Development of automatic system for the seedling transplanter based on machine vision technology

This paper presents an automatic system for seedling transplanter based on machine vision technology. Machine vision technology was used to detect the state of seedlings in the plug tray, locating empty and unhealthy seedling holes, guiding the manipulator to transplant the seedling. The methods of removing noises in binary image were compared, and the more suitable method was selected in this study. The algorithms of image processing were used for capturing the seedling characters, the results showed that the success rate of identification was good. Overlapping of the border seedlings and extruding leaves from neighbouring cells always lead to the identification failures. In this research, tomato seedlings were used as samples, and good identification rate for suitable seedling was obtained (96.2%) when using the second method. The result shows this method could be used in various growing status of seedling, and this technology could be used for automatic seedling transplanting robot.