A spectral index for winter wheat mapping using multi-temporal Landsat NDVI data of key growth stages

Abstract

Winter wheat is one of the most important staple crops in the world. Accurate and timely information on the spatial distribution and temporal change of winter wheat is critical for food security and environmental sustainability. Multi-temporal images and time series data of medium resolution are widely used in winter wheat mapping. However, relatively long revisit times and image noise often result in a deficiency in full time series data. In this paper, a new spectral index, called the winter wheat index (WWI), using multi-temporal Landsat normalized difference vegetation index (NDVI) data of four key growth stages of winter wheat, was proposed to highlight and map winter wheat. Two distinctive NDVI contrasts, each consisting of an NDVI peak and trough, were identified and used in the WWI. The proposed index was evaluated through qualitative and quantitative analyses as well as winter wheat mapping, and compared with three state-of-the-art methods. To map winter wheat using WWI, a Monte Carlo cross validation procedure was adopted to determine the optimal thresholds of the WWI. Visual comparison showed that winter wheat was highlighted by higher WWI values, whereas other land cover types had lower WWI values. The experimental results from quantitative analysis indicated that WWI achieved better separability between winter wheat and other land cover types than the other comparative indices. The proposed WWI also produced more accurate winter wheat mapping results, compared with the state-of-art methods. Therefore, the proposed WWI provides a useful variable for winter wheat mapping, which reduced the dependence on full time series data and the use of noise images, and can be applied in other study areas.