梨冠网蝽数学形态特征的提取与分析
Measurement and analysis of morphometric features of Stephanitis nashi Esaki et Takeya (Hemiptera: Tingidae)
潘鹏亮** 史洪中 尹新明 王国君 刘红敏
点击:1552次 下载:6次
DOI:10.7679/j.issn.2095-1353.2017.064
作者单位:信阳农林学院农学院 信阳 464000
中文关键词:梨冠网蝽,形态特征,雌雄鉴定,数值分类,远程识别
英文关键词: Stephanitis nash, morphological characteristics, gender identification, numerical taxonomy, remote recognition
中文摘要:
【目的】 利用图像处理技术获取昆虫外部数学形态特征,并用于雌雄个体判别,对其远程鉴定具有重要意义。本研究的目的是通过对图像获取技术的研究,获得梨冠网蝽Stephanitis nash Esaki et Takeya雌雄个体差异的参数指标。【方法】 利用中国农业大学IPMist实验室开发的BugShape(v1.0)软件对梨冠网蝽成虫的外部数学形态特征进行提取,并使用SPSS(v22.0)对数据进行分析。【结果】 扫描分辨率高低对部分参数有显著影响,其中周长、中胸宽、中胸宽/腹宽等参数随着分辨率的增加而增大。软件自动测量参数(面积、等效圆半径、偏心率、紧凑度、圆形度)和手动测量参数(中胸宽、腹宽、体长/中胸宽、体长/腹宽、中胸宽/腹宽)在雌雄个体间存在显著差异。在4 800 dpi时,手动测量梨冠网蝽的体长为2.25 mm左右,雌性成虫的中胸宽和腹宽分别比雄性成虫宽0.02 mm和0.19 mm。在1 200 dpi分辨率下,利用手动测量参数对雌雄个体进行判别时,交叉验证判别正确率达99%,仅使用自动测量参数进行判别时,交叉验证判别的正确率为79%,而同时使用两类参数时,交叉验证判别正确率为97%。【结论】 利用图像处理技术对梨冠网蝽进行雌雄识别时扫描的最佳分辨率为1 200 dpi,获取到的部分参数在雌雄个体间存在显著差异,可以作为判别其两性的参考指标。
英文摘要:
[Objectives] To identify diagnostic morphometric features of insect species that allow the sex of individuals to be reliably determined is of great importance to the remote identification of insect species using image processing techniques. The aim of this study was to use image acquisition technology to identify the parameters distinguishing female and male adults of the pear lace bug Stephanitis nash Esaki et Takeya. [Methods] Morphometric characteristics of adults were extracted using the software BugShape v1.0 developed by IPMist Lab. at China Agricultural University and the data were analyzed in SPSS v22.0. Images of adults were created with a flatbed scanner (Epson perfection v370) with a transparency adapter. Fifty samples of each gender were measured at three scanning resolutions (1 200 dpi, 2 400 dpi and 4 800 dpi). [Results] Scanning resolution had a positive effect on some parameters such as perimeter, mesothorax width (MW), and the mesothorax width /abdomen width (MW/AW) ratio, which increased with increasing resolution. Body length/abdomen width (BL/AW), and MW/AW) ratios derived from both automatically measured (area, equivalent radius (ER), eccentricity, compactness and circularity), and manually measured variables (MW, AW, the body length/mesothorax width (BL/MW), significantly differed between female and male adults. At 4 800 dpi scanning resolution, the BL of the pear lace bug was about 2.25 mm and female MW and AW were 0.02 mm and 0.19 mm, respectively, longer than those of males using manually-measured parameters. At 1 200 dpi resolution, discriminant accuracy based on cross validation was 99% using manually measured variables to discriminate the sexes, and 79% using automatically measured variables. However, using both kinds of variables increased the accuracy to 97%. [Conclusion] 1 200 dpi is the best resolution for discriminating male and female pear lace bugs with image processing technology. Some morphometric variables differ significantly between the two sexes, and can therefore serve as standard diagnostic criteria for sexing this species.