Continuous line for automatic diamond color grading evaluation

Abstract

The automatic diamond color grader can carry out color grading on diamonds on a continuous line, and manual color evaluation is not needed. And a sample diamond is conveyed by the input conveying belt and is placed on the mounting seat on the rotating plate by the mechanical arm. Nm mounting seats are arranged on the rotating plate, and Nr colorimetric stones with known color grades and Ns sample diamonds are mounted on the mounting seats. The camera shoots pictures of all diamonds on the rotating plate, then the stepping motor rotates the rotating plate by 360 / Nm degrees, and another picture is shot. The operation is repeated for Nm times, so that the sample diamond and all the colorimetric stones are respectively imaged at all Nm possible rotating positions. Averaging the pixels of all positions of each diamond to obtain x and y average values of each diamond; and the color grade of the sample diamond is the closest colorimetric stone with light x and y average values.

Description

[Technical field]

[0001] The present invention relates to an automatic detection machine, and in particular to an automatic color grading machine for diamonds. [Background technology]

[0002] Diamonds have long been one of the most sought after of all gemstones. Despite their relative scarcity, their widespread availability has fueled consumer demand. Because every diamond is different, their price and value can vary widely.

[0003] The quality of diamonds is based on the 4Cs: color, clarity, cut and carat weight. A diamond with high clarity is more valuable than a diamond with a light yellow color, which in turn is more valuable than a lighter, lighter colored diamond. Usually, someone with professional training and years of experience will compare the diamond to be evaluated with a series of diamond master stones of known color grades.

[0004] Figure 1 shows the color grade of diamonds. Based on color alone, D-grade colorless diamonds are the most valuable, while Z-grade or above (>Z) light-colored diamonds are the least valuable. D, E, and F are colorless, G, H, I, and J are nearly colorless, and K, L, and M are lighter in color. N to R grades have increasingly lighter colors, while light-colored diamonds are divided into S to Z grades.

[0005] Since its introduction in the early 1950s, the Gemological Institute of America (GIA) D-to-Z scale has been used to color grade the vast majority of colorless to light yellow gem-quality polished diamonds and issue laboratory reports. The D-to-Z color grading is based on the observations of a trained observer who compares the diamond to be evaluated to color master stones at known locations on the grading chart. The GIA master stones are at the highest point of their respective grading ranges. Diamonds that are identical to the G master stones are graded G. If the color is slightly off, they are graded F.

[0006] Typically, a trained person will line up known master stones on a matte white background table under a light source with a color temperature of 5500-6500K. Both the diamond to be evaluated and the master stones are examined at a 45-degree angle, with the gemstones directly in front of the light source. The sample diamond to be evaluated is usually examined from several different directions.

[0007] Color can be qualitatively assessed by humans. However, color can also be defined by a color space, such as the International Commission on Illumination (CIE) RGB XYZ, xy, or lightness, aperture, brightness (LAB) color space. Color can be represented by two or three numeric values ​​in a color space. For example, chromaticity in the CIE XYZ color space is specified by two derived parameters x and y, and two of the three standardized values ​​are functions of all three tristimulus values ​​X, Y, and Z. The CIE xy chromaticity diagram shows the brightest colors with the highest saturation. Machines can detect color and report numeric values ​​in the selected color space.

[0008] Figure 2 shows a series of diamond master stones evaluated in the CIE 1931 color space chromaticity diagram. When normalized, the CIE 1931 color space represents color using x, y values ​​between 0 and 1. The same 10 master stones were evaluated using the color tester in three replicates.

[0009] In the first test 12, master stone D has the lowest x, y values, and master stone M has the highest x, y values. The x, y values ​​of master stones E to K increase in sequence, but then the x, y values ​​of master stone L are lower than those of master stone K, although the x, y values ​​of master stone K should be lower than those of master stone L. This regression of the x, y values ​​of master stone L in test 12 is obviously wrong.

[0010] In the second test 14, the x,y values ​​of masterstones D and E are approximately the same, although the x,y value of masterstone D should be lower. Masterstones F and G are too close to each other, and their x,y values ​​should be more different.

[0011] In the third test 16, the x and y values ​​of master stones D and E are almost the same, but they should be different. The x and y values ​​of master stones L and M are also almost the same, but they should be different.

[0012] Although the same 10 master stones were evaluated in Trials 12, 14, and 16, there were deviations in each trial. The x and y values ​​of most of the master stones in Trial 12 were shifted downward and to the left relative to the x and y values ​​in Trial 14. The x and y values ​​in Trial 16 were shifted upward (higher y values) and to the right (higher x values) relative to Trials 12 and 14.

[0013] The shifts in x and y are small, about 0.001 for x and 0.002 for y, or about 0.3%. However, the distance between the color grades of the master stones is smaller, about 0.0004 for x and 0.0003 for y. A more accurate system is needed to evaluate sample diamonds to within 0.5 color grades. While shifts in x and y can be tolerated from test to test, the x and y values ​​of the master stones should rise steadily for successive master stones. There should be no regression.

[0014] Recently, automated diamond color grading machines have begun to appear. Some machines use a manual rotating platform that is slow and difficult to use. These machines can be expensive or have poor accuracy.

[0015] It is desirable to automatically color grade diamonds to within + / -0.5. An automated process and machine is needed to assess the color grade of sample diamonds. For automated diamond color grading, it would be desirable to have a machine that continuously feeds sample diamonds for comparison with known master stones.

Brief Description of the Drawings

[0016] Figure 1 shows the color grades of diamonds.

[0017] FIG. 2 shows a prior art series of diamond master stones evaluated in the CIE 1931 color space chromaticity diagram.

[0018] Figure 3 is a diagram of a diamond color grading machine.

[0019] Figure 4A-4C The rotating plate is shown in more detail.

[0020] Figure 5 The target area on the diamond image is shown.

[0021] Figure 6A-6D Shown is the conveyor system for loading and unloading sample diamonds onto a rotating plate.

[0022] Figure 7A-7E A side view showing the conveyor system loading and unloading sample diamonds on a rotating plate.

[0023] Figure 8 This is a chart that uses a diamond color grader to analyze the chromaticity and x, y color values ​​of contrasting stones.

[0024] Figure 9A-9D It is a flow chart of the operation of the diamond color grading machine.

[0025] Fig.10 It is a diamond color grader with input and output conveyors located on the same side.

[0026] Fig.11 is a side view of a diamond color grader with the input and output conveyors on the same side but at different heights.

[0027] Figure 12A-12B is a side view of another diamond color grader with a circular output sorter.

[0028] Fig.13 yes Figure 12A-12BA top view of another diamond color grader with a ring-shaped output sorter.

[0029] Fig.14 An alternative mounting arrangement on a rotating plate is shown.

[0030] Fig.15 It is another type of diamond color grading machine where the camera is located just above the rotating plate of diamonds. [Specific implementation method]

[0031] The present invention relates to improvements in diamond color grading. The following description is intended to enable one of ordinary skill in the art to make and use the invention in the context of a particular application and its requirements. Various modifications to the preferred embodiments will be apparent to one skilled in the art, and the general principles defined herein may also be applied to other embodiments. Therefore, the present invention is not intended to be limited to the specific embodiments shown and described, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0032] Figure 3 Schematic diagram of a diamond color grading machine. A batch of sample diamonds can be input into the diamond color grading machine, which has a conveyor or robot arm system (not shown) for continuously feeding the sample diamonds and placing the sample diamonds on a rotating plate 30 for comparison with a set of master stones 10 also placed on the rotating plate 30. The stepper motor 24 rotates the rotating plate 30 so that new sample diamonds can be placed on the rotating plate 30 and sample diamonds that have completed the evaluation can be removed.

[0033] The light source 42 below the rotating plate 30 emits uniform white light to illuminate the master stone 10 and the sample diamond on the rotating plate 30. The housing 46 can be a sphere with a flat white surface inside to scatter the light from the light source 42 into the chamber 20 and provide a matte white background. The top surface of the rotating plate 30 can also be a matte white surface.

[0034] The camera 40 is mounted on a bracket 44 on a housing 46. The opening in the housing 46 enables the lens of the camera 40 to capture images of light reflected from the sample diamond and the master stone 10 on the rotating plate 30. The controller or computer 50 can cause the camera 40 to capture images after the computer 50 causes the stepper motor 24 to rotate the rotating plate 30 a desired amount. The images captured by the camera 40 can be analyzed by the computer 50 or sent to another computer or server for further processing.

[0035] Figure 4A-4C The rotating plate is shown in more detail. Figure 4AIn the embodiment, the rotating plate 30 is empty and no diamonds are loaded thereon. A plurality of mounting seats 32 are arranged around the rotating plate 30. Each mounting seat 32 may be a groove in the rotating plate 30 into which a diamond may be placed by a conveyor or a robot arm. The mounting seat 32 may also be a raised mounting seat or a fixing clip, which is retained on the surface of the rotating plate 30 and which can fix the diamond in place when the rotating plate 30 rotates.

[0036] exist Figure 4B In the example, the color stone 10 has been placed in the mounting seat 32. However, there is still one mounting seat 32 that is vacant. In this example, 11 color stones 10 are placed on the rotating plate 30.

[0037] exist Figure 4C In the embodiment, the sample diamond 18 is placed in an empty mount 32. All mounts 32 are occupied by sample diamonds 18 or master stones 10. Each master stone 10 has a different color grade.

[0038] For a set of 11 master stones 10 and one sample diamond, a total of 12 mounts 32 are placed on the rotating plate 30, each of which is spaced apart at a radial angle of 30 degrees (360 / 12). The computer 50 instructs the stepper motor 24 to rotate the rotating plate 30 30 degrees, and then instructs the camera 40 to take a picture of the rotating plate 30 and the sample diamond and master stone 10 thereon. The computer 50 then instructs the stepper motor 24 to rotate another 30 degrees and take another picture. This operation is repeated, and a total of 12 pictures are taken, rotating 12 times, for a total of 360 degrees.

[0039] Figure 5 The target area on the diamond image is displayed. Each time the stepper motor 24 rotates one circle, the camera 40 takes a picture of all the master stones 10 and sample diamonds 18 on the rotating plate 30. The computer 50 detects the part of the picture containing diamonds and divides the picture into 12 images, each containing a diamond. The computer 50 can then further process each single diamond image to select the target area 48 on the master stone 10 or sample diamond 18.

[0040] The target area 48 may be within a single facet or side of the diamond, may include multiple faces, and may include one or more edges of the diamond. The computer 50 may select the target area 48 as an area of ​​consistent color without abrupt color changes, or may simply define the boundaries of the diamond and select the target area 48 as the area within the defined boundaries.

[0041] Figure 6A-6D The conveyor system for loading and unloading sample diamonds on the rotating plate is shown. Fig. 6AIn the embodiment of the present invention, the rotating plate 30 has been pre-installed with 11 master stones 10 and has an empty mount 32. The computer 50 has instructed the stepper motor 24 (not shown) to rotate the rotating plate 30 so that the empty mount 32 is aligned with the input conveyor 60. The sample diamond 18 is moved by the input conveyor 60 until it reaches the end of the input conveyor 60, at which time the sample diamond 18 falls into the mount 32. When the sample diamond 18 falls from the input conveyor 60, the mount 32 can be in a reduced shape to guide the sample diamond 18 into position.

[0042] exist Figure 6B In the embodiment, the rotating plate 30 has been rotated three times, a total of 90 degrees, so that the sample diamond 18 is located at Figure 6B The camera 40 takes three pictures, one picture for each rotation step of the rotating plate 30.

[0043] exist Figure 6C In the embodiment, the rotating plate 30 has been rotated six times, a total of 180 degrees, so that the sample diamond 18 is located at Figure 6C The camera 40 has taken six pictures, and the rotating plate 30 takes one picture every time it rotates one step.

[0044] Then, the computer 50 continues to step the rotating plate 30 until the sample diamond 18 returns to the initial position under the input conveyor belt 60. A total of 12 pictures are taken for the 12 steps of the rotating plate 30. Finally, the computer 50 causes the stepper motor 24 to rotate the rotating plate 30 180 degrees so that the sample diamond 18 is placed under the output conveyor belt 62.

[0045] exist Fig.6D In the embodiment, the sample diamond 18 has been removed from the mount 32 and placed on the output conveyor 62. A robot arm (not shown) can pick up the sample diamond 18 from the rotating plate 30 and place the sample diamond 18 on the output conveyor 62.

[0046] The output sorter 64 has a plurality of bins 66, each of which holds sample diamonds 18 of different color grades. When the computer 50 determines the color grade of the sample diamond 18, the computer 50 commands the output sorter 64 to move the bins 66 so that a target bin 68 of the determined color grade is located below the end of the output conveyor 62. Then, the sample diamond 18 falls into the target bin 68.

[0047] 7A to 7E A side view of the conveyor system loading and unloading diamond samples onto and from the rotating plate is shown. Fig. 7AIn the embodiment of the present invention, the rotating plate 30 has been pre-loaded with the master stone 10 and has an empty mount 32. The computer 50 has instructed the stepper motor 24 (not shown) to rotate the rotating plate 30 so that the empty mount 32 is aligned with the input conveyor 60. The sample diamond 18 moves through the input conveyor 60 until it reaches the end of the input conveyor 60, at which time the sample diamond 18 falls into the mount 32. As the sample diamond 18 falls from the input conveyor 60, the mount 32 can be reduced to guide the sample diamond 18 into place.

[0048] exist Figure 7B In the embodiment of the present invention, the input conveyor 60 has fed the sample diamond 18 into the mount 32. The rotating plate 30 is now fully loaded with the master stones 10 and the sample diamond 18. Each time the stepper motor 24 moves radially one step, the camera 40 takes a picture, and a total of Nt rotations and Nt pictures are taken. Nt = Ns + Nr, where Ns is the number of sample diamonds 18 on the rotating plate 30, and Nr is the number of master stones 10 on the rotating plate 30.

[0049] exist Figure 7C In the embodiment, after the stepper motor 24 rotates radially Nt steps, the computer 50 determines the color grade of the sample diamond 18. The computer 50 instructs the output sorter 64 to move the bin 66 so that the target bin 68 of the color grade determined by the computer 50 is aligned with the end of the output conveyor 62.

[0050] The computer 50 also activates the stepper motor 24 to rotate the rotating plate 30 180 degrees so that the sample diamond 18 is aligned with the output conveyor 62. Then, the computer 50 instructs the robot arm 72 to pick up the sample diamond 18 from the mounting seat 32 on the rotating plate 30. Fig.7D In the embodiment, the computer 50 instructs the robot arm 72 to move the sample diamond 18 to the output conveyor belt 62 and places the sample diamond 18 on the output conveyor belt 62.

[0051] exist Fig. 7E In the embodiment of the present invention, the sample diamond 18 reaches the end of the output conveyor 62 and falls into the target bin 68, which is moved by the output sorter 64 to a position below the end of the output conveyor 62. The target bin 68 may contain other sample diamonds 18 having the same color grade (e.g., G grade).

[0052] Figure 8 It is a chart for analyzing the chromaticity and x,y color values ​​of diamonds using a diamond color grader. During each evaluation run or test, the rotating plate 30 rotates Nt times and the camera 40 takes Nt pictures.

[0053] The order of the 10 master stones DM was correct in each trial, although there was a shift in runs 22 and 26 compared to run 24. Although masters D and E were closer than the other master stone FM, master stone D still had lower x,y values ​​than master stone E. Sample 25 was correctly placed between masters E and F in all three runs 22, 24, and 26, so sample 25 was correctly evaluated as color grade E. Regenerating the x,y values ​​of the master stones for each evaluation run of the sample corrected for possible shifts between runs 22, 24, and 26 that could occur with changes in temperature and other operating conditions.

[0054] Evaluating the color of each master stone at the same time as the sample stone's color serves as an internal reference to correct for color shifts under operating conditions. Therefore, calibration is not required before each evaluation run. The relative order of each master stone in the D to M range remains accurate. The distance and slope between adjacent grades across the baseline for the master stones in the D to M range are repeatable after evaluation runs on different dates. For sample diamond color evaluations, the sample diamond's color grade is consistently within 0.5 after each evaluation run.

[0055] Figure 9A-9D A flow chart showing the operation of a diamond color grader. Fig. 9A In step 202, a total of Nr master stones 10 are placed on the mount 32 or bracket 73 on the rotating plate 30. During the continuous line test, many sample diamonds 18 are placed on and removed from the rotating plate 30, and these master stones 10 remain on the rotating plate 30.

[0056] In step 204, the diamond under test (DUT), i.e., the sample diamond 18, is placed on the rotating plate 30. The input conveyor 60 can move the sample diamond 18 near the rotating plate 30, and then the robot arm 72 picks up the sample diamond 18 and places it in the mounting seat 32 or the bracket 73 on the rotating plate 30.

[0057] In step 206, parameter P is set to the total number Nm of mounts 32 on the rotating plate 30, which may be greater than Nt when some mounts 32 are empty and have no gemstones. Nt is the sum of the number of sample diamonds 18 (Ns) and the number of master stones 10 (Nr).

[0058] At step 208, the camera 40 takes a picture of the initial rotational position of the rotating plate 30 and all of its diamonds. At step 210, the picture is stored for later access and processing by the computer. Then, at step 212, the stepper motor 24 is activated to rotate the rotating plate 30 360 / Nm degrees. This rotation will place the diamonds in the same position as in the previous picture (step 208), but because all diamonds are offset by one position in the direction of rotation, the diamonds in each position are different. At step 214, the parameter P is decremented to keep track of the number of partial rotations.

[0059] At step 216, the photograph of all Nt diamonds on the rotating plate 30 in step 208 is divided into Nt smaller images. Each of these small images generated in step 216 contains only one diamond, while the large photograph in step 208 contains Nt diamonds. Image processing software (such as an object detection program) running on the computer can detect the outline of each diamond against the matte white background of the rotating plate 30 and the housing 46, and use the detected outlines to cut the photograph into smaller images.

[0060] At step 218, while parameter P has not yet reached zero, the camera 40 takes another picture of all Nt diamonds, but rotated to the new position, while each diamond is separated into individual smaller images at step 216. This cycle is repeated Nm times, so each diamond is photographed at Nm positions relative to the camera 40. By rotating the sample diamond 18 through all Nm positions through this cycle, any slight changes in the illumination at each position within the chamber 20 are applied to all Nt diamonds.

[0061] exist Fig. 9B In the example, after taking all Nm photos and dividing each photo into Nt small images, there are Nm*Nt small images. For each of these Nm*Nt images (step 220), a target area 48 is selected in the small image (step 222). The target area 48 can be selected as an area within the diamond, or even an area within a face of the diamond, or an area with uniform color within the diamond outline. Areas in the diamond outline that are saturated, pure black, or have sharp, sudden changes in color can be deleted from the target area 48 because these areas may be reflections or image distortions.

[0062] At step 224, the red, green, and blue (RGB) pixel values ​​of all pixels within the target area 48 are obtained from the small image. Then, at step 226, the computer converts these RGB values ​​into XYZ values ​​in the CIE XYZ tristimulus color space. Then, at step 230, the computer further calculates these XYZ values ​​into xy values ​​in the CIE chromaticity space.

[0063] Then, at step 232, the computer generates the average value x and the average value y of all xy pixel values ​​for all pixels in the target area 48 in the current small image. At step 234, the average value x and the average value y are stored for the small image.

[0064] At step 236, the image processing loop (steps 222-234) is repeated for all small images, for a total of Nm*Nt loops. Once all Nm*Nt small images are processed (step 236), the average values ​​of the Nm*Nt x, y values ​​are stored.

[0065] exist Fig. 9C In the embodiment, by tracking the rotation of the stepper motor 24 and the placement of the sample diamond 18, the computer tracks that a small image corresponds to a certain master stone 10 and the sample diamond 18. For each diamond (step 250), for example, for the sample diamond 18 or for each master stone 10, each diamond has Nt small images and Nt average values ​​of x and y. In step 252, the Nt average values ​​x and y of the diamond are read, and in step 254, the Nt average values ​​x are averaged to obtain the color average value x of the diamond, and the Nt average values ​​y are averaged to obtain the color average value y of the diamond.

[0066] when Fig. 9C At the end, the sample diamond 18 has a color value based on the average value x, y, the master stone 10 of the D-grade diamond has a color value based on the average value x, y, the master stone 10 of the E-grade diamond has a color value based on the average value x, y, and so on.

[0067] exist Fig.9D In step 248, Nr master stones 10 of Nr diamond grades are arranged in grade order according to Nr average values ​​x, y and stored as a reference database. In step 240, the average values ​​x, y of the sample diamond 18 are compared with the average values ​​x, y in the database to find the closest lighter color (left side) master stone. For example, the color grade value of the sample diamond 18 or the master stone 10 can be calculated as the projected position of a line that is fitted by the average values ​​x, y of the master stone 10. In step 242, the master stone grade with the largest value and less than the sample diamond 18 is selected as the closest lighter color reference.

[0068] The left reference is effectively "rounded" to a better color grade, making the color of the closest master stone lighter than the color of the sample gem. The computer sends the color grade determined in step 242 to the output sorter 64, which moves the bins to align the target bin 68 for that grade with the output conveyor 62. In step 244, the robot arm 72 picks up the sample diamond 18 from the rotating plate 30 and places the sample diamond 18 into the target bin 68 on the output conveyor 62. The computer may need to activate the stepper motor 24 to align the sample diamond 18 with the output conveyor 62, and then align the empty tray 73 with the input conveyor 60 again to place the next DUT (steps 246, 204).

[0069] Fig.10 is a diamond color grader with input and output conveyors on the same side. In this alternative, both the input conveyor 60 and the output conveyor 62 are located on the right side of the rotating plate 30. The input conveyor 60 and the output conveyor 62 are located adjacent to each other above the rotating plate 30, or the input conveyor 60 can be located above or below the output conveyor 62. The output conveyor 62 can end above the target bin 68 on the output sorter 64, and the input conveyor 60 can extend above the output sorter 64 to enter the bin (not shown) of the input sample diamonds 18 located outside the output sorter 64. A single robotic arm 72 can move the sample diamonds 18 between the rotating plate 30, the input conveyor 60, and the output conveyor 62.

[0070] Fig.11 is a side view of a diamond color grader with the input and output conveyors on the same side but at different heights. In this alternative, the input conveyor 60 and the output conveyor 62 are both located to the right of the rotating plate 30. The input conveyor 60 and the output conveyor 62 are located adjacent to each other above the rotating plate 30, with the input conveyor 60 located above the output conveyor 62. The output conveyor 62 may end above a destination bin 68 on the output sorter 64, while the input conveyor 60 may extend above the output sorter 64 to enter a bin (not shown) of input sample diamonds 18 located outside the output sorter 64. A single robotic arm 72 may move the sample diamonds 18 between the rotating plate 30, the input conveyor 60, and the output conveyor 62.

[0071] Figure 12A-12B A side view of another diamond color grader with a ring-shaped output sorter is shown. Fig. 12A In the process, the input conveyor 60 moves the sample diamond 18 ′ and places the sample diamond 18 into the empty mount 32 on the rotating plate 30 .

[0072] exist Fig. 12BIn the embodiment, after the evaluation is completed, the robot arm 72 picks up the sample diamond 18 from the rotating plate 30 and places the sample diamond 18 into the target bin 68. The robot arm 72 can gently place the sample diamond 18 into the target bin 68 instead of letting the sample diamond 18 fall into the target bin 68.

[0073] Fig.13 Shows the Figure 12A-12B FIG. 5 is a top view of another diamond color grader with an annular output sorter. The output sorter 64' forms an annular ring around the rotating plate 30, and the bins 66 are arranged on the periphery of the rotating plate 30. The computer 50 instructs the output sorter 64' to rotate relative to the fixed input conveyor 60 and output conveyor 62 to rotate the target bin 68 below the end of the output conveyor 62.

[0074] Fig.14 Another mounting method on the rotating plate is shown. The robot arm 72 picks up the sample diamond 18' from the input conveyor belt 60 and places the sample diamond 18 into the bracket 73. The bracket 73 is a V-shaped support device mounted on the upper surface of the rotating plate 30. The bracket 73 holds the sample diamond 18 or the color master stone 10 on the upper surface of the rotating plate 30 at a standard design angle to the upper surface of the rotating plate 30. When the camera 40 is mounted directly above, as shown in FIG. Fig.15 As shown, the bracket 73 enables the viewing angle to be exactly 45 degrees to meet the GIA viewing standards through the camera 40. The viewing angle is the angle between the camera 40 and the gemstone axis, which is the angle passing through the culet of the gemstone and perpendicular to the table facet (top surface) of the sample diamond 18 or the master stone 10. In addition, when the robot arm 72 places the sample diamond 18 into the bracket 73, the slope of the groove in the bracket 73 helps guide the sample diamond 18 into position in the bracket 73.

[0075] Fig.15 Another type of diamond color grader where the camera is located directly above the diamond rotating plate. When using the V-shaped bracket 73 ( Fig.14 ), the master stone 10 and the sample diamond 18 are at an angle to the rotating plate 30. In order to achieve a viewing angle between the camera 40 and the gemstone axis (the axis passes through the cusp of the sample diamond 18 or the master stone 10 in the bracket 73 and is perpendicular to the table (top surface)), the camera 40 and the bracket 44 are moved to the opening at the top of the housing 46. The light source 42 can also be placed under the rotating plate 30 to provide appropriate lighting for the chamber 20. There can also be other light sources in the housing 46 to provide optimal lighting. [Alternative Embodiments]

[0076] The inventors have also contemplated several other embodiments. For example, various combinations and variations of the apparatus and method are possible. In FIG. 9 , the color conversion, target region 48 detection, and pixel averaging can be performed after all Nm photos are taken, or after each photo is taken. All RGB pixels in the target region can be averaged together to obtain an average RGB value for the target region, and then Fig. 9B In steps 226, 230, and 232, the average RGB value is converted to an average XYZ value and then to an average x,y value. It is also possible to average all pixels after the conversion to XYZ and before the conversion to x,y. It is also possible to perform only one RGB to xy conversion, rather than first converting RGB to XYZ and then to xy. It is also possible to use a three-dimensional R, G, B map instead of a two-dimensional x, y map for color grading without converting to x, y values, but this will reduce accuracy and increase the computational workload.

[0077] The rotating plate 30 can be circular or have other shapes while still providing clearance for rotation within the housing 46. The housing 46 can be a sphere or have other shapes. A sphere is believed to better integrate light and provide a flat background illumination, but other shapes of the housing 46 can provide sufficiently uniform illumination even if they are not as ideal.

[0078] There may be other light sources in addition to light source 42 within housing 46 to provide optimal illumination. The location and number of light sources may be modified, and the intensity and color profile of the light sources may be adjusted, for example under program control. The program may adjust the white balance or color temperature of light source 42 within a certain range, for example 5500-6500K. Testing may be performed to find the optimal settings for light source intensity and color temperature to produce the best background light uniformity within chamber 20 to achieve color grading accuracy.

[0079] An opening or hole may be provided in the side of the housing 46 to allow the input conveyor belt 60 to enter the chamber 20. Alternatively, the input conveyor belt 60 may terminate outside the housing 46, and a robot arm may then pick up a sample diamond 18 from the input conveyor belt 60 outside the housing 46, pass the sample diamond 18 through the hole into the chamber 20, and then place the sample diamond 18 on the mount 32 on the rotating plate 30. This allows the hole in the housing 46 to be smaller, reducing the non-uniformity of the background light within the chamber 20. Fig.10 , 11The alternatives of the embodiment of the present invention may provide better lighting effects because their input conveyor 60 and output conveyor 62 are on the same side. In this way, a single robot 72 can use a hole in the housing 46, without the need to set up a second hole for the output conveyor 62. This hole in the housing 46 can have a door or curtain that closes after the robot 72 loads the sample diamond 18 onto the rotating plate 30. Such a door can provide more uniform lighting within the chamber 20 than leaving a hole in the housing 46. The door can be painted the same matte white as the interior of the housing 46.

[0080] Although an output sorter 64 has been described with a bin 66 and a destination bin 68, the bins may be sacks, bags, or other collection devices with flexible sides. The output sorter 64 may move the bin 66 in various ways other than linear or rotational to select the destination bin 68. The output sorter 64, input conveyor 60, output conveyor 62, and robotic arm 72 may have many variations and physical configurations.

[0081] The robotic arm 72 may have a clamp to grab the sample diamond 18, or may have a small tube with a vacuum to hold the sample diamond 18 during movement. The robotic arm 72 may be rotated in various ways to better align the sample diamond 18 with the mount 32. When the range of motion and stroke length of the robotic arm 72 are large, precise alignment with the sample diamond 18 on the rotating plate 30 is not necessary. The robotic arm 72 can pick up diamonds from multiple mounts on the rotating plate 30, thereby avoiding rotation for alignment. The input conveyor 60 can be used and the robotic arm can be used to place the sample diamond directly into the target bin, thereby eliminating the output conveyor 62.

[0082] The mount 32 can have various shapes and features to better hold and align the sample diamond 18, and can hold the sample diamond 18 at a certain angle to the plane of the rotating plate 30. When the mount 32 has a built-in angle, the position of the camera 40 can be changed from Figure 3 The 45 degree position shown is adjusted so that the viewing angle of the camera 40 is 45 degrees to the axis of the sample diamond 18 and the master stone 10 on the rotating plate 30. Other viewing angles may be used instead of 45 degrees.

[0083] The interior of the housing 46, the surface of the rotating plate 30 and the mounting base 32 or bracket 73 can be painted the same matte white, or can be made of matte white plastic or other materials. There may be some unevenness, such as screw heads, gaps or seams, but it is best to keep these to a minimum to improve light uniformity within the chamber 20.

[0084] Various types of mounts 32 or brackets 73 may be used. Ideally, the same mount 32 or bracket 73 is used for both the sample diamond 18 and the master stone 10, but the sample diamond 18 may be mounted on a bracket 73 that is different from the mount 32 of the master stone 10, but the mount 32 and bracket 73 should not introduce non-uniformities in the image captured by the camera 40, or these non-uniformities should be able to be removed when the target area 48 is selected from the large photo. The target area 48 may be a combination of multiple areas within the diamond, which are included in the image processing and averaging of the diamond. The target area 48 may be non-contiguous, rather than a separate area bounded by a polygon. The mount 32 or bracket 73 may be a support device, or it may be integral to the rotating plate 30.

[0085] The master stones can all be of the same carat weight, preferably the same carat weight as the sample diamond. The cut and clarity of all the master stones in the master stone set can also be the same or similar. The closer the match between the master stone 10 and the sample diamond 18, the better the color grading. However, the automatic color grading system can also work adequately when the match is not ideal, although some accuracy may be lost. The master stones 10 can be placed on the rotating plate 30 in the order of their known color grades, or they can be placed in a random order. However, the program can be instructed when to place the different grade diamonds on the rotating plate 30 so that the program can keep track of the grades of these diamonds. Alternatively, the program can also determine the grade based on the x, y values ​​of the master stones and the order of the x, y values.

[0086] Another method of unloading the diamond sample 18 onto the output conveyor 62 is to install an ejection mechanism on the mounting seat 32 to eject the sample diamond 18 from the mounting seat 32 onto the output conveyor 62. Another method is to install a valve below the mounting seat 32 on the rotating plate 30, and opening the valve allows the sample diamond 18 to fall through the valve on the rotating plate 30 onto the output conveyor 62 or into the target bin 68 located below the rotating plate 30.

[0087] The positions of the mounts 32 or brackets 73 on the upper surface of the rotating plate 30 can be separated by equal angles and the distances from the center of the rotating plate 30 can also be equal. Therefore, when the rotating plate 30 rotates 360 / Nm degrees (where Nm is the number of mounts 32 or brackets 73), the camera 40 will see different master stones 10 or sample diamonds 18 in the same position as the previous diamond. Another method is to use unequal angles, or some diamonds are farther from the center of the rotating plate 30 than other diamonds, and the program can map these changes. The mounts 32 can be non-circular or arc-shaped along a circle. When Nm>Nt, empty mounts without gems can appear. For example, there can be 6 mounts on the rotating plate 30, but only 3 gems. For best results, each gem has the same lighting conditions at each position as much as possible. For example, if there are 6 mounts and the last 3 mounts are empty, gem 1 can be photographed at photo numbers 1, 2, and 3. But gem 2 can be photographed at photo numbers 2, 3, and 6, and gem 3 can be photographed at photo numbers 1, 5, and 6. In this example, although there are only 3 gemstones (Nt=3), the rotating plate 30 can be rotated Nm=6 times. The gemstones can be placed in any empty position on the rotating plate 30.

[0088] Although one sample diamond 18 is shown on the rotating plate 30, there may be more than one sample diamond 18, such as Ns. In addition, the number Nr of master stones 10 may vary, and the total number of diamonds Nt=Ns+Nr may also vary. There may also be empty seats 32 or brackets 73 on the rotating plate 30 that are not occupied by diamonds. The program can filter out these unoccupied mounting seat positions and not include them. When the stepper motor 24 is started, the program can track each master stone 10 and its current position on the rotating plate 30, and once all Nt positions of the rotating plate 30 are captured, the same master stone 10 is averaged at these positions.

[0089] Although the CIE XYZ color space has been described, other color spaces may also be used. For example, the camera 40 may capture pixels using RGB, and then the processor or computer converts RGB to XYZ. XYZ is superior to RGB because Y is brightness or intensity and does not have color information, so in the XYZ color space, color is represented by only two variables, X, Y, instead of three variables, R, G, B. Using 2 variables makes it easier to plot and compare color values. Removing the Z brightness component helps to mask changes in the brightness of the lighting within the chamber 20 and provides more accurate color grading. The computer can convert between multiple color spaces, such as XYZ and xyZ. Other color spaces, such as CIE Lab, hue saturation brightness (HSL), or hue saturation value (HSV), may also be used. Various filters and image processing may also be performed.

[0090] When the camera 40 outputs RGB, use Fig. 9B When the camera 40 can output XYZ, steps 224 and 226 can be skipped or modified.

[0091] Although a camera 40 has been shown, various image sensors may be used, and various lens systems may be used to transmit light to the image sensor. More than one camera 40 may be mounted on the housing 46, so that multiple photos can be taken simultaneously to observe the rotating plate 30 at different angles. The angle of the camera 40 relative to the rotating plate 30 may be adjusted, and various mounting brackets or supports 44 may be used instead.

[0092] Although it has been described that the camera 40 takes a picture every 360 / Nm degrees of rotation of the rotating plate 30, more images can be taken at smaller rotation increments. For example, a picture can be taken every 360 / (2*Nm) degrees, and the average number of images is doubled. This can improve accuracy. Alternatively, the camera 40 can be a video camera that can take a series of continuous frames, and the computer 50 can select the frames according to the desired rotation interval.

[0093] There may be other mounts 32 that do not have diamonds placed in them, and these empty mounts can be ignored by the image processing of the computer 50. When Ns is 2 or greater, the computer 50 can color grade 2 or more sample diamonds during one 360 ​​degree rotation, thereby increasing throughput.

[0094] Although diamonds have been described, other gemstones may be substituted. Although a circular or annular rotating plate 30 has been described, the rotating plate 30 may be elliptical, octagonal, square, or other shapes, but still leave sufficient rotational clearance in the chamber 20. The computer 50 may be a personal computer (PC), a controller, or a processor with a memory (such as a semiconductor main memory), with a large capacity memory (such as a flash drive or a hard drive), and with an I / O interface (such as an interface with a network (such as the Internet)). The computer 50 may offload some processing tasks to an external server, such as specialized image processing tasks. Neural networks may be used for certain processing tasks. The steps in FIG9 may be implemented by a computer executing machine-readable instructions that control the stepper motor 24 and the robotic arm, conveyor belt, and camera.

[0095] Instead of storing various values, we can calculate them. For example, once all Nm*Nt small images ( Fig. 9B In step 236), Nm*Nt average values ​​x and y are stored. Alternatively, the original Nm*Nt x and y values ​​can be stored instead of the average values ​​of x and y. Fig. 9C When necessary, calculate Nt average values ​​x and y.

[0096] The input conveyor 60, robotic arm 72, rotating plate 30 and output conveyor 62 can be used to inspect a continuous line of diamonds as sample diamonds 18. When the bin 66 on the output sorter 64 is full, the machine can be paused to replace the full bin with an empty bin or to remove the sorted diamonds from the output sorter 66. Untested diamonds can be poured into the input bin to deliver the diamonds to the input conveyor 60. Of course, the continuous line of diamonds can be paused for various reasons, such as loading, unloading, error handling, maintenance or work stoppage at the end of the day. Many color graders can be operated manually, without the need for skilled, trained experts to look at each diamond.

[0097] Terms such as upward, downward, above, below, horizontal, vertical, inside, outside, etc. are relative and dependent on the viewing angle and are not meant to limit the invention to a particular viewing angle. The device can be rotated so that vertical becomes horizontal and horizontal becomes vertical, so the terms are dependent on the viewer.

[0098] The background section of the present invention may contain background information about the problem or environment of the present invention, rather than describing the prior art of others. Therefore, the materials contained in the background section are not an admission by the applicant that the prior art is available.

[0099] Any method or process described herein is machine-implemented or computer-implemented, intended to be performed by a machine, computer or other device, and is not intended to be performed solely by humans without machine assistance. Tangible results produced may include reports or other machine-generated displays on display devices such as computer monitors, projection devices, audio generating devices, and related media devices, and may include hard copy printouts that are also machine-generated. Computer control of other machines is another tangible result.

[0100] Any advantages and benefits described do not necessarily apply to all embodiments of the invention. When the word "means" appears in a claim element, applicant intends that the claim element fall within 35 USC § 112(6). Typically, the word "means" is preceded by a label of one or more words. The one or more words preceding the word "means" are a label for the purpose of facilitating reference to the claim element and are not intended to express a structural limitation. Such a means-plus-function claim is intended to cover not only the structures described herein for performing the function and their structural equivalents, but also equivalent structures. For example, although a nail and a screw have different constructions, they are equivalent structures because they both perform the function of fastening. Claims that do not use the word "means" do not fall within 35 USC § 112(6). Signals are typically electronic signals, but can also be optical signals, such as those that can be transmitted over fiber optic lines.

[0101] The above description of the embodiments of the present invention is presented for the purpose of illustration and description. It is not intended to be exhaustive, nor is it intended to limit the invention to the precise form disclosed. Many modifications and variations are possible in light of the above teachings. It is intended that the scope of the present invention is not limited by this detailed description, but rather by the limitations of the claims appended hereto.

Claims

1. An automatic gemstone color grading machine, comprising: a rotating plate having mountings on its surface, each mounting being used to support a gemstone; wherein the gemstones on the mounting include a sample gemstone and a series of reference gemstones, wherein the series of reference gemstones have reference gemstones of different known color grades; The sample gemstone has an unknown color grade that is between at least two reference gemstones of known color grades; a stepper motor for rotating the rotating plate in one or more increments less than a full revolution; An image sensor for capturing an image of the gemstone mounted on the rotating plate; a light source for illuminating the gemstone on the rotating plate; a computer for instructing the stepper motor to rotate the rotating plate by one increment so that the sample gemstone rotates to a position previously occupied by one of the reference gemstones in the field of view of the image sensor; when the sample gemstone rotates sequentially to all positions previously occupied by the reference gemstones on the rotating plate, the computer causes the image sensor to capture an image after each rotation increment; Wherein, as the rotating plate increases in size, each colorimetric gemstone sequentially occupies the position of the sample gemstone and other colorimetric gemstones; The computer executes machine-readable instructions, causing the computer to select pixels for the sample gemstone and each reference gemstone from the image captured by the image sensor after each rotation increment, the computer averages the color index of these pixels for each gemstone at all incremental positions of the gemstone to generate a gemstone average color value for each gemstone, the computer selects a closest reference gemstone from the reference gemstones, the gemstone average color value of the closest reference gemstone having the smallest difference from the gemstone average color values ​​of the sample gemstone and the reference gemstone, and the computer outputs an indication of the closest known color grade of the gemstone as the color grade of the sample gemstone, Thus, the color grade of the sample gemstone is determined by the average of the color indices of the gemstone pixels captured by the image sensor at different rotation increments.

2. The automatic gemstone color grading machine according to claim 1, wherein the gemstone is a diamond; The negative difference between the average color value of the sample gemstone and the average color value of the gemstone closest to the gemstone is the smallest; The sample gemstone is darker in color than the closest gemstone.

3. The automatic gemstone color grading machine according to claim 2, further comprising: a housing that surrounds the rotating plate and enables the image sensor to view the gemstone on the rotating plate; The interior of the housing and the rotating plate are made of matte white material, providing a flat background with uniform color for the image captured by the image sensor.

4. The automatic gemstone color grading machine according to claim 3, wherein the housing is a sphere with openings, including an opening for the image sensor, and an opening for placing the sample gemstone on a vacant mount on the rotating plate and removing the sample gemstone from the mount on the rotating plate after the computer determines the color grade of the sample gemstone.

5. The automatic gemstone color grading machine according to claim 4, further comprising: A mechanical arm is used to pick up the sample gem from the sample gem input end and place the sample gem into an empty mounting seat on the rotating plate. The mechanical arm passes through an opening of the shell.

6. An automatic gemstone color grading machine according to claim 5, wherein after the computer determines the color grade of the sample gemstone, the robotic arm also passes through the housing to pick up the sample gemstone from the mounting seat on the rotating plate, and the robotic arm places the sample gemstone into an output bin, wherein the output bin is filled with sample gemstones of the same color grade determined by the computer.

7. The automatic gemstone color grading machine according to claim 6, further comprising: an output sorter having a plurality of output bins, each output bin being for sample gemstones of a different color grade, the output sorter being for receiving the color grade determined by the computer and moving a target bin having the color grade to an output position proximate to the robotic arm for receiving the sample gemstones from the robotic arm.

8. The automatic gemstone color grading machine of claim 6, further comprising: An input conveyor is used to move sample gemstones from a sample gemstone input bin with unknown color grade to an input position close to the robotic arm so that the robotic arm picks up the sample gemstones and places the sample gemstones on the mounting seat on the rotating plate.

9. The automatic gemstone color grading machine according to claim 3, wherein the image sensor is a camera, which is mounted on the housing at an angle of 45 degrees to the surface of the rotating plate; The light source has a fixed color temperature between 5500K and 6500K.

10. The automatic gemstone color grading machine according to claim 1, wherein the mount is a fixture that holds the gemstone on the surface of the rotating plate at a certain angle; The image sensor is mounted on the housing so that the axis of the gemstone forms an angle of 45 degrees with the image sensor.

11. The automatic gemstone color grading machine according to claim 3, wherein: The series of master color gemstones includes at least 10 gemstones having at least 10 different known color grades; The stepper motor rotates the rotating plate by at least 11 increments, and the image sensor captures images of all at least 11 increments.

12. The automatic gemstone color grading machine according to claim 3, wherein pixels in the image captured by the image sensor are converted into color coordinate values ​​x and y in an xy color space, wherein the computer has an image processor that identifies the boundaries of the gemstones in the image and selects pixels from a target area within the boundaries for averaging; For the sample gemstone and each reference gemstone, pixels of the gemstone in the target area in the images taken from all increments of the rotating plate are averaged.

13. The automatic gemstone color grading machine according to claim 12, wherein the computer selects the closest gemstone by generating differences in color coordinate average values ​​x and y, each difference being a difference between the average values ​​x and y of the sample gemstone and the average values ​​x and y of one of the reference gemstones, The closest gemstone is the gemstone that is closest in color and lighter than the sample gemstone.

14. The automatic gemstone color grading machine of claim 3, wherein each of the mounts on the rotating plate is at the same radial distance from the center of the rotating plate; wherein the mounting seats are all separated by equal distances, and each mounting seat is at an equal distance from the nearest adjacent mounting seat on the rotating plate; wherein the mounting seats are distributed along the circumference of the rotating plate; The camera captures images of all gemstones at all incremental positions of the rotating plate under the same lighting conditions.

15. A system for evaluating diamond color, comprising: a rotating plate, on whose surface there are mounts arranged along a circle, each mount being intended to accommodate a diamond; wherein Nr mountings are pre-mounted with master stones, the master stones being diamonds of known color grades, wherein the Nr master stones have Nr different color grades, wherein Nr is an integer of at least 3; wherein Nm is the total number of mounts on the rotating plate, wherein Nm is an integer of at least 4; wherein Ns mounts load and unload Ns sample diamonds of unknown color grade, wherein Ns is an integer of at least 1, wherein Nt is Ns+Nr; A camera for capturing an image of the diamond on the mounting seat on the rotating plate; a light source for illuminating the diamond on the rotating plate; a chamber that partially surrounds the rotating plate and provides a diffuse light background for the camera; an input loader for loading sample diamonds from an input diamond bin outside the chamber onto the Ns mounts on the rotating plate; an output loader for unloading sample diamonds from the Ns mounts on the rotating plate after evaluation; an output sorter for placing the sample diamonds from the output loader into sorting bins which are sorted according to the assessed color grade of the sample diamonds; A stepper motor, which is used to rotate the rotating plate; A color evaluation routine executed on a computer that controls the stepper motor, the input loader, the output loader, the output sorter, and the camera, the color evaluation routine comprising: instructing the input loader to load a sample diamond onto the rotating plate; (a) instructing the camera to take a picture of the rotating plate, the picture including all diamonds on the rotating plate; (b) instructing the stepper motor to rotate the rotating plate in one increment; Repeat steps (a) and (b) until the rotating plate is incremented Nm times; Split each of the Nm-1 photos into Nt images, each containing a diamond; For each image, selecting a target region on the image, the target region being located within the diamond boundary in the image; Reading pixel values ​​of pixels within the target area and converting each pixel value into color coordinate values ​​x and y; Calculate the average x and y color coordinates of the pixels in the target area of ​​each image; For the sample diamond and each of the master stones, calculate the average value of the gemstone color coordinate x as the average value of the color coordinate x values ​​of all images containing the specific gemstone, and calculate the average value of the gemstone color coordinate y as the average value of the color coordinate y values ​​of all images containing the specific gemstone; Generate a color grade value in x,y color space for the sample diamond and each of the master stones, wherein the color grade value is calculated as the projection position of the fitted line of the gemstone color average value x and the gemstone color average value y of the master stone; Finding a selected master stone having a maximum color grade value, wherein the color of the selected master stone is lighter than the color of the sample diamond; outputting an indication of the known color grade of the selected master stone as the estimated color grade of the sample diamond; instructing the output loader to unload the sample diamonds from the rotating plate, and instructing the output sorter to place the sample diamonds into a sorting bin of the sample diamonds having an assessed color grade, Sample diamonds are thereby loaded and unloaded onto the rotating plate and their color assessed by comparison with the coordinates of a master stone in x,y color space.

16. The system of claim 15, wherein the camera captures images having red, green, and blue (RGB) pixels; Where RGB pixels are converted to x and y color values.

17. The system of claim 15, further comprising: A robotic arm loads or unloads the sample diamond onto the rotating plate through an opening in the chamber.

18. The system of claim 17, wherein said output sorter comprises a rotating ring surrounding said chamber, the rotating ring having an output bin rotated into position.

19. The system of claim 15, wherein the x color value and the y color value of the master stone are regenerated each time a sample diamond is evaluated.

20. A method for evaluating the color of a sample diamond, comprising: placing master stones on a mount on the rotating plate, wherein the master stones are diamonds having known color grades, wherein at least seven of the master stones have at least seven different known color grades; placing a sample diamond on a mount on the rotating plate; (a) instructing a camera to take a picture of the rotating plate, the picture including all diamonds on the rotating plate; (b) instructing the stepper motor to rotate the rotating plate in one increment; Repeating steps (a) and (b) until the rotating plate is incremented at least 7 times; Split each photo into smaller images, each with a diamond; For each smaller image, selecting a target region on the smaller image, the target region being within the diamond boundary in the smaller image; Read the pixel values ​​of the pixels in the target area, and convert each pixel value into a color value x and a color value y; Calculate the color average x and color average y of the pixels in the target area of ​​each smaller image; For the sample diamond and each of the master stones, calculate the gemstone color average x as the average of the color averages x of all the smaller images containing the specific gemstone, and calculate the gemstone color average y as the average of the color averages y of all the smaller images containing the specific gemstone; Generate color grade values ​​in x,y color space for the sample diamond and each master stone, wherein the color grade values ​​are calculated as the projection positions of the fitted lines of the gemstone color average value x and the gemstone color average value y of the master stone; Finding a selected master stone having a maximum color grade value, wherein the color of the selected master stone is lighter than the color of the sample diamond; outputting an indication of the known color grade of the selected master stone as the estimated color grade of the sample diamond; instructing an output loader to remove the sample diamonds from the rotating plate and instructing an output sorter to place the sample diamonds into a sorting bin for which color grade has been assessed, Thereby, the sample diamond is loaded onto and unloaded from the rotating plate and the color is evaluated by comparison with the color value of a reference stone in the x,y color space.

Images