Rice Growth Parameter Estimation Based on Remote Satellite and Unmanned Aerial Vehicle Image Fusion

Precise monitoring of the leaf area index (LAI) and soil–plant analysis development (SPAD, which represents chlorophyll content) at the field level is crucial for enhancing crop yield and formulating agricultural management strategies. Currently, most studies use multispectral sensors mounted on unmanned aerial vehicles (UAVs) to obtain images, whereby the spectral information is utilized to estimate rice growth parameters. Considering the cost of multispectral sensors and factors influencing rice growth parameters, this study integrated satellite remote sensing images with UAV visible-light images to obtain high-resolution multispectral images during key rice growth stages, thereby determining the rice LAI and SPAD on the same day. The vegetation indices and textural features most correlated with rice LAI and SPAD were selected using Pearson correlation analysis, and based on vegetation indices, textural features, and their combinations, regression models were established. The results indicate the following: (1) The fusion of satellite and UAV images, combined with spectral information and textural features, can significantly improve the estimation accuracy of LAI and SPAD compared to using only spectral information or textural features. (2) Sparrow search algorithm-optimized extreme gradient boosting (SSA-XGBoost) regression achieved the highest accuracy, with R2 and RMSE of 0.904 and 0.183 in LAI estimation and 0.857 and 0.882 in SPAD estimation, respectively. This demonstrates that integrating satellite and UAV images, combined with vegetation indices and texture features, can effectively establish rice LAI and SPAD estimation models, using the SSA-optimized XGBoost method, as an effective and feasible solution for precise monitoring of rice growth parameters.