摘要
为评价花生图像质量,首先通过阈值分割算法分割花生图像,接着进行开操作,然后根据连通区域个数得出花生粒数作为评价标准。从3种背景颜色和5种光照强度的试验中得出在黑色背景50～90 lx的光强下效果较好。研究将有助于花生自动检测。
To evaluate the quality of the peanut image,the peanut image was first segmented with thresholding segmentation algorithm.Then the number of peanut in image as evaluation standard was obtained with counting the connected regions after open operation.The experiment including 3 colors and 5 different light intensity showed that the better images are acquired in the light intensity of 50-90 lx with black background.It should be helpful for peanut automatic text.
出处
《安徽农业科学》
CAS
北大核心
2011年第33期20901-20901,20922,共2页
Journal of Anhui Agricultural Sciences
关键词
背景颜色
光强
花生图像
Background color
Light intensity
Peanut image
