DEVELOPMENT AND TESTING OF AN INTELLIGENT TOBACCO LEAF HARVESTING ROBOT BASED ON MACHINE VISION

Abstract

The efficiency and quality of tobacco leaf harvesting are crucial for the economic performance of the tobacco industry. To enhance harvesting efficiency, a non-destructive tobacco leaf harvesting robot based on machine vision and robotics technology was developed. Experimental evaluations of key components demonstrated that the biomimetic flexible gripper based on the fin ray effect has good stiffness when the clamping force is 2.5 N, ensuring stable subsequent harvesting and collection of tobacco leaves. The introduction of a 6+1-axis robotic arm significantly expands the working range compared to the original 6-axis design, effectively covering the height of the tobacco column. The robotic arm's speed notably affects harvesting time (P < 0.001), with 1.2 m/s identified as optimal for balancing recognition efficiency and success rates. Additionally, exposure time plays a critical role in success rates (P < 0.001), achieving peaks of 90.00% in the morning and 83.33% in the afternoon at 40000 μs. These advancements enhance tobacco harvesting technology and provide valuable insights for intelligent crop harvesting.