节点文献

基于图像识别技术的耐旱小麦种子多指标精选技术研究

Study on Seed Separation of Anti-draught-stress Wheat by Multiple Traits Based on Image Recognition Technology

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 贾佳尹誉霖王建华谢宗铭杨丽明孙群

【Author】 JIA jia;YIN Yulin;WANG Jianhua;XIE Zongming;YANG Liming;SUN Qun;Beijing Key Laboratory of Crop Genetic Improvement,China Agricultural University;College of Science,China Agricultural University;Center for Molecular Agro-biotechnology and Breeding,Xinjiang Academy of Agricultural and Reclamation Science;

【机构】 中国农业大学农学与生物技术学院植物遗传育种学系/北京市作物遗传改良重点实验室新疆农垦科学院分子农业技术育种中心中国农业大学理学院

【摘要】 实验以西农979、百农矮抗58和济麦22三个小麦品种为材料,采用Seed Identification软件获取单粒小麦种子的形态物理指标(RGB、Lab、HSB、长度、宽度和投影面积),然后进行垂直玻璃板发芽试验(干旱胁迫),以发芽第8天的幼苗鲜重作为种子活力指标;种子物理指标与种子活力指标的相关分析表明:幼苗鲜重与长度、宽度、投影面积、Green、Saturation和Red等值(极)显著相关。综合单一指标和多指标精选结果,按照投影面积≥12.82mm~2、Green<98两个指标对西农979小麦种子同时进行精选,其精选效果最好;按照15.06mm~≤投影面积≤18.50 mm~2、Saturation≤48两个指标对百农矮抗58小麦种子同时对百农矮抗58进行精选,其精选效果最好;按照12.42mm~2≤投影面积≤16.32mm~2、Red≥123两个指标同时对济麦22小麦种子精选,精选效果最好。研究不仅揭示了计算机图像识别技术在小麦种子多指标精选上应用的可行性,还为小麦种子的加工和精选提供了理论依据和精选参数。~8

【Abstract】 Physical characteristics affiliated to three wheat varieties,XiNong979,BaiNongAiKang58 and JiMai22,which included RGB,Lab,HSB,length,width and projected area were measured using Seed Identification software.Draught-stress germination test was conducted afterwards.Seedling weight on 8th day during the germination test was defined as seed vigour in this experiment.Correlation analysis between seed vigour and physical traits indicated that seedling weight was correlated to seed length,width,projected area,Green,Saturation and Red on a(highly) significant level.After an overall comparison on advantages of single and multiple traits as criterion,seed separation of XiNong979 should follow two physical traits which referred to as projected area ≥12.82 mm2 and Green <98 simultaneously.The double-physical trait,15.06mm2≤projected area≤18.50 mm2 and Saturation≤48,was the best parameter for seed separation of BaiNongAiKang58.Furthermore,12.42mm2≤projected area≤16.32mm2together with Red≥123 were the best for JiMai22.Not only did the conducted experiment discovered the practical utilization of computer image recognition technology in wheat seed separation,but also provided a series of parameters for seed processing and selection,both theoretically and practically.

【基金】 公益性行业(农业)科研专项项目(201303002);新疆生产建设兵团“十二五”项目(2012BD046)
  • 【会议录名称】 中国作物学会作物种子专业委员会2015年学术年会会议论文集
  • 【会议名称】中国作物学会作物种子专业委员会2015年学术年会
  • 【会议时间】2015-08-15
  • 【会议地点】中国陕西杨凌
  • 【分类号】S512.1
  • 【主办单位】中国作物学会作物种子专业委员会
节点文献中: