基于计算机图像识别技术的米饭表面孔隙率测定方法研究

    Research on measuring surface porosity of cooked rice based on computer image technology

    • 摘要: 米饭表面孔隙率能够反映米饭的结构特征、物性和口感。为快速准确地测定米饭表面孔隙率,提出了一种基于图像处理技术的测定方法。以米饭饭粒间的空隙作为研究对象,使用OpenCV库进行图像处理,首先将米饭图像进行灰度化并使用Otsu阈值将图像进行分割处理,然后对二值化图像进行除噪处理,利用连接组件分析算法提取米饭表面孔隙区域,计算孔隙面积占比以得到米饭表面孔隙率,并对算法结果进行了验证。结果表明:对5种不同品种大米(黄河粥、珍珠米、长粒香、京贡一号、五常大米)的米饭表面孔隙率进行计算,得到其Photoshop处理与Otsu处理值之间的相对误差分别为1.45%、1.89%、1.64%、1.42%和0.95%;所测得的米饭表面孔隙率的变异系数均在4%以内,说明数据波动性较小、离散程度较低、测量精度较高,米饭孔隙率测定结果与Photoshop处理结果接近,且重复性良好。该方法能快速高效地测定米饭边界的孔隙率,无须单独取样,且米饭试样自身无破损,可为评价米饭品质提供一种新的无损检测技术手段。

       

      Abstract: The surface porosity of cooked rice can reflect the structural characteristics, physical properties, and taste of cooked rice. To rapidly and accurately determine the porosity of rice surface, this paper proposes a method based on image processing technology. Taking the gaps between rice grains as the research object, the OpenCV library was used for image processing. Firstly, the rice image was converted to grayscale and segmented using the Otsu threshold method. Then, noise reduction was applied to the binarized image. The connected component analysis algorithm was utilized to extract the pore areas on the rice surface, and the porosity ratio was calculated to obtain the porosity of the rice surface. The algorithm results were also verified. The results indicated that after calculating the surface porosity of the five kinds of rice, the relative errors between the Otsu treatment method measured values and the Photoshop treatment method values of the rice test samples were 1.45%, 1.89%, 1.64%, 1.42% and 0.95%, respectively. The coefficients of variation for the measured porosity of rice surfaces all within 4%, indicating low data volatile, low dispersion, and high measurement accuracy. The porosity measurement results are close to those obtained through Photoshop processing and exhibit good reproducibility. The method can quickly and efficiently measure changes in the boundary porosity of cooked rice without the need for sampling and without damaging the rice samples themselves. It provides a new non-destructive testing technique for evaluating rice quality.

       

    /

    返回文章
    返回