基于彩色图像的柑橘糖度无损分析  

Orange Brix' s Nondestructive Analysis Based on Color Image

作　　者：王雪峰[1] 李晓冬[1] 平藤雅之 

机构地区：[1]中国林业科学研究院资源信息研究所,北京100091 [2]日本北海道农业研究中心

出　　处：《林业科学》2013年第10期88-92,共5页Scientia Silvae Sinicae

基　　金：948项目(2011-4-67);日本学术振兴会资助项目(2007-2009)

摘　　要：以柑橘为例,从图像分析角度研究前景图像的分割提取算法以及其与糖度的关系,把基于确定指数阈值的相关分析法引入到图像分割中,实现对不同性质图像的较为准确分割。结果表明:柑橘内部的糖度含量与其外在的某些彩色图像参数之间存在某种明显的依附关系,用双指数方程能够较好地模拟柑橘彩色图像G分量亮度平均值随糖度的变化趋势。在不需要额外增加光源的情况下,仅根据可见光波段的彩色图像无损获取柑橘糖度含量数值是可能的。Taking orange as example,this article does some research on image foreground segmentation algorithm and the relationship between brix and the images. Correlation analysis method based on coefficient of determination is introduced into image segmentation. And more accurate segmentation results of different images were obtained using this method. The research result shows that there is a close relationship between orange brix and its color image＇s external parameters and the variation tendency of the average G intensity according to the change of brix can be well simulated using double exponential equation. It also shows that in the case of no additional light source,it is possible to know orange brix nondestructively based only on its visible waveband color image.

关 键 词：柑橘 糖度 图像分析 无损测量 

分 类 号：S759.8[农业科学—森林经理学]