Alfalfa yield estimation based on time series of Landsat 8 and PROBA-V images: An investigation of machine learning techniques and spectral-temporal features

• Published in: Remote sensing applications Vol. 25; p. 100657
• Main Authors: Azadbakht, Mohsen, Ashourloo, Davoud, Aghighi, Hossein, Homayouni, Saeid, Shahrabi, Hamid Salehi, Matkan, AliAkbar, Radiom, Soheil
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.01.2022
• Subjects: Alfalfa yield estimation; Feature selection; Machine learning; Precision agriculture; Time series images; Precision agriculture; Feature selection; Alfalfa yield estimation; Machine learning; Time series images
• DOI: 10.1016/j.rsase.2021.100657
• ISSN: 2352-9385

Remote Sensing (RS) technology provides regular monitoring of alfalfa farms, as a major source of forage production worldwide. Phenological characteristics derived from time series of RS imagery provide a valuable information source to estimate crop yield accurately. In this study, we computed spectral vegetation indices (SVIs) from time series of Landsat 8 and PROBA-V images to extract temporal characteristics of alfalfa farms throughout the growth periods in three consecutive years in the Moghan plain, Iran. Then, several new spectral-temporal features were developed based on phenological characteristics of alfalfa during the growing season. Such features particularly describe geometry and variations of the temporal curves and are thus invaluable in describing phenological attributes. We conducted several feature selection methods due to the variety of features. Machine learning (ML) methods, including ridge, lasso, Gaussian Process Regression (GPR), Random Forest Regression (RFR), Boosted Regression Trees (BRT), and Support Vector Regression (ν-SVR) were utilized to build inversion models in order to estimate alfalfa yields, where the results showed satisfactory performance of GPR using the selected features by GS (RMSE=1114.0 kg/ha), RReliefF (RMSE=1157.7 kg/ha) and Boruta (RMSE=1210.2 kg/ha) as compared to the complete feature dataset (RMSE=1237.4 kg/ha). Overall, the developed phenological features coupled with feature selection methods resulted in the appropriate performance of the ML methods in alfalfa yield estimation. •A set of spectral-temporal features is introduced to describe phenological characteristics of alfalfa.•Feature selection methods were implemented to identify the most important variables.•Feature selection could acceptably increase cost-effectiveness of the alfalfa yield estimation procedure.•Long gaps between available cloud-free satellite images consistently affect the performance of the inversion methods.•Area under the spectral-temporal indices and features related to the slope of temporal curves were the most common features.