Methods for analyzing mineral species in ores

Abstract

To provide an ore mineral species analysis method that uses an optical microscope, and that can correct an influence caused by the condition for a light source.SOLUTION: An ore mineral species analysis method includes: a first step of photographing ore with an optical microscope to obtain an optical microscope image including image information in a pixel unit; a second step of binarizing the optical microscope image to extract a mineral particle area corresponding to mineral particles; and a third step of, for pixels to be analyzed in the mineral particle area, comparing the image information of the pixels with mineral identification areas set in advance to mineral species to identify the mineral species of the pixels. The image information used in the third step is any one of data of S and V and data of S and L. The mineral identification area is an area set in an S-V space or an S-L space. In performing the third step, a correction is made on the mineral identification area in at least a V direction or an L direction, and subsequently the mineral species is identified.SELECTED DRAWING: Figure 5

Description

Technical Field

[0001] The present invention relates to a method for analyzing mineral species of ores.

Background Art

[0002] Among the ores mined from mines, various minerals are included, from those with high usefulness to those with low usefulness. As an example of a method for analyzing mineral particles contained in such ores, a method for analyzing mineral species using an optical microscope has been conventionally known. In the method for analyzing mineral species using an optical microscope, generally, based on the information obtained by the optical microscope, mineral particles are analyzed by quantification by comparison with an index or quantification by the point counting method.

[0003] Also, as an example of a method for analyzing mineral particles contained in ores, MLA: Mineral Liberation Analyzer is known. MLA can automatically analyze minerals based on SEM-EDS.

[0004] For example, in Japanese Patent Application Laid-Open No. 2019-174473 (Patent Document 1), in non-ferrous metal smelting, by observing the slag obtained by melting the raw material with a digital microscope, a method for analyzing the slag capable of analyzing the concentration of matte grains present in the slag by a simpler method is disclosed.

[0005] Also, as a method for preparing an observation sample for MLA analysis, Japanese Patent Application Laid-Open No. 2016-050918 (Patent Document 2) discloses a method of preparing a sample by applying a predetermined embedding method to mineral powder particles, suppressing the bias in the state of existence of minerals caused by the difference in specific gravity of ore particles, and observing them.

Prior Art Documents

Patent Documents

[0006]

Patent Document 1

[0007] However, conventional methods of analyzing mineral species using optical microscopes have limitations in accuracy because the observation results of the target object vary depending on the light source conditions.

[0008] This invention was completed in view of the above-mentioned problems, and in one embodiment, aims to provide a method for analyzing the mineral species of ore using an optical microscope that can correct for the influence of light source conditions. [Means for solving the problem]

[0009] As a result of various diligent studies to solve the above problems, the inventors have found that the above problems can be solved by utilizing the high qualitative ability of optical microscopes to analyze minerals, using an optical microscope for mineral species analysis, programming identification by image analysis, and performing necessary corrections. Specifically, the invention utilizes the automatic control of a digital microscope to perform wide-field observation, analyzes the observed colors as images, and identifies the mineral species by comparing them with a pre-set mineral identification area. The present invention was completed based on the above findings and is illustrated below.

[0010] [1] A method for analyzing the mineral species of ore, The first step involves photographing the aforementioned ore with an optical microscope to obtain an optical microscope image containing image information at the pixel level, A second step involves binarizing the optical microscope image to extract the mineral particle region corresponding to the mineral particle, A third step involves comparing the image information of each pixel with a pre-set mineral identification region for each mineral species to identify which mineral species each pixel belongs to. Includes, The image information used in the third step is either the S and V data, or the S and L data. The aforementioned mineral identification region is a region set on the SV space or SL space, A method for analyzing the mineral species of ore, characterized in that, in the third step, correction is performed on the mineral identification region in at least the V direction or the L direction before identifying the mineral species. [2] Furthermore, prior to the third step, the process includes an additional step of creating distribution data of an SV plot or SL plot by plotting the S and V data or the S and L data for each pixel in the SV space or SL space for the pixels to be analyzed in the mineral particle region, comparing the peak position of the pixel density in the distribution data with the mineral determination region for each mineral species, and determining the mineral species included in the analysis target. The mineral determination region is a region set on the SV space or SL space. In the aforementioned additional step, the mineral determination region is corrected in at least the V direction or the L direction before determining the mineral species. The third step is characterized by performing mineral species identification only for mineral species that have been determined to be included in the analysis target in the additional step, as described in [1]. [3] The mineral identification region is set so as not to overlap between different mineral species, as described in [2], for the mineral species analysis method of ore. [4] The mineral identification region correction is performed only in the V direction or the L direction, and not in the S direction, as described in any one of items [1] to [3] for the mineral species analysis method of ore. [5] The correction of the mineral identification region is performed only in the V direction or the L direction, and not in the S direction, as described in any one of items [2] to [4] for the mineral species analysis method of ore. [6] Furthermore, it includes a step of calculating the ratio for each mineral species based on the area ratio of each identified mineral species and the known density of each mineral species, and the method for analyzing mineral species of ore according to any one of [1] to [5].

Effect of the Invention

[0011] According to an embodiment of the present invention, there can be provided a method for analyzing mineral species of ore that can correct the influence of the conditions of a light source in a method for analyzing mineral species using an optical microscope.

Brief Description of the Drawings

[0012] [Figure 1] In one embodiment of the present invention, it is a diagram for explaining a procedure for obtaining an optical microscope image of a mineral particle region. [Figure 2] In one embodiment of the present invention, it is a diagram showing an example of the S-V plot distribution of different mineral species. [Figure 3] In one embodiment of the present invention, it is a diagram showing a mineral determination region determined for an ore actually containing mineral particles of both Ccp and Py. [Figure 4] In one embodiment of the present invention, it is a diagram showing a mineral determination region and a mineral identification region determined for an ore actually containing mineral particles of both Ccp and Py. [Figure 5] In one embodiment of the present invention, it is a flowchart showing a procedure for analyzing an ore sample using a mineral identification region. [Figure 6] In one embodiment of the present invention, it is a flowchart showing a procedure for analyzing an ore sample using a mineral determination region and a mineral identification region. [Figure 7] In the S-V plot and the S-L plot, it is a diagram showing the dependencies of V and L on the brightness of the S value, respectively. [Figure 8] In an example of the present invention, it is a diagram showing a method for correcting a mineral determination region in the V direction. [Figure 9] In an example of the present invention, it is a diagram showing a method for correcting a mineral identification region in the V direction. [Figure 10] It is a diagram showing a comparison between an optical microscope image and mineral maps with and without correction in an embodiment of the present invention.

Embodiments for Carrying out the Invention

[0013] Next, embodiments for carrying out the present invention will be described in detail while referring to the drawings. The present invention is not limited to the following embodiments, and it should be understood that design changes, improvements, etc. can be appropriately added based on the ordinary knowledge of those skilled in the art without departing from the gist of the present invention.

[0014] (1. Ore) In the method for analyzing mineral species of ore according to an embodiment of the present invention, the mineral to be analyzed is not particularly limited as long as it is a mineral of ore that can be determined by color, and any mineral can be analyzed. In particular, copper ore is easy to identify by the color of mineral species, so the present invention can be preferably used for analyzing mineral species of copper ore. Examples of copper ore include various minerals contained in copper concentrate (for example, Chalcopyrite, Chalcocite, Covellite, Bornite, Pyrite, Molybdenite).

[0015] The mineral particles contained in the ore are preferably subjected to pretreatment to form a smooth observation surface for observation by the optical microscope image described later. The specific method of pretreatment is not limited, but for example, a method of embedding mineral particles in a resin material to produce a sample-embedded resin in which the mineral particles are fixed in the resin material can be mentioned.

[0016] To embed mineral particles in a resin material, one can use a method in which the mineral particles are placed in a container together with a liquid resin material, the mineral particles and liquid resin material in the container are stirred, and then the liquid resin material is cured. When stirring, it is preferable to rotate the container containing the mineral particles and liquid resin material in a rotation-orbit agitator while simultaneously revolving it in the opposite direction to the rotation. This effectively suppresses particle aggregation, effectively improves the dispersibility of granular samples in the resin material, and allows for obtaining a representative composition without separation or segregation.

[0017] The particle size of mineral particles ranges from, for example, 1 μm to 700 μm, typically from 20 μm to 200 μm, and is usually fairly evenly distributed and uniform. While particle size analyzers can measure particle sizes from, for example, 0.243 μm to 2000 μm, the particle size of mineral particles, as described above, is usually uniformly distributed within this range.

[0018] Various resin materials can be used to embed and fix mineral particles, as long as they can be kept in a liquid state when added to the container and stirred, and can then be cured. Examples include epoxy resins, acrylic resins, and phenolic resins, among which epoxy resins, being thermosetting resins, are preferred.

[0019] The ratio of mineral particles to the liquid resin material placed in the container is preferably 100% to 300% by volume. More preferably, it is 200% to 300% by volume. This means that if the ratio of mineral particles is 100% or more by volume, particle aggregation can be suppressed, and if the ratio of mineral particles is 300% or less by volume, the mineral particles can be sufficiently solidified, suppressing the phenomenon of breakage during surface polishing when the measurement surface is exposed.

[0020] In the above-described pretreatment method, mineral particles can be sufficiently dispersed in the resin material, so it is not necessary to add graphite or the like. Of course, graphite may be added. For further details of this pretreatment method, please refer to Japanese Patent Publication No. 7018804.

[0021] (2. Optical microscope image) In this embodiment, for a sample to be analyzed that contains mineral particles with multiple mineral species, first, an optical microscope image of the ore is taken to obtain an optical microscope image containing image information at the pixel level (first step). The type of optical microscope used to obtain the optical microscope image is not particularly limited, but preferably, an optical microscope equipped with a digital camera can be used.

[0022] The magnification and field of view of the optical microscope image can be appropriately set depending on the complexity of the mineral being analyzed and the required accuracy. However, from the viewpoint of improving the accuracy of mineral analysis as described later, it is preferable to set the conditions so that the size of one pixel (square) of the optical microscope image is 1 μm or less. This is because if the size of one pixel is 1 μm or less, the accuracy of the analysis of the mineral species corresponding to that pixel is high, and the error in the calculation of the quantitative determination of the mineral species for the entire analysis target is small. From this viewpoint, it is more preferable that the size of one pixel is 0.9 μm or less, even more preferable that it is 0.8 μm or less, even more preferable that it is 0.7 μm or less, even more preferable that it is 0.6 μm or less, and even more preferable that it is 0.5 μm or less. The lower limit of the size of one pixel can be set under appropriate conditions depending on the tissue and size of the object being observed.

[0023] (3. Extraction of mineral particle regions) Next, the optical microscope image of the object to be analyzed is binarized, and the bright areas corresponding to the mineral particles are extracted as mineral particle regions (second step). As mentioned above, since the mineral particles to be analyzed have undergone pretreatment, resins, graphites, etc., are present in the optical microscope image. To exclude these resins and graphites, the optical microscope image is binarized. This is because resins and graphites do not reflect much light and become dark areas when binarized. If the size of the optical microscope image is large, the image data may be divided into sizes that can be analyzed and processed. The number of mineral particle regions may vary depending on the particle size of the mineral particles, the size of the optical microscope image, etc.

[0024] Furthermore, the binarization process only requires that the optical microscope image be converted to a grayscale image, provided that the mineral particles to be analyzed can be appropriately extracted; there are no specific restrictions on the conditions. For example, regarding the brightness of the optical microscope image, 0 represents black and 255 represents white, and 128 is used as the threshold. By replacing the brightness of pixels with a value less than 128 with 0 and the brightness of pixels with a value of 128 or more with 255, binarization can be performed. Alternatively, the color image may be converted to grayscale before the binarization process.

[0025] (4. Processing of image information) Next, for each pixel to be analyzed in the mineral particle region, either S and V data or S and L data is selected. If the image information of the optical microscope image is RGB data, the image information consisting of RGB data is converted to HSV data or HSL data on a pixel-by-pixel basis, and from the HSV data or HSL data, either S and V data or S and L data is selected. RGB refers to an additive color space that reproduces a wide range of colors by mixing the three primary colors: red, green, and blue. HSV refers to a color space consisting of Hue, Saturation, and Value / Brightness components. HSL refers to a color space consisting of Hue, Saturation, and Lightness components.

[0026] Considering a space with two or more types of image information as variables (dimensional axes), each pixel contained in the mineral particle region can be plotted at a specific location in that space according to the value of its image information (hereinafter also referred to as pixel data). In the analysis of mineral species described later, the mineral species of each pixel is analyzed based on its position in this space.

[0027] In particular, with regard to HSV data, when the light source used for optical microscope images is white light, the S and V data are the most suitable for identifying mineral species among the three data types: H, S, and V. This is because the differences between the S and V data for different mineral species are significant. On the other hand, with regard to HSL data, when the light source used for optical microscope images is white light, the S and L data are the most suitable for determining mineral species among the three data types: H, S, and L. This is because the differences between the S and L data for different mineral species are significant.

[0028] The following explanation uses the example of selecting S and V values ​​from HSV data as image information for mineral species analysis. Hereafter, the specific value of S will be referred to as the S value, and the specific value of V will be referred to as the V value. The combination of S and V values ​​will be referred to as SV data. Similarly, if S and L values ​​from HSL data are selected, the combination of S and L values ​​will be referred to as SL data. In this specification, descriptions of "SV data" can be interpreted as descriptions of "SL data."

[0029] When SV data from multiple pixels is plotted in a two-dimensional space (hereinafter also called SV space) with S and V as variables (dimensional axes), pixels of the same mineral species tend to concentrate within a certain range in the SV space (this plot of SV data from multiple pixels in the SV space is called an SV plot). Different mineral species will have different ranges of pixel concentration. Therefore, by looking at the range in which a pixel of interest is located, it is possible to identify which mineral species that pixel corresponds to.

[0030] For example, Figure 2 shows a graph plotting the SV data for the mineral species Chalcopyrite (abbreviated as "Ccp") and Pyrite (abbreviated as "Py"). In the SV plot, it can be clearly seen that the regions where the SV data for the two are densely concentrated are different.

[0031] Here, a method used for identifying mineral species in some embodiments of the present invention is described. Some embodiments of the present invention include cases in which mineral species are not determined and cases in which mineral species are determined. In this embodiment, determination means identifying the type of mineral species contained in the mineral particle region, and identification means determining which mineral species each pixel corresponds to based on the image information of each pixel in the mineral particle region. That is, in this embodiment, the mineral species is determined for each pixel in the mineral particle region, but determination is a procedure to pre-identify the mineral species that may be contained in the mineral particle region before pixel-by-pixel identification, if necessary.

[0032] (5. If mineral species is not determined) If mineral species determination is not performed, the third step involves identifying the mineral species for each pixel by comparing the SV data of each pixel in the mineral particle region with a pre-defined mineral identification region for each mineral species in the SV space.

[0033] A mineral identification region is a region in the SV space. A mineral identification region is defined by obtaining SV plots from multiple pixels that are known to be a specific mineral species, and then defining the mineral identification region for that specific mineral species to include those SV plots. For example, the region where SV plots of multiple pixels that are known to be Ccp are distributed is defined as the mineral identification region for Ccp, and the region where SV plots of multiple pixels that are known to be Py are distributed is defined as the mineral identification region for Py.

[0034] Since the combination and frequency of mineral species vary from mine to mine, the mineral identification area may be set for each mine, or it may be set to be broader to cover the differences between mines.

[0035] In this way, by setting a mineral identification region based on the SV plots of multiple pixels whose mineral species are known in advance, and by looking at which mineral species' mineral identification region each pixel's SV data falls into, it is possible to identify the mineral species for each pixel.

[0036] (6. When determining the mineral species) Next, we will explain the case where mineral species determination is performed. Depending on the mineral species, the distribution of SV plots may overlap even for different mineral species. As shown in Figure 2, the distributions of the SV plots for Ccp and Py are not completely separate, and the extent of the distribution (i.e., the area where SV data can exist) partially overlaps. In the above case where mineral species determination is not performed, if the mineral identification areas overlap between different mineral species, it becomes impossible to identify the mineral species. Therefore, the mineral identification areas are set so that they do not overlap between mineral species. In the above-mentioned area where the SV plots of Ccp and Py overlap, the area is divided into two, one area is assigned to the mineral identification area of ​​Ccp, and the other area is assigned to the mineral identification area of ​​Py. Therefore, even if only Py is present and Ccp is not, if a part of the SV data is in the mineral identification area of ​​Ccp, that pixel will be identified as Ccp, and the accuracy of identification will be impaired. In such cases, the mineral species may be determined in advance before identifying the mineral species of each pixel using the mineral identification area.

[0037] In determining mineral species, the inventors focused on the differences in the peak positions of pixel density in the SV plot (hereinafter also referred to as the peak positions of the SV plot). That is, although the distributions of the SV plots for each mineral species often overlap, the peak positions of each distribution (i.e., the positions where pixel data is concentrated) are different, and do not overlap with other mineral species. In a preferred embodiment of the present invention, this characteristic can be utilized to enable more accurate determination.

[0038] Furthermore, if the population of pixels from which the peak position of the SV plot was determined contains only one mineral species, only one peak position will appear. On the other hand, if multiple mineral species exist, multiple peak positions corresponding to those mineral species will appear. In this way, by determining the peak position, it is possible to identify the mineral species included in the population of those pixels, preventing some pixels from being mistakenly identified as other mineral species when only one mineral species actually exists, as in the Py and Ccp example mentioned above.

[0039] In determining the mineral species, a distribution data of pixels (hereinafter also referred to as SV plot distribution data) is created by plotting the SV data of each pixel under analysis in SV space. The pixels under analysis are at least a part of the mineral particle region in question. Based on this distribution data, image information of the peak position of the pixel density is obtained. The pixel population from which the peak position is determined may be all pixels contained in one mineral particle region, or it may be a part of the pixels within one mineral particle region. Alternatively, pixels from multiple mineral particle regions may be added together to form a single population. Here, we will explain the case where the peak position is determined using all pixels contained in one mineral particle region as the population.

[0040] The distribution data of pixels is obtained by plotting the SV data of each pixel in a single mineral particle region in SV space. Using a Gaussian function (normal distribution) as the kernel function, with S and V values ​​as explanatory variables and the number of pixels as the dependent variable, kernel density estimation is performed by appropriately setting the bandwidth, and the S and V values ​​of the peak position of the pixel density can be obtained. However, the method for determining the peak position is not particularly limited; for example, it is possible to draw a circle in SV space and set the center of the circle containing the largest number of pixels as the peak position.

[0041] Once the peak positions are determined, the mineral species are then identified. Specifically, the peak positions of the pixel density in the distribution data of the SV plot are compared with mineral identification regions corresponding to known mineral species to determine the mineral species included in the analysis. The mineral identification region refers to the range set in the spatial image information to identify mineral species based on the peak positions mentioned above. When determining mineral species, only the mineral identification region of the mineral species determined by comparison with this mineral identification region is used in the subsequent identification of mineral species for each pixel using the mineral identification region. In other words, this mineral species determination identifies the mineral species of the mineral identification region used in the subsequent identification of mineral species for each pixel, out of the mineral identification regions of multiple mineral species.

[0042] The mineral identification region is defined using multiple mineral particle regions that are known in advance to be a specific mineral species (e.g., Ccp). Specifically, the peak positions of each of these multiple mineral particle regions are determined by the method described above, and the mineral identification region for that specific mineral species is defined to include the multiple peak positions obtained from these multiple mineral particle regions. The mineral identification region is typically a range enclosed by a certain closed shape. This closed shape can be appropriately defined depending on the distribution characteristics of the peak positions and the required accuracy, and may be, for example, circular, elliptical, or rectangular. The shape, position, and size of this closed shape are set to enclose the aforementioned peak positions of each mineral species. However, the peak positions do not necessarily have to be at the geometric center of the mineral identification region.

[0043] Figure 3 shows the mineral identification regions determined for ore containing both Ccp and Py mineral particles. While Figure 2, mentioned earlier, plots the SV data of each pixel contained within a single mineral particle region, Figure 3 plots the peak positions of multiple mineral particle regions after identifying them according to the procedure described above. That is, in Figure 3, each data point represents the peak position calculated for one mineral particle region. Each mineral identification region (shaded area) encloses these peak positions, but is set so as not to overlap with one another.

[0044] As shown in Figure 3, even for the same mineral species, the peak positions of each mineral particle region are not exactly the same, and a distribution with a certain degree of spread can be observed. Therefore, in order to exclude mineral particle regions that cannot be determined, it is necessary to cover these spreads of distribution when setting up mineral determination regions. Although the characteristics of the spread of distribution differ for each mineral species, when setting up mineral determination regions, it is not necessary to plot the peak positions of the SV plots of all mineral particle regions contained in the ore sample. It is sufficient to plot the peak positions of a sufficient number of mineral particle regions that ensure representativeness in showing the distribution of peak positions for each mineral species. This sufficient number can be set as needed, but for example, it can be set within the range of 30 to 1000.

[0045] Thus, in the SV plot, there may be overlaps in the overall distribution of pixel data for each mineral species (Figure 2), but there is no overlap in the distribution of peak positions between different mineral species (Figure 3). By setting the mineral identification region to cover the distribution of peak positions for each mineral species while avoiding overlaps between different mineral species, it becomes possible to identify mineral species based on peak positions, thus avoiding unidentifiable or overlapping identifications.

[0046] The example in Figure 3 shows Ccp and Py, but the mineral identification area can be set in a similar manner for mineral species other than Ccp or Py. Note that the combination and frequency of mineral species differ from mine to mine, so the mineral identification area may be set for each mine, or it may be set broadly to cover the differences between mines.

[0047] Once the mineral species is determined, the mineral species is identified for each pixel based on the mineral identification region corresponding to each known mineral species, and the mineral species corresponding to the S data and V data of each pixel is identified (third step). However, here, only the mineral identification region of the mineral species determined by comparison with the mineral determination region is used. In other words, when determining the mineral species, it is identified whether each pixel being analyzed is of the mineral species determined by comparison with the mineral determination region.

[0048] In the identification method using mineral identification regions, the mineral identification region encompasses the corresponding mineral identification region and is wider than the corresponding mineral identification region. The mineral identification region is set by defining the mineral identification region for a specific mineral species so that it includes the distribution range of pixel data for multiple mineral particle regions that are known in advance to be that particular mineral species. As a concrete example, as shown in Figure 4, each mineral identification region is set so as to cover the distribution range of the SV data for Ccp and Py respectively (shaded areas in the upper right and lower right figures).

[0049] As mentioned above, even for different mineral species, there is overlap in the distribution of pixel data, resulting in overlapping mineral identification regions. However, if the mineral species of the mineral particle is determined beforehand based on the mineral identification region, it is possible to narrow down which mineral species' mineral identification region to use. Specifically, even if some of the SV data is distributed in the overlapping region of the Ccp and Py SV plots described above, if the peak position of the SV plot is observed only in the Py mineral identification region, it can be determined in advance that the mineral particle contains only Py. In this case, the Py mineral identification region is used for identifying the mineral species for each pixel, and the Ccp mineral identification region is not used. Then, by setting the Py mineral identification region to a wide range that includes the overlapping region of the Ccp and Py SV plots, all pixels can be correctly recognized as Py, avoiding the misidentification as Ccp as described above. Note that, as can be seen in the lower left of Figure 4, not all SV data within a mineral particle region is necessarily included in the mineral identification region, but if it is included in the mineral identification region, these SV data can identify the corresponding mineral species.

[0050] Furthermore, even when narrowing down mineral species using mineral identification regions, it is possible that in rare cases, the mineral identification regions may partially overlap depending on the mineral species. For example, in the above example, if the peak position of the SV plot is observed in both the Ccp and Py mineral identification regions, identification will be performed using both the Ccp and Py mineral identification regions. Therefore, it is necessary to decide how to handle the pixel data contained within the overlapping region. Arbitrary rules can be established as needed for handling such pixel data. For example, the following methods can be considered. (1) Of the multiple mineral species identified using the mineral identification area, identification is first performed using the mineral identification area of ​​one mineral species, and then using the mineral identification areas of the other mineral species. The order in which each mineral identification area is used can be determined arbitrarily. In this case, pixel data within areas where the mineral identification areas overlap will be identified as the mineral species of the last mineral identification area used, and will not be identified as any other mineral species. (2) The distance between the pixel data in the overlapping mineral identification regions and the peak position of the mineral identification region for each mineral species is calculated, and the mineral species corresponding to the peak position with the shorter distance is identified. (3) Modify the shape of the mineral identification area to avoid overlap. When modifying, it is ideal to include as much as possible the distribution range of the pixel data for the mineral species corresponding to each mineral identification area.

[0051] However, in reality, even if overlap occurs, the pixel data contained within the overlapping area represents only a small portion of the total data, so arbitrarily deciding how to handle it has little impact on the accuracy of the analysis.

[0052] Thus, when determining mineral species, it becomes possible to narrow down the types of mineral species contained in the mineral particle region, enabling higher-precision identification for each pixel. On the other hand, regarding pixel-by-pixel identification, the observation results of the object being analyzed change depending on the light source conditions, so it is necessary to perform corrections as described later. In other words, regardless of whether mineral species are determined or not, the accuracy problem caused by the light source conditions exists in common, and the present invention solves this problem by implementing the corrections described later.

[0053] Furthermore, if the composition of the ore to be analyzed is relatively simple, there is little need to pre-select the mineral species using the mineral identification area. As explained in the section on cases where mineral species determination is not performed, it is also possible to directly identify the mineral species for each pixel using multiple mineral identification areas without performing mineral species determination.

[0054] (7. Analysis Procedure) The significance of the mineral determination area and the mineral identification area has been explained above. Below, we will show the procedure for analyzing an ore sample using the mineral determination area and the mineral identification area in the analysis method of this embodiment (Figures 5 and 6).

[0055] First, if the mineral species is already known and there is no need to determine the mineral species, the mineral identification area will be used to identify each mineral particle area. An example of the procedure in this case (when mineral species determination is not performed) is shown in Figure 5.

[0056] First, a correction is made to the mineral identification area to be used (S101). This correction is a procedure to ensure the accuracy of the mineral identification area, and its details will be described later.

[0057] Next, a single mineral particle region is selected for mineral species identification (S102). The mineral particle region is extracted by binarizing the aforementioned optical microscope image.

[0058] Next, in order to perform pixel-by-pixel identification of the image information of the mineral particle region, one pixel is selected (S103).

[0059] Next, the selected pixels are compared against the mineral identification region to identify which mineral species they correspond to (S104). If the pixel data (SV data in the case of an SV plot) is within the mineral identification region, it is identified as corresponding to that mineral species; otherwise, it is identified as a non-mineral.

[0060] Next, it is determined whether the above identification has been completed for all pixels in the mineral particle region. If it has not been completed, the process returns to S103, selects the next pixel, and repeats S103 and S104. If it has been completed, the process proceeds to the next step (S105).

[0061] Next, it is determined whether the above determination and identification have been completed for all mineral particle regions. If not, the process returns to S102, the next mineral particle region is selected, and the above procedure is repeated. If completed, the determination and identification are terminated (S106). In this embodiment, the repeated identification procedure is performed on a single mineral particle region as the unit. However, it is also possible to perform the repeated identification procedure on multiple mineral particle regions as the unit, or to divide a single mineral particle region into multiple subdivided regions and perform the repeated identification procedure on each subdivided region as the unit.

[0062] Furthermore, in another embodiment of the present invention, an example of the procedure for determining the mineral for each mineral particle region (determining the mineral species) is shown in Figure 6.

[0063] First, the mineral identification area and mineral determination area to be used are corrected (S201). This correction is a procedure to ensure the accuracy of the mineral identification area, and its details will be described later.

[0064] Next, a single mineral particle region is selected for the determination and identification of the mineral species (S202). The mineral particle region is extracted by binarizing the aforementioned optical microscope image.

[0065] Next, an SV plot of the mineral particle region is created, and the peak position is calculated using the method described above. Then, the peak position is compared with the mineral identification regions set for known mineral species to confirm which mineral identification region it falls into, and the mineral species is determined (S2021).

[0066] Once the mineral species is identified, one pixel is selected from the image information of the mineral particle region in order to perform pixel-by-pixel identification (S203).

[0067] Next, the selected pixels are compared against the mineral identification region corresponding to the mineral species determined in S202 to determine whether the pixels correspond to that mineral species (S204). If the pixel data (SV data in the case of an SV plot) is within the mineral identification region, it is identified as corresponding to that mineral species; otherwise, it is not identified as a mineral.

[0068] Next, it is determined whether the above identification has been completed for all pixels in the mineral particle region. If it has not been completed, the process returns to S203, selects the next pixel, and repeats S203 and S204. If it has been completed, the process proceeds to the next step (S205).

[0069] Next, it is determined whether the above determination and identification have been completed for all mineral particle regions. If not, the process returns to S202, the next mineral particle region is selected, and the above procedure is repeated. If completed, the determination and identification are terminated (S206). In this embodiment, the repeated determination and identification procedure is performed on a single mineral particle region as the unit. However, it is also possible to perform the repeated determination and identification procedure on multiple mineral particle regions as the unit, or to divide a single mineral particle region into multiple subdivided regions and perform the repeated determination and identification procedure on each subdivided region as the unit.

[0070] In the above explanation, the case in which S and V from HSV data are selected as the image information used for the analysis of mineral species was described as an example. However, the present invention is not limited to this. For example, S and L from HSL data may be selected as the image information used for the analysis of mineral species. When selecting S and L from HSL data, the V data in the above explanation should be read as L data. Specifically, for example, the SV space should be read as the SL space, and the SV plot as the SL plot.

[0071] (8. Correction of mineral identification area and / or mineral identification area) Here, once the mineral determination area and / or mineral identification area are set, it is desirable that similar determination results be obtained even when analyzing a different object under the same conditions. However, slight differences in the angle when placing the object on the optical microscope can cause fluctuations in the brightness of the object, which may result in slight variations in the obtained SV plot (or SL plot).

[0072] Figure 7 shows the dependence of the S value on V or L in the SV plot and SL plot, respectively, when the brightness of the light source of the optical microscope is changed. In other words, Figure 7 shows how the S value and the V value or L value change when the brightness of the object being analyzed changes. As can be seen from the SV plot in Figure 7, if the V value (brightness) is below a certain value, the change in the S value is small for any mineral species, but when the V value is above a certain value, the S value decreases significantly. Therefore, when initially setting the mineral identification region, it is preferable to set the brightness of the light source so that the V value is within a range where the S value does not change significantly for any mineral species, even when the brightness of the object being analyzed changes. In this way, the accuracy of judgment and identification will not be greatly affected even if no correction is made to the S value. On the other hand, in the SL plot, if the L value (luminance) is below a certain value, the change in the S value is small for any mineral species, but when the L value is above a certain value, the S value increases significantly. Therefore, when initially setting the mineral identification area, it is preferable to set the brightness of the light source so that the L value is within a range where the S value does not fluctuate significantly for any mineral species, even if the brightness of the object being analyzed changes. In this way, the accuracy of judgment and identification will not be significantly affected even if no correction is made to the S value.

[0073] On the other hand, if the brightness of the light source changes, the V value or L value will also fluctuate accordingly. Therefore, even for the same mineral species, the peak position of the SV plot will fluctuate in the V direction, and the peak position of the SL plot will fluctuate in the L direction. Thus, it is necessary to perform corrections in the V direction or L direction for both the mineral determination region and the mineral identification region. The correction in the V direction in the SV plot will be explained below as a specific example.

[0074] The mineral identification region and mineral identification region are determined in the SV space based on the average V value of the peaks in the SV plot of a specific mineral species (hereinafter, this V value will be referred to as the reference value). In this embodiment, the case in which Py is selected as the specific mineral species will be explained as an example.

[0075] First, an SV plot of Py is created for one mineral particle region containing Py, and the V value of the peak in that SV plot is obtained. This acquisition of the V value of the peak is repeated for a sufficient number of mineral particle regions, and the average value of the V values ​​at multiple peak positions obtained in this way is used as the reference value for this measurement. The number of repetitions can be set appropriately based on the required accuracy of the analysis, but for example, it can be 30 to 1000 times. By setting the number of repetitions within this range, the reference value is highly reliable as representative of the V values ​​of the peaks of all Py-containing mineral particles in the ore sample.

[0076] Next, the reference value obtained by the above procedure is compared with the reference value before correction, and the difference is taken as the correction amount in the V direction for the entire mineral determination region and the entire mineral identification region. This reference value before correction may be the reference value obtained by the above procedure in the previous analysis, or it may be a reference value adopted at any previous point in time. Figures 8 and 9 show examples of the mineral determination region and mineral identification region of the SV plot before and after correction, respectively.

[0077] In Figure 8, of the two horizontal dotted lines, the upper horizontal dotted line shows the reference value of Py before correction, and the lower horizontal dotted line shows the reference value of Py after correction. The difference between these reference values ​​is taken as the correction amount in the V direction for the entire mineral identification region. In Figure 8, it can be seen that the mineral identification region shown in the graph on the right has shifted slightly in the V direction (in this example, the direction in which the V value decreases) compared to the mineral identification region shown in the graph on the left. In other words, the shape of the mineral identification region is kept the same, and the above correction amount is used to shift it in the V direction before performing the aforementioned determination. The same applies to mineral identification regions other than Py. In the left graph of Figure 8 (before correction), the mineral identification region is positioned based on the reference value before correction (the average V value of the peaks in the SV plot of Py in the previous analysis), so the mineral identification region is shifted in the direction in which the V value increases relative to the peak position of the SV plot. In contrast, in the mineral identification region of the right graph of Figure 8 (after correction), the peak positions of the SV plot are distributed near the center of the mineral identification region.

[0078] In Figure 9, it can be seen that the mineral identification region for Py shown in the graph on the right has shifted slightly in the V direction (in this example, the direction in which the V value decreases) compared to the mineral identification region for Py shown in the graph on the left (horizontal dotted line). The correction amount in the V direction of the mineral identification region is the same as the correction amount in the V direction of the mineral determination region. That is, the shape of the mineral identification region is kept the same, and the entire mineral identification region is shifted by the above correction amount in the V direction before the aforementioned identification is performed. The same applies to mineral identification regions other than Py.

[0079] Therefore, in this embodiment, each time a new analysis is performed, the correction amount is calculated according to the procedure described above, and the mineral determination area and mineral identification area are moved in the V direction by an amount equivalent to the correction amount. In this embodiment, the same correction amount is applied to all mineral determination areas and mineral identification areas, but in other embodiments, the correction amount for each mineral species can be determined individually for the mineral determination area and mineral identification area.

[0080] As shown in Figure 7, unless the brightness of the light source increases drastically, the S value does not change significantly even if the V value changes. Therefore, it is not necessary to move the mineral determination area and mineral identification area in the S direction during correction. This simplifies the correction procedure. Of course, in some embodiments of the present invention, it is also possible to perform correction in the S direction.

[0081] (9. Determination of mineral species (optional)) As mentioned above, since each pixel in each mineral particle region can identify the corresponding mineral species, the same procedure can be applied to the entire optical microscope image to determine the types of minerals and their respective proportions within the entire object being analyzed. Specifically, the pixels of the entire optical microscope image are analyzed using the method described above, and the proportion of pixels corresponding to each mineral species is calculated. Since mineral particles usually have an isotropic and average distribution, if this is considered to correspond to the proportion of volume of each mineral species, it becomes possible to calculate the mass proportion based on the density of each mineral species.

[0082] When analyzing optical microscope images by dividing them, the divided images may be combined before performing the quantitative analysis described above. In this way, even when image data is divided and processed and then combined later, data consistency is maintained, and the method can be easily applied to processing high-resolution, wide-field images.

[0083] In other embodiments of the present invention, the ore to be analyzed is not limited to copper concentrate, but can be applied to any color-distinguishable sample. Furthermore, when there are multiple parameters, it is possible to perform the analysis in combination with methods such as principal component analysis. [Examples]

[0084] The following examples illustrate the present invention and its advantages, but the present invention is not limited to these examples.

[0085] (1. Correction of mineral identification area and mineral identification area) First, a copper concentrate ore sample was prepared as the sample to be analyzed. Prior compositional analysis had revealed that this copper concentrate ore sample contained Chalcopyrite (Ccp), Chalcocite (Clc), Covellite (Cv), Bornite (Bn), Pyrite (Py), and Molybdenite (Mo). Furthermore, the mineral identification region and mineral determination region corresponding to each mineral species were set based on the method described above. The V values ​​of the peak positions in the set mineral identification regions are shown in Figure 8.

[0086] Next, in the SV plot diagram, the two types of mineral identification regions and mineral identification regions were corrected. Specifically, following the method described above, the correction amount in the V direction was uniformly set to -13 for each of the mineral species Ccp, Clc, Cv, Bn, Py, and Mo, and the mineral identification regions were corrected (Figures 8 and 9). In Figures 8 and 9, the differences resulting from the correction amount in the V direction are shown by horizontal dotted lines. For example, in Figure 8, the reference V value is 220, but in Figure 9 after correction, it is 207. The shapes of each mineral identification region and mineral identification region remained unchanged before and after correction.

[0087] (2. Analysis of mineral species) For the above ore samples, a sample embedding resin was prepared using the aforementioned method with a rotational and revolutionary agitator. Specifically, the ore samples were placed in a clear cup together with epoxy resin (EpoCure 2 manufactured by Buehler), and then stirred using a rotational and revolutionary agitator (Awatori Rentaro® manufactured by Thinky Co., Ltd.). After that, the epoxy resin was cured in the air.

[0088] Next, an image of approximately 8 mm × 4 mm was captured using an Olympus DSX500 optical microscope. The obtained optical microscope image was binarized, and only the bright areas were extracted as mineral particle regions (Figure 1). The size of one pixel was 0.92 μm.

[0089] Next, for each of the extracted mineral particle regions, the image information consisting of RGB data at the pixel level was converted to HSV data. Here, an SV plot was created for each extracted mineral particle region after converting from RGB to HSV. For each SV plot, the distribution density was drawn and the peak position was obtained. The obtained peak positions were compared with the mineral identification regions mentioned above to determine the mineral species contained in each mineral particle region.

[0090] Next, based on the aforementioned mineral identification regions, mineral identification was performed for each point on the SV plot. By performing these procedures for all mineral identification regions and stitching the results together, the number of pixels corresponding to each mineral species in the overall optical microscope image could be calculated. The results are shown in Table 1. The determination rate in Table 1 represents the percentage of individual particles in which the mineral species was correctly identified.

[0091] [Table 1]

[0092] As can be seen from Table 1, the detection rate improved. Figure 10 shows a comparison of the optical microscope image with mineral maps with and without correction. These mineral maps were created based on the results of detection and identification using the mineral detection region and mineral identification region described above. As can be seen from Figure 10, without correction, some mineral particle regions could not be identified as chalcopyrite, but with correction, they could be accurately identified.

[0093] As can be seen from the analysis results above, by applying corrections to the mineral determination area and mineral identification area, we were able to obtain compositional information that is quite close to the actual situation.

Claims

1. A method for analyzing the mineral species of ore, The first step involves photographing the aforementioned ore with an optical microscope to obtain an optical microscope image containing image information at the pixel level, A second step involves binarizing the optical microscope image to extract the mineral particle region corresponding to the mineral particle, A third step involves comparing the image information of each pixel with a pre-set mineral identification region for each mineral species to identify which mineral species each pixel belongs to. Includes, The image information used in the third step described above is either the S and V data or the S and L data. The mineral identification region is a region set on the S-V space or the S-L space, A method for analyzing the mineral species of ore, characterized in that, in the third step, correction is performed on the mineral identification region in at least the V direction or the L direction before identifying the mineral species.

2. Furthermore, prior to the third step, the process includes an additional step of creating distribution data of an S-V plot or S-L plot by plotting the S and V data or the S and L data for each pixel in an S-V space or S-L space for the pixels to be analyzed in the mineral particle region, comparing the peak position of the pixel density in the distribution data with the mineral determination region for each mineral species, and determining the mineral species included in the analysis target. The mineral determination region is a region set on the S-V space or the S-L space. In the aforementioned additional step, the mineral determination region is corrected in at least the V direction or the L direction before determining the mineral species. The method for analyzing the mineral species of ore according to claim 1, characterized in that the third step involves identifying the mineral species only for those mineral species determined to be included in the analysis target in the additional step.

3. The mineral species analysis method for ore according to claim 2, wherein the mineral determination area is set so as not to overlap between different mineral species.

4. The method for analyzing the mineral species of ore according to claim 1 or 2, wherein the correction of the mineral identification region is performed only in the V direction or the L direction, and not in the S direction.

5. The method for analyzing the mineral species of ore according to claim 2 or 3, wherein the correction of the mineral determination region is performed only in the V direction or the L direction, and not in the S direction.

6. Furthermore, the method for analyzing the mineral species of ore according to claim 1 or 2, further comprising the step of calculating the mass percentage of each mineral species based on the area percentage of each identified mineral species and the known density of each mineral species.

Images