Sign in Subscribe
Machine learning

Improved Estimation of Nitrogen Use Efficiency in Maize via Multi-Source Data Fusion

Strategic Insights

— 1 min read

Scientists at Beijing Key Laboratory have developed a method to accurately estimate nitrogen use efficiency in maize using a combination of multi-spectral and LiDAR data acquired by a multi-sensor UAV platform. The technique, which utilizes machine learning algorithms to analyze the data, has shown significant improvements in accuracy compared to traditional methods. According to the research, the multi-source data fusion approach enables rapid, non-destructive nitrogen use efficiency assessment in maize, supporting efficient breeding and precision nitrogen management strategies.

Key Takeaways:

  • The research aims to improve the estimation accuracy of nitrogen use efficiency in maize using a multi-sensor UAV platform and machine learning algorithms.
  • The study utilizes three machine learning algorithms: Partial Least Squares Regression (PLSR), Random Forest Regression (RFR), and Support Vector Machine Regression (SVR) to construct nitrogen use efficiency estimation models.
  • The results demonstrate distinct differences in nitrogen utilization efficiency (NUtE) and nitrogen agronomy efficiency (NAE) among maize cultivars at critical growth stages.
  • The Random Forest Regression (RFR) method obtained the highest model validation accuracy with an average Rtest^2 = 0.68 and RMSEtest = 6.66 kg kg-1.
  • The average accuracy of multi-source data fusion was improved by 20.21% compared to a single data source.
  • The RFR+MS+LiDAR method for nitrogen utilization efficiency (NUtE) estimation obtained the highest model accuracy in the two-year validation dataset with Rtest^2 = 0.86 and RMSEtest = 8.5 kg kg-1.
  • The method proposed in this study mitigates the impact of canopy spectral saturation during the late growth stages of maize, enhancing the accuracy of nitrogen use efficiency estimation.

Statistics:

  • Average accuracy of multi-source data fusion improved by: 20.21%
  • Average Rtest^2 for RFR method: 0.68
  • Average RMSEtest for RFR method: 6.66 kg kg-1
  • Accuracy of RFR+MS+LiDAR method for nitrogen utilization efficiency estimation: Rtest^2 = 0.86, RMSEtest = 8.5 kg kg-1

Sources:

  • Improved Estimation of Nitrogen Use Efficiency In Maize From the Fusion of Uav Multispectral Imagery and Lidar Point Cloud. European Journal of Agronomy, 2025;168.
  • Beijing Key Laboratory Digital Plant, Beijing 100097, People's Republic of China.

Read more