Automated Ripe Fruits Detection and Sorting Technique

At small-scale industries, there is a need of automation for sorting fruits and vegetables. With the technology of computer vision and robotics, we can easily automate the process and reduce repetitive manual labour.

For our undergraduate design project, my teammates and I considered this broad application and proposed an automated ripe fruits detection and sorting technique. This project was supervised by

Prof. Jyoti Singhai wherein I worked as a group leader with five of my colleagues.

We used watershed algorithm to segment fruits from the background and calculated the centroid for each unit. To get maximum accuracy while avoiding varied illumination effects, we calibrated color parameters manually through trial and error.

Our algorithm performed real-time automated segmentation of raw and ripe fruits through images from a raspberyy pi camera. We further combined this segmentation technique with our 4-dof robotic arm which is capable of grasping, lifting and placing objects. We improved the motor acceleration profile to perform smooth transitioning between different poses of the end-effector.

Overall, we successfully implemented the hardware interfacing and software architecture for a robotic arm assembly intended to sort raw and ripe fruits.