Development of Automatic Weeding Thinner for Direct Sowing Sugar Beet (Part 1)

Evaluation of Correct Distinction Rate by Weeding and Thinner Simulations

Abstract

The goal of this study is to develop the automatic weeding thinner using the machine vision for the direct sowing sugar beet. The computer simulations conducted to require the recognition rates of the sugar beet and weeds. As results; the suitable forcible thinning distance (the minimum distance to be vacated) was 10cm (average intrarow spacing; 12cm, set intrarow spacing after the weeding thinner operation; 24cm). If the obtained recognition rates of the sugar beet, ‘87.2%’ and the weeds, ‘94.4%’ were promised in the field test, the loss of sugar beet would be less than 5%. The performance of the automatic weeding thinner was affected by the recognition rate of weeds rather than that of the sugar beet.