Forest Height Estimation Based on P-Band Pol-InSAR Modeling and Multi-Baseline Inversion

• Main Authors: Xiaofan Sun, Bingnan Wang, Maosheng Xiang, Liangjiang Zhou, Shuai Jiang
• Format: Article
• Language: English
• Published: MDPI AG 2020-04-01
• Series: Remote Sensing
• Subjects: P-band polarimetric interferometric synthetic aperture radar (Pol-InSAR); forest vertical structure; Gaussian vertical backscatter (GVB) volume; multi-baseline optimization
• Online Access: https://www.mdpi.com/2072-4292/12/8/1319

The Gaussian vertical backscatter (GVB) model has a pivotal role in describing the forest vertical structure more accurately, which is reflected by P-band polarimetric interferometric synthetic aperture radar (Pol-InSAR) with strong penetrability. The model uses a three-dimensional parameter space (forest height, Gaussian mean representing the strongest backscattered power elevation, and the corresponding standard deviation) to interpret the forest vertical structure. This paper establishes a two-dimensional GVB model by simplifying the three-dimensional one. Specifically, the two-dimensional GVB model includes the following three cases: the Gaussian mean is located at the bottom of the canopy, the Gaussian mean is located at the top of the canopy, as well as a constant volume profile. In the first two cases, only the forest height and the Gaussian standard deviation are variable. The above approximation operation generates a two-dimensional volume only coherence solution space on the complex plane. Based on the established two-dimensional GVB model, the three-baseline inversion is achieved without the null ground-to-volume ratio assumption. The proposed method improves the performance by 18.62% compared to the three-baseline Random Volume over Ground (RVoG) model inversion. In particular, in the area where the radar incidence angle is less than 0.6 rad, the proposed method improves the inversion accuracy by 34.71%. It suggests that the two-dimensional GVB model reduces the GVB model complexity while maintaining a strong description ability.