Abstract: Seedling perpendicularity is one of the most important targets of evaluating transplant machine performance. Three-Dimensional vision system, seedling image were obtained after calibrating camera by computer. Using image processing technique, the angle information of seedling was represented by a line in two dimensional space. Through line reconstruction in three dimensional space based on two cameras and lines, measurement of seedling perpendicularity was completed. The experiments reveal that there is no more difference between this method and manual work.