Kaiming Fu

Biography

Kaiming (Kimi) Fu is a fifth-year doctoral student in the Electrical and Computer Engineering at the University of California, Davis. He is working with guidance from Prof. Stavros G. Vougioukas and Prof. Brian N. Bailey. Before joining UC Davis, he received the M.S. degree in Mechanical Engineering from Purdue University in West Lafayette, Indiana. His research is interdisciplinary, incorporating his expertise in engineering and computer science to focus on machine learning, computer vision and agricultural robotics. His research interests include:

  • Multispectral and Synthetic Imaging for Tiny Object Detection.
  • Integrated 2D and 3D Fruit Mapping for Optimized Harvesting Simulation and Planning.
  • Simulation Design and Optimization for Robotics, along with Conducting Interference Analysis using CUDA and Voxel Technology.
  • Exploring Diverse Applications of Helios (A Versatile Modeling Framework).

Ongoing Research Projects

  • Walnut Detection Through Deep Learning Enhanced by Multispectral Synthetic Images
    • We are using YOLOv5 to detect walnuts, trained on a enriched dataset that includes both real and synthetic RGB and NIR images. The challenge lies in accurately distinguishing between walnuts and leaves due to their similar shapes, colors, and textures.
  • Model-based Design of Mechanized Orchard-harvesting Systems
    • We are currently building model-based design tools to enable researchers and developers to investigate the interrelationships among orchard layout, tree canopy geometry and spatial fruit distribution, harvester design, and worker activities. Such tools can accelerate the development of next-generation orchard mechanization and automation systems.

Challenges

  • Annual Farm Robotics Challenge [More Information]
    • Team Leader. Collabrating with Dr. Juan Villacrés and Dr. Guilherme De Moura Araujo on Supervising.
    • Grand Prize Winner among National-wide Universities and Colleges.
    • Brief Intro:
      • Farm Robotics Challenge is open to any university or college in the U.S. Student teams will be asked to address a production farming topic on any crop or size of farm, with a desired focus on small farms, by automating an essential farm-related task using the farm-ng robotics platform.
  • “Inceptio-Tsinghua AIR Cup” Autonomous Driving Challenge [More Information]
    • 1st Prize Winner among 1067 teams.
    • Brief Intro:
      • Autonomous Driving Challenge provides scenarios and data for high-speed heavy trucks and complex urban roads, while the data for the competition questions are all derived from real scenarios. The goal is to develope an efficient decision-making algorithm for self-driving semi-trucks and sedans to navigate complex urban and highway environments collision-free.