Time-weighted dynamic time warping analysis for mapping interannual cropping practices changes in large-scale agro-industrial farms in Brazilian Cerrado

• DOI: 10.1016/j.srs.2021.100021
• Publication Date: 2021-03-07T20:38:00Z
• Authors: Michel E.D. Chaves; Marcelo de C. Alves; Thelma Sáfadi; Marcelo S. de Oli...; Michelle C.A. Picoli; Rolf E.O. Simoes; Guilherme A.V. Ma...
• Keywords: Crop Rotation; Thematic map; Moderate-resolution imaging spectroradiometer; Land Cover; Dynamic Time Warping

Methods for crop phenology detection using time series analysis have provided accurate information large agricultural areas in shorter processing times, which can be useful agronomic management and supply chain monitoring. Given the dynamics Brazilian Cerrado, with alternating type plantings, successions, rotations, as well climate practices variation between harvest periods, these methods detecting subtle land use cover changes at farm field scales, improving thematic classifications near real-time In this study, Time-Weighted Dynamic Time Warping method was applied to recognize patterns Moderate Resolution Imaging Spectroradiometer (MODIS) classification, identifying successions rotations level a large-scale agro-industrial agglomerate of farms located Cerrado. We detected analyzed temporal cropping training samples classify MODIS images, robust ground truth data set validation. The distinguished Cotton-fallow, Soybean-cotton, Soybean-maize, Soybean-millet an overall accuracy above 85% all evaluated periods. Seasonal variations fields, caused by interannual succession rotation, were detected. demonstrated benefit creating spatial vector supporting decision-making several contexts,