THE DESIGN AND IMPLEMENTATION OF A COMPUTER-VISION BASED INTELLIGENT SYSTEM FOR THE REFRACTION INDEX IDENTIFICATION OF GEMSTONES

Refractometer is used to quantitatively determine the degree of light refraction of gemstones. It is one of the most common and effective scientific and teaching instruments in mineralogical laboratories. However, the traditional prismatic refractometer is mainly manual operation with complex operation steps. Its accuracy depends largely on the proficiency of the experimenter. In this research, an intelligent system for optical property identification is developed with cutting-edge technologies, such as opto-mechatronics, computer vision, K-Means arithmetic. The system is composed of seven parts, including refractometer, base, sample rotating apparatus, liquid injection device, video capture device, image recognition software, and general control module. To meet the needs of teaching classroom demonstration, rapid identification of crystals, study of optical axis characteristics and orientation processing of gemstones, four working modes are designated. The inspection results show that the absolute error of the parallel measurements of unknowns is less than 0.0002, which meets the requirements of the Chinese national standard. Therefore, this system can be widely used in various scientific and teaching laboratories.