An Automatic Identification Method for Fluorite Minerals Based on a Consumer Digital Camera

Abstract

This invention discloses an automatic identification method for fluorite minerals based on a civilian digital camera, belonging to the fields of geological remote sensing and automated mineral processing technology. The method includes the following steps: S1, taking a picture under natural visible light to obtain image P1; S2, taking a picture under ultraviolet light to obtain image P2; S3, setting a position calibration flag; S4, combining images P1 and P2 to form a new raster set M; S5, performing principal component analysis on the new raster set M to obtain the analyzed principal component PC, and determining the principal component of the fluorescent mineral; S6, determining the anomaly threshold and cutting out anomalies; S7, completing the fluorescent mineral identification process. By using a common digital camera to take pictures of the fluorescent minerals to be detected under ultraviolet and natural visible light, the hardware cost is low. The determination of the anomaly threshold and the cutting out of anomalies result in a low detection error rate and broad application prospects.

Description

Technical Field

[0001] This invention belongs to the fields of geological remote sensing and automated mineral processing technology. Specifically, it relates to an automatic identification method for fluorite minerals based on a civilian digital camera. Background Technology

[0002] During the geological prospecting stage, the fluorescence of some fluorite minerals is not obvious in the field. It is necessary to create a relatively dark environment and use ultraviolet light to excite the fluorescence of these minerals, which are then identified visually. However, this method relies on the naked eye, resulting in low accuracy and the potential for damage to the eyes and skin from ultraviolet light. Furthermore, the procedures are complex, and the identification of weak fluorescence largely depends on experience and visual observation, leading to significant differences in results among different technicians.

[0003] In the mineral processing stage, it is necessary to separate useful fluorite minerals from useless surrounding rocks and remove or reduce harmful impurities. Currently, in the field of automated mineral processing technology, methods such as visible light or infrared ore machine vision (CN201710281598.X), intelligent mineral processing based on machine vision technology (CN201610366841.3), visible light image sorting of ores based on Adaboost machine learning (CN201610715882.9), ore machine vision recognition devices and methods based on ultraviolet fluorescence (CN201710298894.0), and ore machine vision recognition based on ultraviolet fluorescence (CN201811003956.1) are commonly used. In the field of ore beneficiation and sorting, the main detection technologies include visible light detection, near-infrared detection, laser detection, XRF (X-ray fluorescence spectrometry), and XRT (X-ray transmission). However, the above methods have the following shortcomings:

[0004] (1) The identification of fluorescent minerals is carried out using infrared, near-infrared detection, laser detection, XRF (X-ray fluorescence spectroscopy) and XRT (X-ray transmission) technologies. However, the hardware costs are high and the mineral processing machinery is prone to damage, resulting in high replacement costs.

[0005] (2) When using visible light, ultraviolet light, infrared machine vision and other methods for ore identification, there are still some technical problems, including: the qualitative identification of ore relies on the color, texture and luster characteristics of the ore surface, which has a high detection error rate; the ore image sorting by machine learning requires complex ore identification training, which is complicated to operate and difficult to implement. The calculation parameters or cutting thresholds for ore identification need to be summarized through a lot of practice.

[0006] To address the aforementioned issues, we propose an automatic identification method for fluorite minerals based on civilian digital cameras. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention provides an automatic identification method for fluorite minerals based on a civilian digital camera, thereby solving the problems mentioned in the background section.

[0008] To achieve the above objectives, the present invention provides the following technical solution: an automatic identification method for fluorite minerals based on a civilian digital camera, comprising the following steps:

[0009] S1. The controller shuts off the ultraviolet light source and controls the camera to take pictures of the fluorescent mineral to be detected under natural visible light, thus acquiring photo P1.

[0010] S2. The controller controls the ultraviolet light source to turn on, the ultraviolet light source is aimed at the fluorescent mineral to be detected to irradiate it, and the camera is controlled to take a picture of the fluorescent mineral to be detected to obtain the picture P2.

[0011] S3. Set the position calibration mark. The position calibration mark is located within the camera's field of view. Perform geometric correction on the photo according to the position mark point.

[0012] S4. Combine photos P1 and P2 to create a new raster set M;

[0013] S5. Perform principal component analysis on the new raster set M, obtain the feature vector, obtain the analysis results, obtain the principal component PC of the analysis, and judge the principal component of fluorescent minerals to obtain the principal component pc of fluorescent mineral identification.

[0014] S6. After determining the principal component pc for fluorescent mineral identification, determine the anomaly threshold and cut out the anomalies;

[0015] S7. Complete the fluorescent mineral identification process.

[0016] To further optimize this technical solution, the method requires the use of hardware devices during the automatic identification process. These hardware devices include a civilian digital camera, an ultraviolet light source, a controller, a PC or handheld computer, and a position calibration marker.

[0017] To further optimize this technical solution, in S1 and S2, the acquired photos P1 and P2 include the red, green and blue bands of the camera, that is, photo P1 is denoted as {R1, G1, B1} and photo P2 is denoted as {R2, G2, B2}.

[0018] To further optimize this technical solution, in step S2, when the ultraviolet light source is directed to irradiate the fluorescent mineral to be tested, the irradiation conditions are the same or similar to the natural visible light irradiation conditions in step S1, and the ultraviolet light source is turned on to irradiate the fluorescent mineral to be tested uniformly or approximately uniformly.

[0019] To further optimize this technical solution, when the camera's shooting position and angle are fixed, the position calibration flag in step S3 does not need to be set in the fixed state, that is, step S3 is skipped, and the process proceeds directly from step S2 to step S4.

[0020] To further optimize this technical solution, in step S3, when performing geometric correction, higher-order transformation, quadratic transformation, and affine transformation methods can be used to calibrate the positions of photos P1 and P2 using calibration crosshairs, so as to keep the pixel error of the same object within one pixel.

[0021] To further optimize this technical solution, in S4, the new grid set M is denoted as {P1,P2}, or it can be denoted as {R1,G1,B1,R2,G2,B2}.

[0022] To further optimize this technical solution, in step S5, the principal component PC is denoted as {pc1, pc2, pc3, pc4, pc5, pc6}. The eigenvector is selected based on the characteristics of the fluorescent mineral in each band of M under visible and ultraviolet light. After absorbing ultraviolet light energy, the fluorescent mineral enters an excited state and immediately de-excites and emits outgoing light with a wavelength longer than the incident light. The wavelength λ of the outgoing light is obtained and mapped to the (R,G,B) spectral range. When λ∈R wavelength range, the eigenvector is selected as R2>R1; when λ∈G wavelength range, the eigenvector is selected as G2>G1; when λ∈B wavelength range, the eigenvector is selected as B2>B1. Based on the result of the eigenvector selection rule, the principal component of the fluorescent mineral is determined, and the principal component pc is selected from the principal component table PC{pc1, pc2, pc3, pc4, pc5, pc6}.

[0023] To further optimize this technical solution, in S6, determining the abnormal threshold includes the following specific content: after determining the principal component pc of fluorescent mineral identification, a histogram test is performed on the band to determine the average value X of the band, the standard deviation is σ, and the cutting threshold M = X + 3σ.

[0024] To further optimize this technical solution, in S6, the cutting anomaly includes the following specific content: after determining the cutting threshold M, the part of the principal component pc≥M is cut.

[0025] Compared with existing technologies, this invention provides an automatic identification method for fluorite minerals based on a civilian digital camera, which has the following advantages: by using a common digital camera to photograph the fluorescent minerals to be detected under ultraviolet light and natural visible light, the hardware equipment is inexpensive. By synthesizing the photos and performing principal component analysis, the abnormal threshold is determined and abnormalities are cut off, resulting in a low detection error rate and broad application prospects. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating an automatic identification method for fluorite minerals based on a civilian digital camera proposed in this invention.

[0027] Figure 2 This is a diagram showing the placement of fluorite minerals in an automatic identification method for fluorite minerals based on a civilian digital camera proposed in this invention.

[0028] Figure 3 The principal component histogram of fluorite minerals is provided for the automatic identification method of fluorite minerals based on civilian digital cameras proposed in this invention.

[0029] Figure 4 This image shows the results of extracting fluorescent sample information for a method for automatic identification of fluorite minerals based on a civilian digital camera, as proposed in this invention. Detailed Implementation

[0030] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] Example:

[0032] Please see Figure 1 An automatic identification method for fluorite minerals based on civilian digital cameras includes the following steps:

[0033] S1. The controller shuts off the ultraviolet light source and controls the camera to take pictures of the fluorescent mineral to be detected under natural visible light, obtaining photo P1.

[0034] The placement of the test specimen for fluorescent minerals is as follows: Figure 2 As shown, the camera is a common civilian color digital camera, and the ultraviolet light source is a UV lamp. In an environment where the UV lamp is turned off, sufficient light source should be ensured to facilitate the acquisition of photos.

[0035] S2. The controller activates the ultraviolet light source, illuminating the fluorescent mineral to be tested. The camera then takes a picture of the mineral, obtaining image P2. During the acquisition of images P1 and P2, the position of the fluorescent mineral remains unchanged.

[0036] S3. Set position calibration markers. The position calibration markers should be located within the camera's field of view, and there should be at least three position calibration markers within the range. Perform geometric correction on the photo according to the position calibration markers.

[0037] S4. Combine photos P1 and P2 to create a new raster set M.

[0038] S5. Perform principal component analysis on the new raster set M, obtain the feature vector, obtain the analysis results, obtain the principal component PC of the analysis, and judge the principal component of fluorescent minerals to obtain the principal component pc of fluorescent mineral identification.

[0039] In determining the principal component of a fluorescent mineral, the fluorescent mineral emits visible light upon ultraviolet irradiation, thus exhibiting energy enhancement in at least one band within the P2 band. Therefore, the eigenvector with opposite signs and the largest sum of absolute values ​​among (red light 1, red light 2), (green light 1, green light 2), and (blue light 1, blue light 2) is selected as the principal component of the fluorescent mineral (red light 2, green light 2, and blue light 2 are generally positive values). The sixth eigenvector shown in Table 1 conforms to the characteristic of (green light 1, green light 2) with opposite signs and the largest sum of absolute values, and can be used as the principal component pc for identifying the fluorescent mineral.

[0040] S6. After determining the principal component pc for fluorescent mineral recognition, the anomaly threshold is determined and anomalies are cut off. The anomaly extraction results are as follows: Figure 4 As shown.

[0041] S7. Complete the fluorescent mineral identification process.

[0042] Specifically, this method requires the use of hardware devices during the automatic identification process. These devices include a civilian digital camera, an ultraviolet light source, a controller, a PC or handheld computer, and a position calibration marker.

[0043] Specifically, in S1 and S2, the acquired photos P1 and P2 include the camera's red, green and blue bands, that is, photo P1 is denoted as {R1, G1, B1} and photo P2 is denoted as {R2, G2, B2}.

[0044] Furthermore, P1 can also be denoted as {red light 1, green light 1, blue light 1}.

[0045] Specifically, in step S2, when the ultraviolet light source is directed at the fluorescent mineral to be tested, the irradiation conditions are the same or similar to the natural visible light irradiation conditions in step S1, and the ultraviolet light source is turned on to irradiate the fluorescent mineral to be tested uniformly or nearly uniformly.

[0046] Specifically, when the camera's shooting position and angle are fixed, the position calibration flag in step S3 does not need to be set in the fixed state, that is, step S3 is skipped and the process proceeds directly from step S2 to step S4.

[0047] Specifically, in S3, when performing geometric correction, higher-order transformation, quadratic transformation, and affine transformation methods can be used to calibrate the positions of photos P1 and P2 using calibration crosshairs, so as to keep the corresponding pixel error of the same object within one pixel.

[0048] Specifically, in S4, the new grid set M is denoted as {P1,P2}, or it can be denoted as {R1,G1,B1,R2,G2,B2}.

[0049] Specifically, in S5, the principal component PC is denoted as {pc1, pc2, pc3, pc4, pc5, pc6}. The eigenvectors are selected based on the characteristics of the fluorescent mineral in each band of M under visible and ultraviolet light. After absorbing ultraviolet light energy, the fluorescent mineral enters an excited state and immediately de-excites and emits outgoing light with a wavelength longer than the incident light. The wavelength λ of the outgoing light is obtained and mapped to the (R, G, B) spectral range. When λ∈R wavelength range, the eigenvector is selected as R2>R1; when λ∈G wavelength range, the eigenvector is selected as G2>G1; when λ∈B wavelength range, the eigenvector is selected as B2>B1. Based on the eigenvector selection rule, the principal component of the fluorescent mineral is determined, and the principal component pc is selected from the principal component table PC{pc1, pc2, pc3, pc4, pc5, pc6}. The eigenvector table corresponding to the principal component PC{...} is shown in Table 1, which is the eigenvector table corresponding to a certain intermediate experimental data.

[0050] Table 1. Principal Component Eigenvectors

[0051] Feature vector Red Light One Green Light 1 Blue Light 1 Red Light II Green Light II Blu-ray II First eigenvector -0.116626 -0.407928 -0.537247 -0.384689 -0.434775 -0.440845 Second eigenvector -0.07171 -0.424685 -0.585559 0.433597 0.391027 0.361543 Third eigenvector 0.963995 0.072899 -0.201591 0.09948 -0.056195 -0.108193 Fourth eigenvector -0.12266 -0.028457 0.121265 0.697868 -0.003465 -0.694555 Fifth eigenvector -0.071346 0.434387 -0.323278 -0.314784 0.676189 -0.381299 Sixth eigenvector 0.178419 -0.677069 0.456765 -0.260774 0.444598 -0.188256

[0052] like Figure 3 As shown, specifically, in S6, determining the abnormal threshold includes the following: after determining the principal component pc of the fluorescent mineral identification, a histogram test is performed on the band to determine the average value X of the band, the standard deviation σ, and the cutting threshold M = X + 3σ.

[0053] Specifically, in S6, the cutting anomaly includes the following: after determining the cutting threshold M, the portion of the principal component pc≥M is cut.

[0054] This invention relates to an automatic identification method for fluorite minerals based on civilian digital cameras. By using common digital cameras to photograph the fluorescent minerals to be detected under ultraviolet light and natural visible light, the hardware equipment is inexpensive. Through image synthesis and principal component analysis, the abnormal threshold is determined and abnormalities are cut off, resulting in a low detection error rate and broad application prospects.

[0055] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0056] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An automatic identification method for fluorite minerals based on a civilian digital camera, characterized in that, Includes the following steps: S1. The controller shuts off the ultraviolet light source and controls the camera to take pictures of the fluorescent mineral to be detected under natural visible light, thus acquiring photo P1. S2. The controller controls the ultraviolet light source to turn on, the ultraviolet light source is aimed at the fluorescent mineral to be detected to irradiate it, and the camera is controlled to take a picture of the fluorescent mineral to be detected to obtain the picture P2. The acquired photos P1 and P2 include the camera's red, green and blue bands, i.e., photo P1 is denoted as {R1, G1, B1} and photo P2 is denoted as {R2, G2, B2}. S3. Set the position calibration mark. The position calibration mark is located within the camera's field of view. Perform geometric correction on the photo according to the position mark point. S4. Combine photos P1 and P2 to create a new raster set M; The new raster set M is denoted as {P1,P2}, or as {R1,G1,B1,R2,G2,B2}; S5. Perform principal component analysis on the new raster set M, obtain the feature vector, obtain the analysis results, obtain the principal component PC of the analysis, and judge the principal component of fluorescent minerals to obtain the principal component pc of fluorescent mineral identification. The principal components PC are denoted as {pc1, pc2, pc3, pc4, pc5, pc6}. The eigenvectors are selected based on the characteristics of the fluorescent minerals in various wavelength bands M under visible and ultraviolet light. After absorbing ultraviolet light energy, the fluorescent minerals enter an excited state and immediately de-excite and emit outgoing light with a wavelength longer than the incident light. The wavelength λ of the outgoing light is obtained and mapped to the (R, G, B) spectral range. When λ∈R wavelength range, the eigenvector is selected as R2>R1; when λ∈G wavelength range, the eigenvector is selected as G2>G1; when λ∈B wavelength range, the eigenvector is selected as B2>B1. Based on the eigenvector selection rules, the principal components of the fluorescent minerals are determined, and the principal component pc is selected from the principal component table PC{pc1, pc2, pc3, pc4, pc5, pc6}. S6. After determining the principal component pc for fluorescent mineral identification, determine the anomaly threshold and cut out the anomalies; Determining the anomaly threshold includes the following specific steps: After determining the principal component pc for fluorescent mineral identification, perform a histogram test on the band corresponding to the principal component pc, determine the average value X of the band, the standard deviation σ, and the cutting threshold M = X + 3σ. S7. Complete the fluorescent mineral identification process.

2. The method for automatic identification of fluorite minerals based on a civilian digital camera according to claim 1, characterized in that, This method requires the use of hardware devices during the automatic identification process. These devices include a civilian digital camera, an ultraviolet light source, a controller, a PC or handheld computer, and a location calibration marker.

3. The method for automatic identification of fluorite minerals based on a civilian digital camera according to claim 1, characterized in that, In step S2, when the ultraviolet light source is directed to irradiate the fluorescent mineral to be tested, the irradiation conditions are the same or similar to the natural visible light irradiation conditions in step S1, and the ultraviolet light source is turned on to irradiate the fluorescent mineral to be tested uniformly or nearly uniformly.

4. The method for automatic identification of fluorite minerals based on a civilian digital camera according to claim 1, characterized in that, When the camera's shooting position and angle are fixed, the position calibration flag in step S3 does not need to be set in the fixed state. That is, step S3 is skipped, and the process proceeds directly from step S2 to step S4.

5. The method for automatic identification of fluorite minerals based on a civilian digital camera according to claim 1, characterized in that, In step S3, during geometric correction, higher-order transformation, quadratic transformation, and affine transformation methods are used. The positions of photos P1 and P2 are calibrated using a calibration crosshair to ensure that the pixel error corresponding to the same object is kept within one pixel.

6. The method for automatic identification of fluorite minerals based on a civilian digital camera according to claim 1, characterized in that, In S6, the cutting anomaly includes the following specific content: after determining the cutting threshold M, the part of the principal component pc≥M is cut.

Images