Optical seed sorting for removal of weed seed contaminants from cool‐season grass seed

Maintaining high seed purity is essential for certification and commercial marketing of cool‐season grass seed; however, conventional post‐harvest conditioning equipment often provides incomplete removal of grass weed contaminants because of similarities in seed morphology. Optical seed sorting offers an alternative approach by combining high‐resolution imaging with machine vision to classify and separate individual seeds based on visual characteristics. This study evaluated the performance of a commercial optical seed sorter for removing annual bluegrass ( Poa annua L.) from perennial ryegrass ( Lolium perenne L.) and roughstalk bluegrass ( Poa trivialis L.) from tall fescue ( Schedonorus arundinaceus (Schreb.) Dumort.) seed lots across increasing weed seed contamination levels. Weed seed removal increased with increasing weed seed contamination for both crop–weed combinations. Approximately 84% of P. annua and 69% of P. trivialis were removed from contamination levels evaluated. Crop seed loss occurred in all seed lots and increased with greater weed seed contamination; however, crop seed recovery remained greater than 91% for both grass seed crops. The ability to separate Italian ryegrass ( Lolium multiflorum Lam.) from L. perenne and S. arundinaceus and rattail fescue ( Vulpia myuros (L.) C.C. Gmel) from fine fescue ( Festuca rubra L. subsp. rubra ) was also qualitatively assessed; however, preliminary evaluations indicated that L. multiflorum could not be reliably separated from L. perenne or S. arundinaceus because of similarities in seed appearance, while elongated awns prevented consistent separation of V. myuros from F. rubra . These results demonstrate that optical sorting can complement conventional mechanical conditioning by reducing persistent grass weed contaminants when improvements in seed purity justify modest losses in marketable crop seed.