基于机器视觉判别瓜类嫁接苗成活的试验研究

Experimental Study on Discriminating the Survival of Cucurbits Grafted Seedlings Based on Machine Vision

嫁接育苗是解决蔬菜生产土传病害的重要手段,嫁接苗愈合是嫁接育苗生产的重要环节,而目前嫁接苗愈合成活状态的判断主要依靠人工观察接穗真叶是否出现,以及根据愈合状态判断迟滞,降低了嫁接苗规模化生产时愈合装置的利用率。为此,提出了基于机器视觉技术在瓜类嫁接苗愈合期实时观测嫁接苗生长状态的方法,通过比对愈合初期嫁接苗的接穗子叶相似度判断嫁接苗的愈合成活状态。试验研究表明:未成活嫁接苗在嫁接后的2~3天接穗子叶形态急剧变化,相似度下降至0%;而成活嫁接苗接穗子叶形态变化缓慢,相似度在30%~40%之间波动。该方法与接穗真叶观察法相比,可预计提前1~2天确定嫁接苗的成活状态,缩短了嫁接苗的愈合周期,提高了嫁接苗愈合装置的利用率。

Grafting seedlings is an important means to solve soil-borne diseases in vegetable production.The healing of grafted seedlings is an important link in the production of grafted seedlings.At present,the judgment of grafted seedlings’survival state mainly depends on the artificial observation of the appearance of true scion leaves.This judgment method has a delay in judging the healing state,which affects the efficient utilization of the healing device in large-scale production of grafted seedlings.In order to monitor the growth status of grafted seedlings of cucurbits and identify the survival grafted seedlings,this paper proposes to use machine vision technology to compare the similarity of the scion cotyledon of grafted seedlings at the early stage of healing.The results showed that the cotyledon morphology of dead grafted seedlings changed sharply from 2 to 3 days after grafting,and the similarity decreased to 0%.However,the morphology of the survival grafted seedlings changed slowly and the similarity fluctuated between 30%and 40%.Compared with the traditional method,this method can discriminate the survival state of grafted seedlings 1-2 days ahead of schedule,so as to speeding up the healing cycle of grafted seedlings and improving the utilization rate of the healing device of grafted seedlings.