Abstract—We construct and use wearable sensing systems and various cameras to analyze the characteristics of the motions of trained workers and beginners (sometimes including semi-beginners) in non-specific agricultural jobs, and the differences between them. In recent sequential studies, we developed multitudinous, coverall analysis systems to address various agricultural challenges. We have been contributing to them with investigations verifying the accuracy and utility of our kinematic direct sensing and semi-original program-based visual analysis systems for workers and trainers engaged in the pruning of tree branches using special small saws. Pruning tasks include cutting tree branches and forming shapes to improve ventilation for efficient nourishment and promotion of tree growth. Other purposes of these tasks are to make the trees appear beautiful and to prevent illnesses and breeding of noxious insects. The research analysis is based on nine selected optical flow (OF)–based numerical items (features) used in many other scientific fields. These are extracted from OF vectors calculated from the differences between two successive frames of the obtained digital visual data. The targeted experimental field is situated in the Graduate School of Agriculture of the University of Tokyo in Japan, where the targeted trees are common and adequate for the trials. The targeted task of pruning tree branches is one of the most common movements worldwide, which is why our measurements and proposed indicators are expected to be useful in the future in agricultural fields, especially in developing countries and trend agricultural schools. 
 

Index Terms—tree pruning, visual data analysis, OpenCV, optical flow–based factors, wearable sensing modules

Cite: Shinji Kawakura and Ryosuke Shibasaki, "Optical Flow–Based Study Related to Outdoor Tree Pruning Using OpenCV Utilities and Captured Visual Data," Journal of Advanced Agricultural Technologies, Vol. 6, No. 1, pp. 78-82, March 2019. Doi: 10.18178/joaat.6.1.78-82