应用机器视觉技术筛选射干种子精选指标的研究
Belamcanda chinensis (L.) DC.seeds selection based on machine vision technology

为提高射干种子活力,利用Seed Identification(种子形态自动化识别软件)快速获取300粒射干种子的物理特性(颜色信息、长度、宽度、投影面积),并确定出适宜的种子精选参数。结果表明:射干幼苗苗长与种子颜色信息R(Red,红基色)、S(Saturation,饱和度)、I(Intensity,亮度)、a(从红色至绿色的范围)值呈显著相关,鲜重与种子R、S、I、a值呈极显著相关。经计算得出按R≤40、S≤5、I≤17、a≤0对射干种子进行精选,最多可将发芽率提高9.9%。验证试验结果表明:按以上指标精选后的射干种子发芽率由初始的56.0%分别提高至65.0%、62.0%、64.0%、59.0%;简易活力指数由初始的0.010 9分别提高到了0.016 1、0.017 8、0.015 6、0.012 7。研究表明按R≤40、S≤5、I≤17、a≤0对射干种子进行精选可提高射干种子活力。
In order to improve seed vigor of Belamcanda chinensis (L.) DC.seeds, the physical characteristics (length, width, projection area, color information) of B.chinensis (L.) DC.seeds was quickly identified by the Seed Identification software.The results showed that:Fresh weight, seedling length and simple activity index of B.chinensis (L.) DC.seedling increased gradually with the decrease of R(Red), S(Saturation), I(Intensity) and a(its position between red and green), and reached a significant or highly significant correlation.Selected by R≤40, S≤5, I≤17, a≤0, germination rate can be increased by 9.9% at most.The verification test shows that:After selection according to the above indicators, seed germination rate was improved from 56.0% to 65.0%, 62.0%, 64.0%, 59.0%;simple activity index was improved from 0.010 9 to 0.016 1, 0.017 8, 0.015 6, 0.012 7.These findings suggested that selection of seeds by R≤40, S≤5, I≤17, a≤0 can improve seed vigor.