基于面向对象特征提取的植物叶片面积测量方法

发布时间：2018-03-22 12:25

本文选题：叶面积测量　切入点：面向对象　出处：《西北农林科技大学学报(自然科学版)》2017年05期 　论文类型：期刊论文

【摘要】：【目的】提出一种基于面向对象图像特征提取的植物叶面积测量方法,为快速、高精度地测量野外采集植物的叶片面积提供支持。【方法】以扫描图像为基础,借鉴遥感影像的面向对象图像特征提取的思想,获得扫描对象的矢量轮廓,以此计算其面积,并采用AutoCAD绘制的7种多边形进行重复试验,以验证该方法的精确性;然后进一步对青蒿(Artemisia carvifolia)、臭蒿(Artemisia hedinii)、苜蓿(Medicago sativa)3种植物叶片进行重复试验,并与矢量化方法、监督分类方法进行对比,分析该方法在实际叶片测量中的稳定性和计算效率。【结果】利用基于面向对象图像特征提取的植物叶面积测量方法,在进行标准几何图形的面积测量时,该方法的相对误差皆小于1.86%;与矢量化方法、监督分类方法相比,该方法在测量真实植物叶片面积时具有更高的稳定性,而且耗时都小于20s,用时最短;该方法采用IDL模块设计,可实现叶片面积的自动批量处理。【结论】基于面向对象特征提取的植物叶片面积测量方法,是叶片面积高精度及批量自动化测量的一种新途径。
[Abstract]:[objective] to propose a method for measuring plant leaf area based on object oriented image feature extraction, which provides support for rapid and high precision measurement of leaf area of plants collected in the field. [methods] based on scanning images, Based on the idea of feature extraction of object oriented image of remote sensing image, the vector contour of scanned object is obtained, and the area of scanning object is calculated. Seven polygons drawn by AutoCAD are used for repeated experiments to verify the accuracy of the method. Then the leaves of Artemisia carvifolia, Artemisia hediniiae and Medicago sativa)3 species were tested and compared with the vectorization method and the supervised classification method. The stability and computational efficiency of this method in actual blade measurement are analyzed. [results] using the method of plant leaf area measurement based on object oriented image feature extraction, when carrying out the area measurement of standard geometric figure, The relative error of this method is less than 1.86. Compared with vectorization method and supervised classification method, this method is more stable in measuring the leaf area of real plants, and it takes less than 20 s, and the time is the shortest. This method is designed by IDL module. The automatic batch processing of leaf area can be realized. [conclusion] the method of plant leaf area measurement based on object oriented feature extraction is a new way to measure leaf area with high precision and batch automation.
【作者单位】：西北农林科技大学资源环境学院;西北农林科技大学水土保持研究所;浙江省水利水电勘测设计院;中国科学院教育部水土保持与生态环境研究中心;
【基金】：国家自然科学基金项目(41271297)
【分类号】：Q948;TP391.41

【相似文献】