Ore particle granularity detection device and method based on flexible vibration disc

Abstract

The invention provides an ore particle granularity detection device and method based on a flexible vibration disk, and the device comprises the flexible vibration disk which is used for vibrating an ore material at a certain frequency and amplitude, so that the ore material is uniformly distributed and is in a proper posture; the image acquisition device is used for continuously shooting the ore materials on the flexible vibration disc to obtain ore images; the image processing device is used for carrying out preprocessing, segmentation, feature extraction and particle size analysis on the ore image and outputting the ore particle size distribution condition and an accumulative distribution curve; and the result output device is used for displaying the ore image and the granularity analysis result. According to the invention, while material accumulation is prevented, real-time, accurate and efficient detection of attitude distribution of ore materials is realized, and the method has the advantages of low cost of detection equipment, easy maintenance, strong interference resistance, high accuracy and stability of detection results, and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of particle size detection, in particular to a particle size composition detection method and system for irregularly shaped ore particle groups combining a flexible vibrating disc device and machine vision technology. BACKGROUND

[0002] Ore particle size is a key evaluation indicator for measuring the degree of ore crushing, and is also a main influencing factor affecting the separation effect in the gravity separation section of mineral processing. The traditional ore (3-50m) particle size detection often uses a screening method, which has a complicated measurement procedure and a long lag time. At present, machine vision technology has been applied to a certain extent in the field of ore particle size detection. By using machine vision technology, non-contact, high-precision and high-efficiency detection of ore particle size can be realized. Machine vision technology usually uses an image sensor to obtain image information of a projection surface of an ore particle, and extracts the particle size information of the ore through processing and analysis of the image. However, in the actual application process of the particle size detection method based on the vision technology, the measured material cannot be avoided to be stacked, shielded and in a bad posture (standing or vertical), which leads to a large loss and error of the particle size detection result. According to the shape characteristics of different shaped ore particles, the present application proposes a classification method for ore particles suitable for machine vision recognition, and predicts the ore posture in the recognition result through experiments to count the posture distribution of different shaped particles after vibration. We use a flexible vibrating disc device to effectively disperse the ore particle groups, study the influence of different vibration intensities and vibration frequencies on the dispersion degree of the ore particles, and set an evaluation index for the position dispersion effect of the ore particles to determine whether the effective dispersion is achieved after vibration in the present application. In order to exclude the influence of particles in a bad posture on the particle size detection result, we divide all the ore into four categories according to the shape characteristics of the ore, with the degree of one-way extension (length) and two-way extension (flatness) as the division index, and experimentally and statistically count the posture distribution probability of ore particles of different shape categories, so as to correct the particle size of the ore in a bad posture. Finally, we propose a flexible vibrating disc-based ore material dispersion detection system integrating posture information. SUMMARY

[0003] The present application aims to overcome the shortcomings of the prior art, and provides a flexible vibrating disc-based ore particle size detection device and method. The system can prevent material stacking while realizing real-time, accurate and efficient detection of the posture distribution of ore material, and has the advantages of low cost, easy maintenance, strong anti-interference, high accuracy and stability of the detection result, etc.

[0004] In order to achieve the above object, the present application adopts the following technical scheme: A kind of ore particle size detection device based on flexible vibration disc, comprising: Flexible vibration disc is used to vibrate ore material at a certain frequency and amplitude, so that it reaches uniform distribution and suitable posture; The working principle of flexible vibration disc is to use resonance and interference principle of coherent wave to achieve the process of material moving to a certain direction to form arrangement;Flexible vibration disc uses voice coil motor as power source, and controls the vibration frequency of feeding disc to approach the inherent frequency of material by controlling voice coil motor through flexible vibration disc controller, so as to cause resonance of material and make material in active (easy to move) state.

[0005] Image acquisition device is used to continuously shoot ore material on flexible vibration disc and obtain ore image; Image processing device is used to pre-process, segment, feature extract and size analyze ore image, and output ore size distribution and cumulative distribution curve; Result output device is used to display ore image and size analysis result.

[0006] As preferred, the flexible vibration disc comprises: A load tray for carrying ore material, the surface of which is covered with a layer of antifouling material; A voice coil motor for driving the tray to vibrate, four voice coil motors (equivalent to four vibration points) are distributed at the four inner corner positions of the flexible vibration disc, the voice coil motor includes magnet and coil, when the coil passes through alternating current, the coil generates force under the action of magnet, so that the vibration disc vibrates; A controller for controlling the frequency, amplitude and phase of coil current, so as to control the frequency, intensity and direction of vibration of flexible plate.

[0007] As preferred, the image acquisition device comprises an industrial camera, which comprises: A lens for imaging ore material on flexible vibration disc, which has adjustable focal length, aperture and viewing angle to adapt to different shooting distance, lighting condition and shooting range; An interface for transmitting electrical signal to computer, which supports high-speed, high-stability data transmission protocol, such as USB3.0, GigE, etc., to ensure real-time and integrity of image.

[0008] As preferred, the image processing device comprises a computer host and display.

[0009] A detection method using the above ore particle size detection device based on flexible vibration disc, comprising the following steps: First step: vibration dispersion The ore particles are fed onto a flexible vibrating plate, and the vibrating plate is vibrated by driving a voice coil motor. The ore particles to be tested will disperse with each other during the vibration process. Step 2: Image Capture The raw images of the ore material on the flexible vibrating plate are captured using an industrial camera and transmitted to the computer of the image processing unit. Step 3: Image Preprocessing The original image undergoes grayscale conversion, noise removal, and binarization. The original color image is converted to grayscale to better identify the boundaries and textures of ore particles. A bilateral filtering algorithm is used to filter the image, removing noise while preserving the edges of the ore particles. Binarization is performed using a combination of integral image processing and local thresholding to extract the target ore region; this algorithm is characterized by its adaptability to illumination and high efficiency. Morphological operations are used to optimize the binary image, smoothing the edges of the ore particles and removing noise from the image. Step 4: Image Segmentation The watershed algorithm, the concave point detection algorithm, and the SVM-based ore texture recognition model algorithm were used on the preprocessed image to effectively determine the image edge of a single ore particle and accurately segment the material in the image into a single particle. Step 5: Dispersion Assessment The segmented particle swarm image is subjected to coordinate calibration. The centroid coordinates of each particle are obtained by detecting the outer contour of the particles. The Euclidean distance between the position of each particle and the average position of all particles in the swarm is calculated as the raw data, and its coefficient of variation (CV), i.e., the standard deviation of the raw data, is calculated. The average of the original data The ratio of the two values ​​is used to evaluate the overall dispersion of the particle group; if the obtained coefficient of variation is less than 30%, the vibration dispersion step is restarted and the above process is repeated. Step 6: Particle size calculation Based on the segmentation results of each ore particle, the minimum bounding rectangle of its projection is obtained, and the two sides of this minimum bounding matrix are extracted as l and b, respectively. The geometric mean diameter of the particle is calculated as the particle size D for original image recognition. The calculation formula is as follows: .

[0010] Step 7: Attitude Distribution and Granularity Correction The three-axis diameters l, b and h of each ore particle are obtained through image recognition, and on this basis, the elongation and flatness of the particle are calculated, which are used as the basis for classifying the particle in shape. Then, the final attitude distribution result of the particle group is obtained by multiplying the detected particle size with the probability of occurrence of the undesirable attitude in each shape and the correction coefficient. The attitude distribution result of the particle group will help to correct the particle size recognition result.

[0011] Preferably, the dispersion evaluation in the fifth step includes: through position recognition and coordinate statistics of the ore particles after vibration dispersion, the degree of each particle away from the center position is obtained by calculating the Euclidean distance between the coordinate position of each ore particle and the average coordinate position of all particles The formula is as follows:

[0012] The of all particles is counted and calculated, and the average and the standard deviation are calculated. The formula is as follows:

[0013] The coefficient of variation (CV) is obtained by calculating the ratio of the standard deviation to the average, which is used as the evaluation index of the position dispersion of the ore particle group. The calculation formula is as follows:

[0014] The coefficient of variation CV can reflect the dispersion degree of the whole particle group, and the larger the value is, the more uniform the dispersion between the particles is. It is used as the dispersion evaluation index of the ore particle group after vibration. Through experiments, it is found that when the coefficient of variation is more than 30%, the stacking and shielding of the ore particles will not have a great influence on the particle size recognition result.

[0015] Preferably, the attitude distribution and particle size correction in the seventh step are as follows: the ore particles are classified and counted according to the shape characteristics of the ore, and the attitude distribution of the ore is predicted according to the proportion of different shapes, so as to correct the particle size of the particles in undesirable attitude. The distribution of the ore particles in shape is obtained through visual recognition according to the obtained ore particle image. The shape classification is based on the three-axis diameters of the ore particles: height h, width b and length l. The flatness m and elongation n of the particles are set, and the formula is as follows:

[0016] ​​According to the flatness m and the elongation n of the particles, the morphology of the particles is divided into four types, namely, block, sheet, long sheet and column, as shown in the following table:

[0017] Through repeated experiments, the proportion of the undesirable posture distribution (side standing, vertical standing) of the above four different shaped ore particles after vibration is obtained, and the particle size correction coefficient is as follows:

[0018] According to the identified shape proportion of the to-be-measured ore particle group, the undesirable posture appearance proportion and the corresponding correction coefficient are multiplied, and the corrected ore particle size result is obtained, and the calculation formula is as follows:

[0019] Similarly, the following can be obtained:

[0020] wherein represents the original image recognition particle size of the particle, , , and respectively represent the particle size results of the four types of particles after correction.

[0021] Beneficial effects: The present application can prevent material accumulation while realizing real-time, accurate and efficient detection of the posture distribution of ore materials, and has the advantages of low cost, easy maintenance, strong anti-interference, high accuracy and stability of the detection results. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 is a structural schematic diagram of the ore particle size detection device based on the flexible vibration disc of the present application; Figure 2 is a flow chart of the ore particle size detection method based on the flexible vibration disc of the present application; Figures 3-5 is an example diagram of the ore particle dispersion process and dispersion degree value in the flow of the present application; wherein Figure 3 is an example diagram at the beginning of vibration, Figure 4 is an example diagram in the vibration dispersion, Figure 5 is an example diagram after the vibration dispersion is completed; Figures 6-8 is a particle size composition diagram of the particle size recognition result of the present application, and a comparison between the uncorrected recognition result and the artificial screening result; wherein Figure 6 is the result of the present application, Figure 7 is the uncorrected result, Figure 8 is the artificial screening result.

[0023] Explanation of reference numerals in the attached figures: 1 is a flexible vibrating plate, 2 is an image acquisition device, 3 is an image processing device, and 4 is a result output device. Detailed Implementation

[0024] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0025] like Figure 1 As shown, a particle size detection device for ore based on a flexible vibrating plate includes: Flexible vibratory feeder 1 is used to vibrate ore materials at a certain frequency and amplitude to achieve uniform distribution and a suitable posture; the flexible vibratory feeder 1 includes: A material-carrying pallet for holding ore materials, the surface of which is covered with a layer of anti-fouling material; A voice coil motor is used to drive the tray to vibrate. Four voice coil motors (equivalent to four vibration points) are distributed at the four inner corners of the flexible vibrating plate. The voice coil motor includes a magnet and a coil. When an alternating current is passed through the coil, the coil generates a force under the action of the magnet, causing the vibrating plate to vibrate. A controller is used to control the frequency, amplitude, and phase of the coil current, thereby controlling the frequency, intensity, and direction of the vibration of the flexible plate.

[0026] Image acquisition device 2 is used to continuously photograph the ore material on the flexible vibrating plate to acquire ore images; The image acquisition device 2 includes an industrial camera, which comprises: A lens for imaging ore materials on a flexible vibrating disk, with adjustable focal length, aperture and angle of view to adapt to different shooting distances, lighting conditions and shooting ranges; An interface for transmitting electrical signals to a computer, supporting high-speed, highly stable data transmission protocols such as USB 3.0 and GigE to ensure the real-time performance and integrity of images.

[0027] Image processing device 3 is used to preprocess, segment, extract features and analyze particle size of ore images, and output ore particle size distribution and cumulative distribution curve; The image processing device 3 includes a computer host and a monitor.

[0028] Result output device 4 is used to display ore images and particle size analysis results.

[0029] like Figure 2 As shown, a detection method using the above-mentioned ore particle size detection device based on a flexible vibrating plate includes the following steps: Step 1: Vibration Dispersion The ore particles are fed onto a flexible vibrating plate, and the vibrating plate is vibrated by driving a voice coil motor. The ore particles to be tested will disperse with each other during the vibration process. Step 2: Image Capture The raw images of the ore material on the flexible vibrating plate are captured using an industrial camera and transmitted to the computer of the image processing unit. Step 3: Image Preprocessing The original image undergoes grayscale conversion, noise removal, and binarization. The original color image is converted to grayscale to better identify the boundaries and textures of the ore particles. A bilateral filtering algorithm is used to filter the image, removing noise while preserving the edges of the ore particles. Binarization is performed using a combination of integral image processing and local thresholding to extract the target ore region; this algorithm is characterized by its adaptability to illumination and high efficiency. Morphological operations are used to optimize the binary image, smoothing the edges of the ore particles and removing noise from the image. Step 4: Image Segmentation The watershed algorithm, the concave point detection algorithm, and the SVM-based ore texture recognition model algorithm were used on the preprocessed image to effectively determine the image edge of a single ore particle and accurately segment the material in the image into a single particle. Step 5: Dispersion Assessment The segmented particle swarm image is subjected to coordinate calibration. The centroid coordinates of each particle are obtained by detecting the outer contour of the particles. The Euclidean distance between the position of each particle and the average position of all particles in the swarm is calculated as the raw data, and its coefficient of variation (CV), i.e., the standard deviation of the raw data, is calculated. The average of the original data The ratio of the two values ​​is used to evaluate the overall dispersion of the particle group; if the obtained coefficient of variation is less than 30%, the vibration dispersion step is restarted and the above process is repeated; the vibration dispersion and dispersion evaluation results are as follows: Figures 3-5 As shown; The dispersion assessment specifically includes: identifying the location and calculating the coordinates of the ore particles after vibration dispersion, and calculating the coordinate position of each ore particle. Average coordinate position of all particles European distance To determine the degree to which each particle is far from the center, the formula is as follows:

[0030] Statistically analyze and calculate the values ​​of all particles. And calculate its average. , and standard deviation The formula is as follows:

[0031] The coefficient of variation (CV) is obtained by calculating the ratio of the standard deviation to the average, which is used as an evaluation index of the dispersion degree of the ore particle group position, and the calculation formula is as follows:

[0032] The coefficient of variation CV can reflect the dispersion degree of the whole particle group, and the larger the value is, the more uniform the dispersion between the particles is. It is used as an evaluation index of the dispersion degree of the ore particle group after vibration. Through experiments, it is found that when the coefficient of variation is more than 30%, the accumulation and shielding of the ore particles will not have a great influence on the particle size recognition result. Step 6: Particle size calculation According to the segmentation result of each ore particle, the minimum circumscribed rectangle of its projection is obtained, and the two sides of the minimum circumscribed rectangle are extracted as l and b. The geometric mean diameter of the particle is calculated as the particle size D of the original image, and the calculation formula is as follows: The measured particle size data is classified according to the particle size, and in this example, 0-8mm, 8-16mm and 16-25mm are classified as three particle size groups. After the particle size results are counted, the particle size composition calculation method of each particle size group is as follows: Number percentage = (number of particles in the particle size group / total number of particles in all particle size groups) x 100%; Step 7: Attitude distribution and particle size correction The three-axis diameter l, b and h of each ore particle is obtained through image recognition, and on this basis, the elongation and flatness of the particle are calculated, which are used as the basis for classifying the shape of the particle. Then, the detected particle size is multiplied by the probability of occurrence of the bad attitude in each shape and the correction coefficient, and the final attitude distribution result of the particle group is obtained. The attitude distribution result of the particle group will help to correct the particle size recognition result. In this example, columnar limestone particles are selected as the demonstration.

[0033] The attitude distribution and particle size correction are as follows: the ore particles are classified and counted according to the shape characteristics of the ore, and the attitude distribution of the ore is predicted according to the proportion of different shapes, so that the particle size of the particles in the bad attitude is corrected. According to the obtained ore particle image, the distribution of the shape of the ore particle is obtained through visual recognition. The shape classification is based on the three-axis diameter of the ore particle size: height h, width b and length l. The flatness m and elongation n of the particle are set, and the formulas are as follows:

[0034] ​According to the flatness m and the elongation n of the particles, the shape of the particles is divided into four types, namely, block, sheet, long sheet and column, as shown in the following table:

[0035] Through repeated experiments, the proportion of the undesirable posture (side standing, vertical standing) of the above four different shaped ore particles after vibration is obtained, and the particle size correction coefficient is as follows: According to the identified shape proportion of the to-be-measured ore particle group, the undesirable posture appearance proportion and the corresponding correction coefficient are multiplied, and the corrected ore particle size result is obtained, and the calculation formula is as follows:

[0036] Similarly, we have: wherein represents the original image recognition particle size of the particle, , , and respectively represent the particle size results of the four types of particles after correction.

[0037] Step 8: Result verification The detection results of the to-be-measured particle group composed of ore particles by using the particle size detection method of the present application, the general machine vision particle size detection method and the artificial screening method are compared to evaluate the accuracy of the present application. The results are as follows: In addition, the particle size composition statistical graphs obtained by the above three methods are shown in Figure 6 , Figure 7 and Figure 8 .

[0038] The applicant declares that the above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto. It should be understood by those skilled in the art that any changes or replacements within the technical scope disclosed by the present application can be easily thought of by those skilled in the art, and all fall within the protection scope and disclosure scope of the present application.

Claims

1. A flexible vibrating tray based device for detecting the size of ore particles, characterized by: The utility model relates to a kind of flexible vibration plate for ore material, which comprises the following steps: Flexible vibration plate, for vibrating ore material with certain frequency and amplitude, so that it reaches uniform distribution and appropriate posture; The flexible vibration plate comprises: A load tray for carrying ore material, the surface of which is covered with a layer of anti-fouling material; A voice coil motor for driving the tray to vibrate, four voice coil motors are distributed at the four inner corner positions of the flexible vibration plate, the voice coil motor includes a magnet and a coil, when the coil is connected with alternating current, the coil generates force under the action of the magnet, so that the vibration plate vibrates; A controller for controlling the frequency, amplitude and phase of the coil current, thereby controlling the frequency, intensity and direction of the vibration of the flexible plate; Image acquisition device, for continuously shooting the ore material on the flexible vibration plate and obtaining ore images; The image acquisition device comprises an industrial camera, which includes: A lens for imaging the ore material on the flexible vibration plate, with adjustable focal length, aperture and angle of view; An interface for transmitting electrical signals to a computer, which supports high-speed and stable data transmission protocol; Image processing device, for pre-processing, segmentation, feature extraction and particle size analysis of ore images, and outputting ore particle size distribution and cumulative distribution curve; The image processing device comprises a computer host and a display; Result output device, for displaying ore images and particle size analysis results.

2. A method of detecting the size of ore particles using the flexible vibrating tray based ore particle size detection device as claimed in claim 1, characterized in that: The utility model relates to a kind of flexible vibration plate for ore material, which comprises the following steps: First step: vibration dispersion Put ore particles onto the flexible vibration plate, and vibrate the vibration plate by driving the voice coil motor, the ore particles will disperse from each other during the vibration process; Second step: image shooting Use an industrial camera to shoot the original image of the ore material on the flexible vibration plate, and transmit it to the computer of the image processing device; Third step: image pre-processing Convert the original image to grayscale, remove noise and binarize it; Convert the original color image to a grayscale image to better identify the boundaries and textures of ore particles; Use bilateral filtering algorithm to filter the image, which removes noise while preserving the edges of ore particles; Use integral image and local thresholding method to binarize the image, which extracts the ore target area and has the characteristics of illumination adaptability and high efficiency; Use morphological operation to optimize the binarized image, smooth the edges of ore particles and remove noise in the image; Fourth step: image segmentation Use watershed algorithm, concave point detection-based algorithm and SVM-based ore texture recognition model algorithm to effectively determine the image edges of individual ore particles, and accurately segment the material in the image into individual particles; Fifth step: dispersion evaluation The segmented particle group image is coordinate calibrated, the centroid coordinates of each particle are obtained through particle outer contour detection, the Euclidean distance between the position of the particle and the average position of all particles in the particle group is calculated as the original data, the coefficient of variation CV, i.e. the ratio of the standard deviation of the original data to the average of the original data, is calculated to evaluate the dispersion degree of the whole particle group; if the obtained coefficient of variation is less than 30%, the vibration dispersion step is started again, and the above process is repeated If the obtained coefficient of variation is less than 30%, the vibration dispersion step is started again, and the above process is repeated​ Sixth step: particle size calculation According to the segmentation results of each ore particle, obtain its minimum bounding rectangle projection, and extract the two sides of the minimum bounding rectangle as l and b; Calculate the geometric mean diameter of the particle as the original image recognition particle size D, the calculation formula is as follows: ； Seventh step: posture distribution and particle size correction The triaxial diameter l, b and h of each ore particle is obtained through image recognition, on the basis of which the elongation and flatness of the particle are calculated to classify the particle in shape, and then the detected particle size is multiplied by the probability of occurrence of the undesirable posture in each shape and the correction coefficient to obtain the final posture distribution result of the particle group.

3. The method of detecting the size of the ore particles based on the flexible vibrating tray according to claim 2, characterized in that: The dispersion assessment in the fifth step includes: identifying the location and calculating the coordinates of the ore particles after vibration dispersion, and calculating the coordinate position of each ore particle. Average coordinate position of all particles European distance To determine the degree to which each particle is far from the center, the formula is as follows: ； All particles are counted and their diameters are calculated and their average is calculated with the standard deviation according to the following formula: ； The coefficient of variation is obtained by calculating the ratio of the standard deviation to the average, and is used as an evaluation index of the position dispersion of the ore particle group, and the calculation formula is as follows: ； The coefficient of variation CV reflects the dispersion degree of the whole particle group, and the greater the value, the more uniform and dispersed the particles are; it is used as an evaluation index of the dispersion of the ore particle group after vibration; when the coefficient of variation is greater than 30%, the accumulation and shielding of the ore particles will not have a great influence on the particle size recognition result.

4. The method of detecting the size of the ore particles based on the flexible vibrating tray according to claim 3, characterized in that: The posture distribution and particle size correction in the seventh step are as follows: the ore particles are classified and counted according to the shape characteristics of the ore, the posture distribution of the ore is predicted according to the proportion of different shapes, and the particle size of the particles in undesirable posture is corrected; the distribution of the ore particles in shape is obtained through visual recognition according to the obtained ore particle image; The shape classification is based on the three-axis diameter of the ore particle size: height h, width b and length l; the flatness m and elongation n of the particle are set, and the formula is as follows: ； According to the flatness m and elongation n of the particle, the shape of the particle is divided into four types: block, sheet, long sheet and column, as shown in the following table: ； The proportion of undesirable posture distribution of the above four types of ore particles after vibration and the particle size correction coefficient are as follows: ； According to the proportion of each shape of the detected ore particle group, the undesirable posture occurrence proportion and the corresponding correction coefficient are multiplied to obtain the corrected ore particle size result, and the calculation formula is as follows: ； Similarly, we have: ； wherein The original image of the particle represents the particle size, , , and respectively represent the particle size results of the four kinds of shape particles after correction.

Images