A Multi-Year Study on Rice Morphological Parameter Estimation with X-Band Polsar Data

YUZUGULLU O., Erten E., HAJNSEK I.

APPLIED SCIENCES-BASEL, cilt.7, sa.6, 2017 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 7 Sayı: 6
  • Basım Tarihi: 2017
  • Doi Numarası: 10.3390/app7060602
  • Dergi Adı: APPLIED SCIENCES-BASEL
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Anahtar Kelimeler: polarimetry, SAR, precision agriculture, rice monitoring, stochastic optimization, metamodels, radiative transfer models, electromagnetic scattering models, POLARIMETRIC SAR INTERFEROMETRY, POLYNOMIAL CHAOS EXPANSIONS, PADDY-RICE, HEIGHT, VEGETATION, RETRIEVAL, MODEL
  • İstanbul Teknik Üniversitesi Adresli: Evet

Özet

Rice fields have been monitored with spaceborne Synthetic Aperture Radar (SAR) systems for decades. SAR is an essential source of data and allows for the estimation of plant properties such as canopy height, leaf area index, phenological phase, and yield. However, the information on detailed plant morphology in meter-scale resolution is necessary for the development of better management practices. This letter presents the results of the procedure that estimates the stalk height, leaf length and leaf width of rice fields from a copolar X-band TerraSAR-X time series data based on a priori phenological phase. The methodology includes a computationally efficient stochastic inversion algorithm of a metamodel that mimics a radiative transfer theory-driven electromagnetic scattering (EM) model. The EM model and its metamodel are employed to simulate the backscattering intensities from flooded rice fields based on their simplified physical structures. The results of the inversion procedure are found to be accurate for cultivation seasons from 2013 to 2015 with root mean square errors less than 13.5 cm for stalk height, 7 cm for leaf length, and 4 mm for leaf width parameters. The results of this research provided new perspectives on the use of EM models and computationally efficient metamodels for agriculture management practices.