一种基于图像特征值算法的叶面积测定方法

提出一种基于图像特征值算法的叶面积测定简化方法。应用扫描图像RGB三原色灰度值分离理论,根据植物叶片扫描图像像素点的分布特征,选用蓝色灰度值作为特征值,以扫描图像灰度中间值127作为叶面积图像与背景图像灰度值的判读指标,通过叶片像素点的分布比例计算叶片面积。将已知面积的矩形绿纸片分别随机裁剪成多个碎片,用本文方法测定碎片面积,并分别计算每个叶片的碎片面积之和进行系统精度验证,测定结果与标准面积的相对误差小于0.5%。采集60个水稻叶片分别采用本文方法和复印称重法测定叶片面积,对本文方法进行进一步验证,相关性分析结果表明,二者相关系数r=0.997 1,达极显著水平。本文方法具有较高测定精度,满足叶面积测定要求。

Measurement of leaf area based on algorithm of image characteristic value

A new method for leaf area measurement based on the algorithm of image characteristic value was studied in this research.Leaf area was measured based on the distribution of the scanning picture pixels and the histogram of the picture gray value.RGB tricolor gray value was separated from the picture and blue was selected as image characteristic value.And then the leaf area was determined by measuring the proportion of every leaf pixel.When green shredding units,with known total area,was measured by this method,the relative error of the results between the calculated area and the actual one was less than 0.5%.The correlation coefficient of leaf area measured by the traditional copy method and the scanning method was extremely high(r=0.997 1).