Census Data Mining for Land Use Classification

Abstract

This chapter presents spatial data mining techniques applied to support land use mapping. The area of study is in Sao Paulo municipality. The methodology is presented in three items: extraction, transformation and first analysis; knowledge discovering and supporting rules evaluation; image classification support. The combined inferences resulted in a good improvement in the digital image classification with the contribution of Census data. 17.1 Content Structure The intent of this study is to describe the use of spatial data mining of Brazilian Census data as a support to land use mapping using digital image classification . To describe the procedures and evaluate the results obtained the following items will be described: • Land use: what is it and how can it be mapped; • Remote sensing images as a tool to land use mapping; • Digital image processing as a technique to classify land use; • Characteristic of Census data to understand land use distribution; • Data warehouse and spatial data mining of Census data as a support to land use mapping; • Integration of data warehouse and spatial data mining and digital image processing to classify land use; • Results and Discussion; • Findings and perspectives. E. Roma Neto, D. S. Hamburger Av. Eng. Euseio Stevaux, 823 04696-000, Sao Paulo, SP, Brazil, e-mail: elias.rneto@sp. senac.br,diana.hamburger@gmail.com