刘博艺,唐湘滟,程杰仁.基于多生长时期模板匹配的玉米螟识别方法[J].计算机科学,2018,45(4):106-111, 142
基于多生长时期模板匹配的玉米螟识别方法
Recognition Method for Corn Borer Based on Templates Matching in Muliple Growth Periods
投稿时间:2017-05-03  修订日期:2017-08-06
DOI:10.11896/j.issn.1002-137X.2018.04.016
中文关键词:  虫害识别,玉米螟,直方图反向映射,模板匹配,Hu矩
英文关键词:Insect pests recognition,Corn borer,Reverse mapping of histogram,Template matching,Hu moment
基金项目:本文受国家自然科学基金(61363071,61762033),海南省自然科学基金(617048),海南大学博士启动基金(kyqd1328),海南大学青年基金(qnjj14444),海南大学研究生实践创新项目基金资助
作者单位E-mail
刘博艺 海南大学信息科学技术学院 海口570228
中国科学院大学 北京100000 
 
唐湘滟 海南大学信息科学技术学院 海口570228 tangxy36@163.com 
程杰仁 海南大学信息科学技术学院 海口570228
海南大学南海海洋资源利用国家重点实验室 海口570228 
 
摘要点击次数: 245
全文下载次数: 184
中文摘要:
      玉米螟是玉米种植中的主要虫害之一。为了解决人工识别的劳动强度大,且识别不够准确、及时的问题,文中提出了一种在自然场景下基于多生长时期模板匹配的不同生长时期亚洲玉米螟的识别方法。该方法首先对获取到的图像进行数学形态学预处理;其次利用直方图反向映射法和多模板图像得到总的概率图像;然后利用约束空间的大津法对二值图像进行轮廓提取,并根据周长和面积特征进行初步筛选;最后结合基准轮廓,利用Hu矩等特征选出符合亚洲玉米螟特征的轮廓,进而得出识别结果并以三角形标记。实验和理论分析证明,在复杂自然场景图像中,该方法不仅时效性强,而且具有很好的识别准确度,能够有效降低不同生长时期的玉米螟颜色变化带来的影响。
英文摘要:
      Corn borer is one of main pests encountered in the corn.In order to solve the problems of high labor intensity,inaccuracy,and being not in time in artificial recognition,a novel identification method for Asiatic corn borer was proposed under natural scenes in this paper,which is based on reverse mapping of histogram and multi-template matching of contours.Firstly,this method performs mathematical morphology preprocessing for the obtained image.Secondly,the total probability image is obtained by using reverse mapping of histogram method and multi-template images,and then image contour can be extracted quickly and accurately by using constraint Otsu,and can be preliminary filtrated according to the perimeter and area characteristics of corn borer.Finally,the contours matched with the characteristics of Asiatic corn borer are selected by using Hu moment characters between multiple reference contours and the obtained target contours,and then identification results with triangle mark are obtained.The experimental results and theoretical analysis show that the proposed method has high timeliness and high recognition accuracy in complex natural scenes.
查看全文  查看/发表评论  下载PDF阅读器