.png){kind=icon}
.png){kind=icon}
Return
A PolSAR rotation model for crop classification and soil moisture retrieval in complex agricultural environments
.png){kind=icon}
PRE
AI
DOI๏ผ10.1080/01431161.2026.2664867.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En ไธญๆ
In complex agricultural scenes, the structure, orientation and dielectric properties of scatterers exhibit significant heterogeneity, necessitating the development of highly adaptive scattering models to avoid parameter estimation biases caused by mismatches between models and actual conditions. The adaptability of such models is largely driven by polarization rotation mechanisms. Current research primarily focuses on polarization orientation angle (POA) rotation related to the ๐23 (๐23_๐
๐). However, the effects of ๐12-related rotation (๐12_๐
๐) and ๐13-related rotation (๐13_๐
๐) have yet to be systematically investigated. This limitation limits model adaptability and compromises the accuracy of both crop classification and soil moisture (SM) retrieval. To address this gap, this study develops a rotation scattering model (ROM) that integrates ๐23_๐
๐, ๐12_๐
๐ and ๐13_๐
๐. The proposed ROM aims to leverage its inherent adaptability to improve the accuracy of crop parameter retrieval. UAVSAR data covering the Winnipeg region of Manitoba, Canada, were utilized to evaluate the ROMโs effectiveness in varied SM and vegetation coverage conditions. Experimental results demonstrate that the model parameters of ROM carry clear physical meanings, effectively characterizing both structure, randomness and dielectric constant of targets. These physically interpretable parameters contribute to improved crop classification accuracy, yielding an overall increase of 3.59% compared to conventional general scattering models. Furthermore, the modelโs adaptive power transformation capability across polarization channels allows it to better accommodate variations in surface roughness. In SM retrieval, the ROM can achieve a root mean square error (RMSE) of 8.39% and a correlation coefficient of 0.71. Importantly, experiments using UAVSAR data under varying moisture revealed that for optimal parameter inversion, PolSAR data acquisition should be scheduled to avoid periods of low moisture content.
Keywords:
Polarimetric SAR
scattering mechanisms
crop classification
soil moisture inversion
complex agricultural environments
Journal
IF:
2.6
Papers:
1.2W
Citations:
2.7W
View more organizations