach was utilized to automatically identify the present and absent of endmember classes for each pixel (Table 2)

Summary

Introduction

Spectral mixture analysis (SMA) is an inverse process that decomposes the spectra of a mixed pixel into areal fractions based on the pure spectra of its component land covers ( termed as endmembers) [1]. Some studies have been conducted for selecting the type and number of endmember classes, and several endmember class models have been proposed, such as vegetation-soil-shade (V-S-S) for estimating vegetation fraction in desert region [4], vegetation-impervious surface-soil (V-I-S) [5], vegetation-low albedo-high albedo (V-L-H) [6], and vegetation-low albedo-high albedo-soil (V-L-H-S) [7] for mapping urban biophysical composition. It is straightforward, only fixed endmember sets were. Zhang et al [9]

Results

Discussion

Conclusion