Sookjai Promprasansuk 1, Khongdet Phasinam 1,2, Naphat Albutt 3, Pongthep Poungthong 4, Dowroong Watcharinrat 5, Kittitat Yinyuan 1 and Wirote Ritthong 1,2,3,6,*
1Faculty of Engineering and Technology, Shinawatra University, Pathum Thani 12160, Thailand; 2Faculty of Engineering and Technology, Pathumthani University, Pathum Thani 12000, Thailand; 3Faculty of Engineering, Bangkok Thonburi University, Bangkok 10170, Thailand; 4School of Engineering, King Mongkut’s Institute of Technology Ladkrabang, Bangkok 10520, Thailand; 5Faculty of Education, Shinawatra University, Pathum Thani 12160, Thailand; 6Research Fellow, INTI International University, Negeri Sembilan 71800, Malaysia
*Corresponding author: wirote.r@siu.ac.th
This work introduces a low-cost, machine-vision-guided four-axis robotic arm for automated pick-and-place sorting of leafy vegetables in Controlled Environment Agriculture (CEA). The system consists of an articulated robotic arm, SG90 servo actuation, a PCA9685 PWM driver, a pneumatic parallel gripper, an Arduino-based controller, an ESP32 Wi-Fi data logging, and an OpenCV-based pipeline for HSV color classification. Three lettuce cultivars (Red Oak, Green Oak, and Red Coral) were tested in 27 trials (3 sessions, 3 trials per cultivar). The system achieved an overall classification accuracy of 96.3%, with one Red Coral sample misclassified due to low color saturation near the HSV threshold boundary. The mean cycle times were 26.73, 32.64, and 37.03 s for Red Oak, Green Oak, and Red Coral lettuce, respectively. The prototype demonstrated stable repeatability, no specimen drops, and no visible leaf damage at a gripping force of 0.8 N. The results demonstrate the practical viability of reducing repetitive manual sorting tasks in small and medium greenhouse operations. However, threshold-based HSV classification is sensitive to lighting and cultivar color variability. Future work should integrate CNN-based classification, improved end-effector compliance, and validation in long-duration greenhouse trials.