Xiaohang Su

Papers

Total Citations

177

H-Index

About

Xiaohang Su is a researcher whose work sits at the intersection of computer vision, deep learning, and agricultural robotics. His primary research focuses on developing efficient, real-time neural network architectures for semantic segmentation and object detection in complex, unstructured environments. Su’s most significant contribution is the application of lightweight deep learning models to practical engineering challenges, particularly in agriculture and mobile robotics. His highly cited 2023 paper, “Lightweight detection networks for tea bud on complex agricultural environment via improved YOLO v4,” has garnered 132 citations, demonstrating its substantial impact on precision agriculture. In this work, Su adapted the YOLOv4 architecture for the challenging task of detecting tea buds in natural, cluttered settings, achieving a balance between speed and accuracy that is critical for real-world deployment. He has also pioneered novel architectures for mobile robots, such as the Efficient Dual-Branch Bottleneck Network (EDBNet) for real-time semantic segmentation with CCD cameras, and the Multi-level Enhancement Layers Network (MELNet), which leverages the Broad Learning System (BLS) framework for efficient scene understanding. Through these contributions, Su has advanced the state of the art in deploying deep learning on resource-constrained platforms for autonomous navigation and agricultural monitoring.