Traditional melissopalynology is a time-consuming and subjective process, often taking 4-6 hours per sample. We present an automated, high-throughput microscopy system that integrates H∞ robust mechanical control with advanced deep learning pipelines for the precise counting, classification, and morphological analysis of pollen grains from Bio Bio region in south central territory in Chile. Our system employs U2-Net for salient object detection and a DINOv2 Vision Transformer backbone trained via Deep Metric Learning for classification. By integrating Gradient-Weighted Attention, the model provides human-interpretable texture and diagnostic feature annotations. The system achieves a 95.8% classification recall and a 6x processing speedup compared to manual expert analysis.
@article{arxiv.2604.16743,
title = {Automated Palynological Analysis System: Integrating Deep Metric Learning and $U^{2}$-Net Detection in $H\infty$ bright field microscopy},
author = {J. Staforelli-Vivanco and R. Jofré and B. Muñoz and V. Salamanca and P. Coelho and I. Sanhueza and L. Viafora and C. Toro and J. Troncoso and M. Rondanelli-Reyes and I. Lamas},
journal= {arXiv preprint arXiv:2604.16743},
year = {2026}
}