Study on visual inspection for sampling using laser cutting in corn breeding

Objective To cut corn seed using laser guided by machine vision technology, and to automate sampling of molecular marker-assisted corn breeding.

Method A system for quality inspection and posture recognition based on machine vision was designed. An algorithm was developed to inspect the shape quality and posture of corn seeds. The characteristic parameters of corn seeds were extracted to estimate if the corn seeds were qualified or not, the unqualified ones were removed, and the postures of the qualified ones were calculated so that the posture adjustment device could adjust each seed to the ideal cutting position.

Result The time of algorithm process was about 0.6 s and the mean absolute error of attitude angle was 2.00°, which met the actual precision requirements.

Conclusion This method is helpful to realize the automation of sampling using laser cutting in molecular marker-assisted corn breeding.