摘要
可见光图像可以反应植物的表型特征,在高通量表型测量的要求下,提出一种从数字图像中提取绿色植物区域的方法。首先对原图像运用高斯函数进行滤波预处理,然后利用图像的绿、红、蓝分量进行色彩分割,并通过设置阈值对图像进行二值化处理,针对二值化处理后的杂散点采用形态学开运算算法去除,最后对未提取出的植物区域的孔洞进行填充。从实验结果可以看出,原图像中的绿色植物区域被完整地提取出来。该方法能快速有效地提取图像中的绿色植物区域,实现对植物的无损测量。
Visible light images can reflect the phenotypic characteristics of plants. A method to extract the green plant from digital image on the requirement of high throughput phenotyping measurement is presented. First, the original image is filtered by using Gaussian function. Then the green, red and blue components of the image are used to make colour segmentation. Binarisation of the image is processed by setting the threshold. For desultory points incurred from the binarisation processing, the morphological opening operation algorithm is applied to re- move them. Finally the cavities in the plant areas which have not been extracted are filled. From the experimen＇ml result it can be seen that the green plant area in original image is completely extracted. The method can quickly and effectively extract the green plant area from image, and achieves the non-destructive measurement of the plant.
出处
《计算机应用与软件》
CSCD
北大核心
2013年第10期266-268,283,共4页
Computer Applications and Software
关键词
植物提取
色彩分割
形态学算法
Plant extraction Colour segmentation Morphological algorithm
