In Japan, ninety seven percent of sugar beets are in cultivation by transplanting. About 200 man-hour per hectare of labor is still required in this culture. To decrease more labor in sugar beets growing, the direct sowing culture will be recommended. In direct sowing culture, because of manual work, nearly 40 man-hour per hectare of labor is put into thinning work at present. To decrease the labor in direct sowing culture, automation of thinning work is essential.
In order to develop an automatic sugar beet thinner, the algorithm for serial image capturing, crop extraction and crop selection to be thinned was developed. The indoor model experiment was conducted to evaluate a feasibility of the algorithm. The experiment apparatus consists of video camera (SONY XC-009), PC (NEC PC9801Xa) with image capture board (I/O DATA GV98X) and one-board computer for driving mechanism. A few misjudge was occurred at processing of thinning selection. The mean location error of thinning signal, however, was about 6mm, and max. error was 35mm. As a result of discussion, it was concluded that the thinning algorithm would have a certain practicability.