A Review of Crop Classification Using Satellite-Based Polarimetric SAR Imagery

Crop area information can provide decision support for agricultural production management, and it is an important basis for formulating grain policy and economic plan. Remote sensing has an irreplaceable role in large-scale agricultural monitoring, in which optical remote sensing is greatly influenced by cloud and rain weather. It is often unable to obtain clear and clear optical remote sensing images in the critical period of crop growth, which seriously images the accuracy and timeliness of the classification of crop remote sensing. Synthetic aperture radar (SAR) capable of monitoring ground objects throughout the day and all weather, Polarimetric SAR data also include scattering matrix, geometric details and permittivity information, it is sensitive to the geometry and height of vegetation, which can make up for the lack of optical remote sensing and has unique advantages in crop identification and monitoring. So full polarization SAR has wide application demand and great potential in crop remote sensing monitoring. The effective use of full polarimetric SAR data to classify crops is of great academic and practical value for promoting the greater role of radar technology in national agricultural remote sensing monitoring and agricultural supply side structural reform. This paper summarizes the different data types and polarization decomposition methods of polarization SAR used in crop identification. According to the current research, there are the following deficiencies: first, most of the research subjects are rice, and dry land crops are less studied. Secondly, the identification of dry land crops is not accurate enough, and the average recognition accuracy is less than 85%. Finally, there are few researches on the scattering mechanism, which leads to poor rationality and poor universality. First, determine the scattering mechanism of dry land crops, and use scattering mechanism to improve the accuracy and universality of remote sensing recognition of dryland crops. Second, how to make use of the special imaging method of SAR to optimize the design is applicable to polarimetric SAR classification algorithm. Third, how to better integrate with optical remote sensing and other multi-source data, these three points will become the key problem to be solved in polarimetric SAR crop classification in the future.