Enhanced Neural Network for Rapid Identification of Crop Water and Nitrogen Content Using Multispectral Imaging

Precision irrigation and fertilization in agriculture are vital for sustainable crop production, relying on accurate determination of the crop’s nutritional status. However, there challenges optimizing traditional neural networks to achieve this accurately. This paper aims propose a rapid identification method water nitrogen content using optimized networks. addresses difficulty backpropagation network (BPNN) structure. It uses 179 multi−spectral images crops (such as maize) samples model. Particle swarm optimization (PSO) is applied optimize hidden layer nodes. Additionally, proposes double−hidden−layer structure improve model’s prediction accuracy. The proposed PSO−BPNN model showed 9.87% improvement accuracy compared with BPNN correlation coefficient R2 predicted was 0.9045 0.8734, respectively. experimental results demonstrate high training efficiency lays strong foundation developing precision plans modern holds promising prospects.