基于RGB色彩空间的花岗岩组成成分边界曲线识别与提取

Boundary curve recognition and extraction of granite composition based on RGB color space

花岗岩是由石英、长石和云母3种物质组成,因此对花岗岩的组成成分之间的边界曲线进行识别和提取,对分析花岗岩力学性能显得尤为重要。普通的阈值分割方法(如二值分割、分水岭阈值分割等)多数只适用于分割2种介质,因而难以对花岗岩石英、长石和云母3种介质进行有效分割。已有的研究对花岗岩组分的分割结果较为粗糙,且存在误识别等问题。本实验制作4组100mm×100mm的花岗岩切片,基于RGB(红绿蓝)色彩空间的彩色图像处理技术,分别对切片图像的RGB灰度值进行分割,能够实现对花岗岩的石英、钾钠长石和云母进行精确区分。并利用边缘检测技术对其边界曲线进行有效识别和矢量提取,为后续建立数值模型和有限元分析提供了数据基础。

Granite is composed of Quartz, Feldspar and Mica. So the boundary curve recognition and extraction of Quartz, Feld-spar and Mica are essential for the analysis of mechanical properties. Most threshold methods （such as the binary segmentation, watershed threshold segmentation, etc. ） apply only to segment two mediums. Thus it is difficult to segment the three mediums of Quartz,Feldspar and Mica. The existing research on the effect of the segmentation of the granite is rough, and there are some problems such as false identification Four 100 mmX 100 mm granite sections were prepared for this experiment. In this experiment, the gray value of granite image is segmented, which is based on the RGB （red green blue） color space of the color image processing technology. The Quartz, Feldspar and Mica were successfully distinguished. The boundary curves were also successful identified and extracted by using edge detection technique. It provides the basis for the analysis of the failure mechanism and mechanical properties of granite.