In the vast landscapes of Inner Mongolia, where the medicinal plant Glycyrrhiza uralensis, commonly known as licorice, thrives, a groundbreaking study led by Ao Zhang from the College of Pharmacy at Inner Mongolia Medical University is revolutionizing agricultural practices. Zhang and his team have harnessed the power of unmanned aerial vehicle (UAV) multispectral sensors to monitor and predict the growth and yield of this valuable medicinal plant, offering a novel solution for standardized production.
The study, published in *Frontiers in Plant Science* (translated as “Plant Science Frontiers”), addresses a critical gap in current agricultural technologies. Traditional water and fertilizer monitoring systems often fall short in meeting the demands of large-scale cultivation. “Our goal was to develop an integrated model tailored specifically for medicinal plants like Glycyrrhiza uralensis,” Zhang explains. By collecting UAV multispectral images and extracting vegetation indices (VIs), the team synchronized these data with field phenotypic indicators (PIs) such as plant height, tiller number, soil plant analysis development values, and nitrogen content.
The research employed advanced machine learning algorithms, including backpropagation neural networks (BP), support vector machines (SVM), and random forests (RF), to evaluate and predict growth dynamics and yield. The results were impressive. The random forest algorithm and the excess green index (EXG) demonstrated remarkable versatility in growth monitoring and yield prediction. “We found that the random forest algorithm and the excess green index were particularly effective,” Zhang notes. “They provided high-precision predictions, with the green leaf index in the BP algorithm achieving the highest accuracy of R² = 0.94 for plant height prediction.”
The study also revealed that integrating multiple indicators significantly enhanced model performance. Vegetation indices and phenotypic indicators exhibited comparable predictive capacity for yield, with multi-indicators integrated modeling achieving R² values of 0.87 under RF algorithms for VIs and 0.81 using BP algorithms for PIs. “Plant height emerged as the central predictor, but other parameters provided supplementary diagnostic value through complementarity effects,” Zhang adds.
The implications of this research are far-reaching, particularly for the energy sector. Medicinal plants like Glycyrrhiza uralensis are not only valuable for their pharmaceutical properties but also for their potential in bioenergy production. Efficient monitoring and yield prediction can optimize resource allocation, reduce costs, and enhance productivity. “This method offers a practical, time-efficient, and high-precision approach for growth monitoring and yield prediction,” Zhang states. “It can be a game-changer for standardized production of medicinal plant resources.”
As the world increasingly turns to renewable and sustainable resources, the ability to monitor and predict crop growth with precision becomes ever more critical. Zhang’s research paves the way for future developments in agricultural technology, offering a blueprint for integrating advanced remote sensing and machine learning techniques into large-scale cultivation practices. The study not only advances our understanding of medicinal plant cultivation but also sets a new standard for agricultural innovation.