Teknofest 2023 Contest – Computer Vision Project
Collaborative vision backend for an Autonomous Harvesting Robot, showcased at Teknofest 2023.
Description
Contest project integrating YOLOv8 detection with a Flask UI to demonstrate autonomous harvesting scenarios.
Objective / Aim
Automate detection, classification, and localization to improve harvesting efficiency.
Role & Responsibilities
Co-developed the vision backend, built dataset workflows, implemented cropping/color pipelines, and integrated Flask UI.
Features / Implementation
GPU-trained YOLOv8 models; custom preprocessing/augmentation; real-time demo via Flask.
Challenges & Solutions
Heterogeneous datasets handled with augmentation and preprocessing; tuned YOLOv8/GPU for live demo performance.
Impact / Results
Ranked 25th; showcased to 2.5M+ visitors; validated feasibility of vision-driven harvesting.