We conduct research on measurement systems and environment recognition technologies based on high-speed image processing in field robots. Especially targeting autonomous driving and agricultural robots, we aim to establish a new architecture that combines real-time performance and advanced recognition.

In-vehicle image processing

• Inter-vehicle distance measurement using spatio-temporal shared filter with stereo vision
• Separation and identification of overlapping vehicles using contour information
• Seamless tracking based on hybrid recognition
• Traffic light detection based on high-frequency blinking pattern extraction

Plant recognition

• Weed identification using deep learning in red perilla cultivation

---

[Home]