ABSTRACT
This chapter analyses the theory of different mixture models along with a quantitative and intercomparative assessment of the popular unmixing algorithms is performed based on the evaluation of the derived LC fractional estimates in a comprehensive and rigorous manner on the NASA Earth Exchange (NEX) computing platform. It presents the survey, definition, and analysis of the state-of-the-art unmixing algorithms for subpixel classification. The chapter also presents the experimental results and their evaluation from the computer-simulated data followed by assessment of the algorithms on Landsat data of an agricultural scenario and of an urban setup. It discusses the evaluation results along with the merits and demerits of the algorithms with concluding remarks. In the unmixing jargon, nonnegativity and sum-to-one constraints are termed abundance nonnegativity constraint (ANC) and abundance sum-to-one constraint (ASC). The chapter explains the results from the unmixing algorithms on computer-simulated data and Landsat data of an agricultural landscape near Fresno and an urban agglomeration in San Francisco.
