Study on Scattering Characteristics of Dryland Crops using Multi-temporal Polarimetric RADARSAT-2 Imagery

Abstract

Dryland crops have a long planting history in China. They are planted in a wide range and account for a high proportion of the total grain output. Quick and accurate acquisition of Dryland Crop Planting area can provide decision support for agricultural production managers, provide basis for food policy formulation, and provide guarantee for national food security. Different from the initial stage of rice growth, the underlying surface water layer and rice plant can form obvious dihedral angle, which can produce strong backscatter to microwave. Intercropping and interplanting of dryland crops are common, and the planting structure is complex, so it is difficult to identify them. At present, there is a lack of research on scattering characteristics of Dryland crops, and the universality of recognition methods is also poor, which leads to the low accuracy of dry land crop recognition based on SAR data. The purpose of studying the scattering characteristics of dryland crops and their changes with time is to provide basis for the identification of dryland crops and improve the classification accuracy. This paper chooses Jizhou City, Hebei Province as the research area, and takes corn and cotton as the research objects. The full polarization RADARSAT -2 data of three phases in 2018 (July 17, August 7 and September 24) were used. The changes of basic scattering characteristics (average scattering angle, entropy, volume scattering, dihedral angle scattering, surface scattering) of crops with different target decomposition methods (Cloude-Pottier, Freeman, Yamaguchi) were studied and compared, and the proportion of basic scattering power and its changing trend at different growth stages were analyzed. The results showed that the counter-entropy of the two crops changed little in the whole growth period, and mainly consisted of surface scattering and volume scattering. For corn, with the growth of crops, the entropy and average scattering angle increased first and then decreased, the proportion of surface scattering power decreased from 67% to 48% and then increased to 55%, and the proportion of volume scattering power increased from 33% to 52% and then decreased to 45%. On August 7, the volume scattering power is greater than the surface scattering power. For cotton, with the increase of crop growth entropy and average scattering angle, the proportion of surface scattering power decreases from 66% to 54%, and the volume scattering power increases from 33% to 46%. The surface scattering power is larger than volume scattering in the whole growth period. This study will help to determine the scattering mechanism of corn and cotton, and provide reference for the study of scattering characteristics of other dryland crops.