 This study investigates the effects of minimum noise fraction MNF, transformation on crop classification using multi-temporal radar SAT, two polarimetric SAR data in southwestern Ontario, finding that the elements of coherency matrix performed best, and an overall accuracy of 90% can be achieved with two images acquired in September and October. The MNF transformation improved classification accuracy, making it competitive with the integration of optical and polarimetric SAR data. This article was authored by Chen Huoliao, Jean-Faye Wang, Xiaodong Huang and others.