基于HSV颜色空间的矿物识别研究*

矿物识别是工艺矿物学的研究基础，针对矿物颜色作为镜下矿物鉴定主要依据的特点，为提高矿物识别效率，降低人工识别成本，提出了一种HSV颜色空间的阈值分割方法。利用双边滤波对采集的矿物显微图像进行预处理，消除噪声干扰，针对不同矿物颜色亮度的差异，在HSV颜色空间下分离H、S、V三通道的颜色分量图提取颜色阈值，并通过阈值分割操作获得目标矿物区域。对磁铁矿和黄铜矿共存的显微图像进行识别，并与传统的大津法和基于HSV颜色空间的阈值分割方法对比。结果表明，基于HSV颜色空间的矿物识别方法能够准确的识别区分开磁铁矿和黄铜矿，分割结果与人工标注的矿物位置基本符合，准确率达95%以上，且提高了识别速度，是机器视觉代替人眼视觉在矿物识别方面的一次探索。

Study on Color Mineral Recognition in HSV Color Space

Mineral identification is the research basis of process mineralogy. According to the characteristics of mineral color as the main basis for microscopic mineral identification, in order to improve the efficiency of mineral identification and reduce the cost of manual identification, a threshold segmentation method of HSV color space is proposed. According to the difference of color brightness of different minerals, the color component images of H, s and V channels are separated in HSV color space, the color threshold is extracted, and the target mineral area is obtained through threshold segmentation. The microscopic images of the coexistence of magnetite and chalcopyrite are recognized and compared with the traditional Otsu method and the threshold segmentation method based on HSV color space. The results show that the mineral recognition method based on HSV color space can accurately distinguish magnetite and chalcopyrite. The segmentation results are basically consistent with the manually marked mineral location, with an accuracy of more than 95%, and improve the recognition speed. It is an exploration of machine vision instead of human vision in mineral recognition.