You are here:
Comparison of Sub-pixel Classification Approaches for Crop-specific Mapping
Citation:
Shao, Y. AND R. S. LUNETTA. Comparison of Sub-pixel Classification Approaches for Crop-specific Mapping. Presented at 2009 Geoinformatics, Fairfax, VA, August 12 - 14, 2009.
Impact/Purpose:
Presentation
Description:
The Moderate Resolution Imaging Spectroradiometer (MODIS) data has been increasingly used for crop mapping and other agricultural applications. Phenology-based classification approaches using the NDVI (Normalized Difference Vegetation Index) 16-day composite (250 m) data product is among the most promising for automated processing. Most MODIS-NDVI crop mapping applications to date have focused on per-pixel classification methods; while the sub-pixel crop patterns and proportions have not been thoroughly exploited. The objective of this paper was to implement and compare three sub-pixel classification approaches for estimating sub-pixel crop proportions using MODIS-NDVI data. The sub-pixel unmixing approaches included: (a) Multilayer Perceptron (MLP) regression algorithm, (b) Multilayer Perceptron (MLP) classification algorithm, and (c) Regression tree. The sub-pixel proportions were estimated for three major crop types including corn, soybean, and wheat; throughout the entire 480,000 km2 Laurentian Great Lakes Basin. Accuracy assessments were conducted using the cropland data layer (CDL) developed by the National Agricultural Statistics Service (NASS) to provide reference data corresponding to calendar year 2007. The performances of the three sub-pixel classification approaches were compared based on a number of factors including the size of the training set, model complexity, and the generalization potential over large geographic regions. We found that Multilayer Perceptron (MLP) regression algorithm provided the best overall performance and the Multilayer Perceptron (MLP) classification algorithm the poorest. The implementations of MLP classification approaches were time-consuming compare to the Regression tree approach. Also, the Regression tree had greater interpretability; especially when smaller regression trees were employed.