Optimizing land cover change detection using combined pixel-based and object-based image classification in a mountainous area in Mexico

Inventories of past and present land cover changes form the basis for future conservation strategies and landscape management. In this study Landsat images of a mountainous area in Mexico are used in an object-based and pixel-based image classification. The land cover categories with the highest individual classification accuracies determined with these two methods are extracted and merged into combined land cover classifications. Seven land cover categories were extracted and combined into single combined best classification layers. Comparison of the overall classification accuracies for 1999 and 2006 of the pixel-based (0.74 and 0.81), object-based (0.77 and 0.71) and the combined (0.88 and 0.87) classifications shows that the combination method produces better results. These combined classifications then form the input for change detection between the two years, by applying post-classification object-based change analysis using image differencing. It is concluded that post-classification object-based change detection analysis leads to an improved land cover change detection result with an overall accuracy of 0.77. This approach has the potential to be applied in similar mountain areas using medium resolution imagery for land cover change analysis.