Improvement of the SWAP-Hapke Model and Evaluation of Multiangle Retrieval Accuracy

Soil water content (SWC) is an important reference for reflecting agricultural drought conditions and irrigation management. Accurate acquisition and monitoring of SWC information is crucial for water-saving agriculture and soil protection. Optical remote sensing technology has the advantages of high efficiency, continuity, noncontact, and wide regional coverage, which is one of the most commonly used SWC estimating methods. First, this work studies the best modeling band and puts forward the band that is sensitive to SWC but insensitive to the illumination and view angle as the modeling band. Then, the process of the internal reflection of incident radiation caused by the water layer around soil particles is added to improve the SWAP-Hapke model. This work innovatively proposed using absorbance instead of reflectance to select the modeling band and establish an improved SWAP-Hapke model. In addition, a complete multiangle and multi-SWC gradient dataset is used to improve the estimation accuracy of model parameters. Finally, the multiangle retrieval accuracy of the improved SWAP-Hapke model was evaluated. The results show that the improved SWAP-Hapke model based on absorbance has obviously higher retrieval accuracy in different view and illumination directions, and its accuracy is less affected by the illumination and view geometry. The average <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">R²</i> of the improved SWAP-Hapke model established by using reflectance and absorbance can reach 0.89 and 0.92, and the average of the root-mean-square error is 4% and 2.4%. In addition, the five optimal bands for establishing the SWC retrieval model are the absorbance at 1989, 1462, 2357, 2190 nm, and reflectance at 1460 nm. Finally, the improved SWAP-Hapke model is applied to the public AAV dataset, and the retrieval results are relatively consistent with measured SWC.