|
Crop lodging is a phenomenon in which plant stems change from a naturally upright state to a permanent dislocation caused by external factors. It is a common disaster in agricultural production and will directly affect the yield and quality of crops. It needs to be discovered and intervened in time. Among the current remote sensing methods, optical remote sensing is limited by the uncertainty of spectral changes, and it is difficult to distinguish lodging from other complex influencing factors such as drought, pests and diseases, and it cannot be observed at night, which is easily affected by clouds and rain. Synthetic aperture radar (SAR) technology has the capability of all-day and all-weather monitoring. However, the methods using SAR in existing research are mostly simple exponential structures and thresholds for extraction, which have poor applicability. Since polarimetric SAR can obtain multi-channel scattering characteristics of crops, this paper proposed a method of crop lodging region detection using polarimetric SAR images. First, the H/A/Alpha polarization decomposition is performed to obtain a series of polarization features, and then the SVM classifier is established by using the polarization features and the backscattering intensity of different polarization modes. Finally, the lodging detection of crops in the area is completed. The experimental results show that the crop lodging area extracted by Polarimetric SAR and Support Vector Machine (SVM) is consistent with the field survey results, and it is a reliable means for monitoring the growth status of large-scale crops.
|
