Soil Moisture Inversion Based on Data Augmentation Method Using Multi-Source Remote Sensing Data
Abstract
Soil moisture is an important land environment characteristic that connects agriculture, ecology, and hydrology. Surface soil (SSM) prediction can be used to plan irrigation, monitor water quality, manage resources, estimate agricultural production. Multi-source remote sensing a crucial tool for assessing SSM in areas. The field-measured sample data are required model building accuracy assessment of inversion using data. When the samples insufficient, severely affected. An method suitable small size was proposed. alpha approximation employed expand measured offer more training models. Then, feature parameters were extracted from Sentinel-1 microwave Sentinel-2 optical data, optimized three methods, which Pearson correlation analysis, random forest (RF), principal component analysis. common machine learning models training, RF, support vector regression, genetic algorithm-back propagation neural network, built retrieve SSM. Comparison experiments carried out between various optimization methods experimental results showed after augmentation, enhanced, combination utilizing RF screening had higher accuracy, with coefficient determination 0.7256, root mean square error 0.0539 cm3/cm3, absolute 0.0422 respectively. proposed finally invert regional study area. indicated good performance applications size.