Portable real-time crop health assessment for diseases and nutrient issues with lightweight neural networks and RGB captures

Biotic and abiotic stresses reduce crop yields, yet existing image‑based diagnostic models are often crop‑specific, computationally intensive, and poorly interpretable. This study present: (i) a multi-crop diagnostic platform integrating lightweight convolutional neural networks (CNNs) with management recommendations, (ii) a hybrid ultralightweight CNN architecture with stacked triple attention mechanisms : (SE) squeeze-and-excitation, (CBAM-Lite) convolutional block attention module, and coordinate attention, optimized for efficient edge deployment; and (iii) a crop transferability index (CTI), a single scalar, to quantify cross-crop generalization. This framework leveraged eight transfer‑learning CNNs to detect 34 stress classes across 11 crops, assembling 47,263 RGB images (44,623 field &amp; 2640 web-scraped), randomly stratified split into training/validation/test sets (70/15/15), and applied 5-fold cross-validation on the training set for model tuning. Gradient‑weighted class activation saliency maps (Grad‑CAM) and Platt-scaling prediction confidence scores were used to localize stress‑related regions and enhance model interpretability. Achieving 99.9 % test accuracy, the hybrid model, ConvNeXtV2_atto, significantly outperformed baseline CNNs, matched ResNet18_weight, and offered 8–14 times fewer parameters than conventional backbones with the most expeditious CPU inference (27.7 FPS). Statistical analysis (Friedman test with Nemenyi post hoc, p < 0.001) confirmed the superiority of the hybrid, ConvNeXtV2_atto and ResNet18_weights backbones. The CTI up to 45.11 % and edge-device performance metrics further reinforce these findings, with MobileNetV2, V2PlantNet and SqueezeNet1_1 remaining competitive alternative due to their lower computational overhead. However, expanding work on zero-shot and unsupervised domain adaptation is needed, as Fréchet inception distance explains only 10–11 % of CTI variance (R² = 0.10–0.11).