论文部分内容阅读
目的:实现林下山参的数字化鉴别。方法:应用Matlab 2014a数字图像处理工具箱对林下山参图像信息进行处理。从林下山参的芦、体、纹、皮、须入手,提取图像的形状特征、颜色特征及纹理特征共18个,构建数字化林下山参数据库模型。结果:本研究选择的林下山参特征属性集可满足林下山参数字化鉴别,林下山参识别率为96%。结论:本研究在机器视觉下提取林下山参的大量特征,能够以无损的方式实现林下山参的快速检测。
Objective: To realize digital identification of understory ginseng. Methods: Apply the Matlab 2014a digital image processing toolbox to deal with the image information of understory ginseng. From the undergrowth of ginseng, body, grain, skin, to be started, the extraction of the image shape, color and texture features a total of 18, to build digital forest under the ginseng database model. Results: The characteristic attribute set of understory ginseng selected in this study can meet the characterization identification of understory parameters. The identification rate of understory ginseng is 96%. Conclusion: In this study, we extracted a large number of characteristics of Grifola frondosa under machine vision and can quickly detect Grifola frondosa in a non-destructive way.