Seed size, shape and colour as computer imaging markers of germination quality

摘要

In the two last decades, new techniquesbased on machine vision systems havebeen developed to assist researchers andanalysts in seed quality testing and sorting, in order to try to overcome some ofthe operational limitations of the standardmethods of testing (AOSA, 2002; ISTA,2005). Various methods use CCD-cam-eras and flat bed scanners to acquire seedimages, high power computers, and fastimage analysis software packages that allow rapid data processing and storage ofdata on hard disk (Chen and Sun, 1991).These outcomes suggest computerisedimage analysis is a promising techniqueas an approach to studying seed biology,with further potential in seed quality testing procedures. This paper is not intendedas an overview of a wide range of applications of image analysis techniques onplant product evaluation, carried out indifferent laboratories; the author suggestsa previous review article (Dell'Aquila,2006a) in this respect. Rather, this paperfocuses on our experiences in the development of a computerised image analysissystem to study seed morphology, germination and radicle growth.