食用油红外光谱的分形维数特征

Fractal geometry feature of edible oil FTIR

以傅立叶变换红外光谱为研究对象,探讨了图像分形理论在傅立叶变换红外光谱识别中的应用,提出一种利用分形理论提取食用植物油红外光谱特征的算法.通过对傅立叶变换红外光谱分形维数的比较,达到区分不同光谱的目的,利用开源计算机视觉库OpenCV实现了食用植物油红外光谱分形维数的计算算法,以玉米油和芝麻油的红外光谱为例对该方法进行了说明.实验表明:利用分形维数特征可以对玉米油和芝麻油的真伪进行很好的鉴别。

A feature extraction of edible oil FTIR based on fractal geometry is proposed and theoretical calculation formulas of fractal dimension are introduced in this paper.The calculation method of box-counting dimension based on FTIR is analyzed.Through the comparison of FTIR fractal dimensions,the different spectral could be distinguished.Based on the open source computer vision library OpenCV,the FTIR of edible vegetable oils fractal dimension calculation algorithm has been finished.Taking the corn and sesame oil as examples.The detecting result indicates that the FTIR of corn oil and sesame oil can be discriminated effectively by the difference of box-counting dimension.