基于近红外图像纹理分析检测作物叶片含水率的研究

    Foundamental Study on Determination of Leaves' Moisture Content Based on Near Infrared Image Texture Analysis

    • 摘要: 利用水分子对960nm附近的红外光有较强的吸收这一特性,采用图像处理技术中纹理分析的灰度梯度共生矩阵法分别对两种作物(苞菜和青菜)叶片的近红外图像求出各自的特征量。实验结果表明,特征量与叶片干基含水率之间的线性关系不仅和特征量有关,而且和拍摄图像时的实验因素(如摄像头的工作距离,光圈,光强等)有关,另外还取决于梯度图像灰度级的数目。

       

      Abstract: Water molecule absorbs light near 960 nm more than that of others. With this characteristic, texture anlysis was applied to near infrared image of two species vegetable leaves. This work is an initial research from which some physiological information of vegetable were obtained by image processing technology. Characteristic values were obtained from near infrared image of vegetable leaves by greylevel gradient occurrence matrices approach. Linear relationships between characteristic values and dry basis moisture content of leaf varied not only with varied characteristic values but also with experiment factors ( i.e. work distance of ICCD, diaphragm, intensity of light, etc.). The relationships are also related to the grayscale of gradient image.

       

    /

    返回文章
    返回