Method and system for performing color grading on faceted colored gemstone by using digitization technology
By combining digital technology and artificial intelligence models, the subjectivity problem in the color grading of faceted colored gemstones has been solved, achieving more accurate and consistent color ratings. It integrates the advantages of human visual perception and spectral data, improving the objectivity and consistency of grading.
Patent Information
- Application Number
- CN202410584634.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-13
- Publication Date
- 2025-11-14
AI Technical Summary
Existing technologies for color grading of faceted colored gemstones are greatly affected by subjective factors, making it difficult to achieve consistency and accuracy. Furthermore, spectral data analysis has not fully integrated human visual perception and human brain color perception.
By combining digital technology and artificial intelligence models, the gemstone is fixed in position by a mechanical module, while a vision module acquires high-resolution images and spectral data. Computer vision is used in conjunction with a robotic arm to perform color grading, and the process is calibrated with reference to expert evaluation and industry standards.
It achieves more accurate and consistent color grading, reduces the differences in human rating, and improves the consistency and objectivity of grading results.
Abstract
Description
Technical Field
[0001] This invention relates to the field of colored gemstone identification, specifically to a novel method and system for color grading of cut colored gemstones, i.e., faceted colored gemstones (hereinafter also referred to as "gemstones" or "colored gems"). Background Technology
[0002] Colored gemstones are natural products, highly commercialized due to their beautiful appearance and rarity. Colored gemstones typically entering the market come with a certificate from a gemological laboratory, and the color grade indicated on the certificate significantly impacts their marketability and price.
[0003] Grading faceted colored gemstones typically requires the expertise and long-term experience of skilled gemologists. It involves considering factors such as color, clarity, cut, and carat weight. Color grading, in particular, is highly subjective and influenced by factors like cut quality and clarity, making consistency difficult. In practice, inconsistent gradings from different gemologists are common, and even the same gemologist may yield conflicting results. While objective grading methods based on spectral data analysis have been applied, they do not incorporate human visual perception and color perception, thus failing to fully elucidate human perception of color appearance and not representing universal preferences.
[0004] Therefore, a comprehensive approach is needed to combine objective data with subjective perception in order to achieve a more accurate and consistent color grading measurement. Summary of the Invention
[0005] This invention provides a method and system for color grading faceted colored gemstones. The system consists of three modules that digitally mimic the color grading process of an appraiser: a mechanical module for fixing and adjusting the position of the gemstone; a visual simulation module that uses a digital imaging system and spectrophotometer to record the optical performance of the gemstone under multiple CIE standard illumination conditions in the form of high-resolution images, videos, and spectral curves; and an output module, the core of which is a pre-trained and continuously learning artificial intelligence model. This model is based on human color perception mechanisms and uses large amounts of data on gemstone type characteristics and color grade opinions as a training source to provide color grading opinions.
[0006] Detailed description of the invention: The output module involves a pre-trained artificial intelligence model that processes color grading requests using data from the visual simulation module. The model is built upon extensive research into human visual perception and gemstone color judgment: it utilizes machine learning and pattern recognition algorithms to establish a correlation between digitized images and spectral data and human color perception and gemstone grading, thereby achieving accurate and consistent color ratings. To ensure consistency in rating results, its reference database, or database used for training, is integrated with expert evaluations and industry standards to periodically calibrate the module's output. The model can be expanded as the input data accumulates.
[0007] The visual simulation module, also known as the computer vision and spectral analysis module, is designed to use a digital imaging system and a non-contact spectrometer in conjunction with a robotic arm to acquire high-resolution images (pictures and videos) and spectral power distributions of gemstones in a black-and-white field of view from multiple angles under various CIE standard illumination conditions (especially A, D, and F). The spectral data is used for CIE color value calculations, including brightness, saturation, and hue. According to GB / T 32862-2016 Sapphire Grading and GB / T 32863-2016 Ruby Grading, CIE standard illumination D with a color temperature of 4500-5500K and a color rendering index of not less than 90 is used as the primary light source. CIE standard illumination A is designed to reveal the color-change properties of certain gemstones, such as alexandrite and color-change sapphire. CIE standard illumination F is used to reveal the fluorescence properties of certain gemstones, such as ruby, jade, and red spinel; when used as the illumination source, the spectrophotometer acts as a fluorometer to record the fluorescence spectrum curves of the gemstones. In addition to determining the location of data acquisition, computer vision systems also use images and videos to determine the gemstone cutting method and craftsmanship level.
[0008] Robotic arm operating unit: A precision robotic arm equipped with a gripping function can place the gemstone at a suitable distance based on the position of the imaging system and light source, and adjust the angle according to imaging requirements. This operation is linked by computer vision and the robotic arm, with the gemstone's table surface perpendicular to the imaging unit as the initial position, and provides multiple tilt angles to ensure full coverage of the gemstone's surface and internal characteristics.
Claims
1. Claim 1: A big data AI model: Based on extensive research into human visual perception and gemstone color judgment, this AI model utilizes machine learning and pattern recognition algorithms to establish the correlation between digitized images and spectral data and human color perception and gemstone grading, achieving accurate and consistent color ratings. To ensure consistency in rating results, its reference database, or training database, is integrated with expert evaluations and industry standards for periodic calibration of the module's output. The model can expand with the accumulation of input data.
2. Claim 2: Equipped with multiple CIE standard illumination sources, primarily A, D, and F, this digital imaging and spectrophotometer system integrates photoelectric conversion elements to convert light signals into digital information and spectral power distribution curves. When illuminating gemstones using CIE standard illumination F, the spectrometer also functions as a fluorometer. The spectral and imaging equipment is calibrated using a standard white light source.
3. Claim 3: The robotic arm works in conjunction with computer vision to determine the correct distance and position of the gemstone relative to the light source and imaging system, mimicking the observation process of an appraiser by adjusting the gemstone's position from multiple angles to obtain sufficient digital information.