This study aimed to improve the efficiency of sugarcane yield prediction and growth surveys by collecting image data using drones and crawler robots and developing a data analysis method using AI. A practical yield prediction model was constructed, and basic verification of a stalk diameter estimation method using an autonomous mobile robot in the field was performed, and good accuracy results were obtained within a limited data collection period.

Sugarcane is the staple crop in the southwestern islands of Japan, and highly accurate yield predictions are extremely important for sugar factory operation planning, but growth status surveys have relied exclusively on visual inspection by experienced workers. However, traditional visual inspections are becoming difficult due to the aging of workers and a lack of successors. Therefore, we aimed to establish a new method of grasping yield and growth information by automatically collecting data using drones and robots and analyzing it with AI.

Kotetsu Arakawa(Drone)

Ryusei Shimizu(Drone)

Shota Kikuchi(Crawler)