CN113624645B - Device and method for detecting particle flow distribution - Google Patents
Device and method for detecting particle flow distribution Download PDFInfo
- Publication number
- CN113624645B CN113624645B CN202110810094.9A CN202110810094A CN113624645B CN 113624645 B CN113624645 B CN 113624645B CN 202110810094 A CN202110810094 A CN 202110810094A CN 113624645 B CN113624645 B CN 113624645B
- Authority
- CN
- China
- Prior art keywords
- particle
- image
- area
- index value
- flow
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 239000002245 particle Substances 0.000 title claims abstract description 241
- 238000009826 distribution Methods 0.000 title claims abstract description 48
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000012545 processing Methods 0.000 claims abstract description 24
- 230000011218 segmentation Effects 0.000 claims abstract description 15
- 238000003709 image segmentation Methods 0.000 claims abstract description 13
- 238000001914 filtration Methods 0.000 claims abstract description 10
- 238000004458 analytical method Methods 0.000 claims abstract description 8
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 7
- 230000009466 transformation Effects 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 238000003672 processing method Methods 0.000 claims abstract description 3
- 238000001514 detection method Methods 0.000 claims description 16
- 238000003860 storage Methods 0.000 claims description 13
- 239000000843 powder Substances 0.000 claims description 9
- 241001292396 Cirrhitidae Species 0.000 claims description 7
- 230000007246 mechanism Effects 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 239000004744 fabric Substances 0.000 claims description 4
- 239000003973 paint Substances 0.000 claims description 4
- 238000009499 grossing Methods 0.000 claims description 2
- 238000012544 monitoring process Methods 0.000 claims description 2
- 238000012360 testing method Methods 0.000 claims description 2
- 230000002708 enhancing effect Effects 0.000 claims 1
- 238000010998 test method Methods 0.000 claims 1
- 238000005516 engineering process Methods 0.000 abstract description 5
- 238000000605 extraction Methods 0.000 abstract description 3
- 239000008187 granular material Substances 0.000 description 28
- 230000008859 change Effects 0.000 description 7
- 238000011160 research Methods 0.000 description 6
- 230000008901 benefit Effects 0.000 description 4
- 239000002184 metal Substances 0.000 description 4
- 238000002360 preparation method Methods 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 3
- 230000031700 light absorption Effects 0.000 description 3
- 239000000126 substance Substances 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 239000011148 porous material Substances 0.000 description 2
- 238000013519 translation Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 238000005452 bending Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 239000003292 glue Substances 0.000 description 1
- 230000008676 import Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 239000002086 nanomaterial Substances 0.000 description 1
- 239000013618 particulate matter Substances 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume or surface-area of porous materials
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N11/00—Investigating flow properties of materials, e.g. viscosity, plasticity; Analysing materials by determining flow properties
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N11/00—Investigating flow properties of materials, e.g. viscosity, plasticity; Analysing materials by determining flow properties
- G01N2011/006—Determining flow properties indirectly by measuring other parameters of the system
- G01N2011/008—Determining flow properties indirectly by measuring other parameters of the system optical properties
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume or surface-area of porous materials
- G01N2015/03—Electro-optical investigation of a plurality of particles, the analyser being characterised by the optical arrangement
Landscapes
- Chemical & Material Sciences (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Dispersion Chemistry (AREA)
- Image Processing (AREA)
Abstract
The invention relates to a device and a method for detecting particle flow distribution. The processing method comprises the following steps: (1) image preprocessing: carrying out image decomposition, gray level transformation, noise filtration and edge sharpening enhancement on the acquired image; (2) Image segmentation, namely obtaining a binary image of the particle region through threshold segmentation; (3) feature processing: and detecting the characteristic parameters of the particle image to obtain the particle flow distribution uniformity analysis. The method improves the extraction precision of image parameters through preprocessing, analyzes the particles quickly and accurately through a graph processing technology, calculates the influence weight of different parameters on the flow distribution uniformity of the particles through a coefficient of variation method, and judges the flow condition of the particles according to the influence weight.
Description
Technical Field
The invention discloses a device and a method for detecting particle flow distribution, and belongs to the technical field of particle detection.
Background
With the development of technology being more and more advanced, researchers are also going deep into analyzing and studying particulate matters in powder. The granular materials are distributed in the fields of industry, agriculture, medicine and the like, and the detection analysis and research of the granular materials from various aspects are greatly helpful for mastering the objective rule of the granular materials.
The detection and analysis of the particle substances in the past mainly take physical characteristics such as shape, size and the like as main characteristics, and the methods cannot accurately calculate the characteristic parameters of the particles. The research method is also paid more and more attention by researchers, and is basically one of the necessary means in the particle detection analysis and research process. The image processing technology can assist researchers in observing and analyzing the particle nano structure, quantitatively analyze the properties of particle substances, extract microscopic multidimensional information of powder particles and is more beneficial to realizing real-time online detection. At present, a detection system for analyzing the particulate matter by utilizing image research is verified, and has the advantages of high speed and good repeatability. The rapid and accurate particle analysis based on the image shows that the technology has great development potential for particle detection, and the particle substance analysis based on the image can be developed more deeply in the future.
Disclosure of Invention
The invention provides a device and a method for detecting particle flow distribution, and aims to provide a system for detecting particle flow distribution uniformity based on an image, aiming at the defects of low precision, low efficiency and the like in the existing particle detection technology, so that the characteristic parameters of particles can be rapidly and accurately calculated.
The technical scheme of the invention is as follows: a device for detecting particle flow distribution is characterized by structurally comprising a particle flow system, an image acquisition system and an image processing system, wherein the particle flow system comprises a particle storage hopper, a strong light source and a particle collecting box; the particle storage bin hopper is used for containing particle powder objects, the particle powder objects are fixed through the support, particles flow out of the bottom of the hopper and fall into a particle collection box arranged at the bottom of the support, the background of a particle flow system is black, a strong light source is adopted for irradiation in the flow process, the flow state of the particles is collected through an industrial high-speed camera and a collection card, the particles are monitored and stored in real time through image data collection software, and finally, the collected particle flow images are subjected to image processing through image processing software.
The background of the particle flowing system is black light absorption flocking cloth or a background plate sprayed with black matt paint; the strong light source is a white light LED lamp with power being more than or equal to 40W, and the LED lamp is fixed through a long hose universal lamp holder clamp and can adjust the position and the angle of the light source; the industrial high-speed camera is fixed through the movable support, and the shooting position and angle can be adjusted.
The device is characterized by also comprising a start-stop aperture control structure, wherein the structure comprises a single chip microcomputer controller, a steering engine and an aperture rotation control mechanism, the aperture rotation control mechanism is provided with a plurality of circular rings with different inner diameters, the outer diameter of each circular ring is equal to the outer diameter of the hopper outlet, the aperture rotation control mechanism is rotatably arranged on the steering engine through a rotating shaft, and the steering engine is fixedly arranged at the bottom of the support, so that the circular rings with different apertures can be in close contact with the hopper outlet of the particle storage bin; the control signal input end of the steering engine is connected with the control signal output end of the single chip microcomputer controller, and the single chip microcomputer controller controls the steering engine to rotate at a fixed angle to switch the circular ring, so that the diameter of the outlet of the particle storage hopper is controlled.
Three pins of the single chip microcomputer are connected with three fixing pins at the end of the steering engine through DuPont wires, the DuPont wires are fixedly kept still through buckles, VCC is connected with the positive electrode of a power supply, a signal wire is connected with the IO pin of the single chip microcomputer, and the power ground is connected with the negative electrode of the steering engine.
The industrial high-speed camera is a Lingyun photon 4M180-CL type industrial high-speed camera, a Lingyun photon Myutron series high-resolution industrial lens is matched with the lens, and an OR-X4C0-XPF00 series acquisition card of the Lingyun photon is adopted as an acquisition card; the image data acquisition software adopts the cooperation of Sapera Cam Expert and Stream7, and the image processing software is Halcon machine vision software.
The working method of the image processing software comprises the following steps:
(1) Image preprocessing: carrying out image decomposition, gray level transformation, noise filtration and edge sharpening enhancement on the acquired image;
(2) Image segmentation, namely obtaining a binary image of the particle region through threshold segmentation;
(3) Characteristic processing: and detecting the characteristic parameters of the particle image to obtain the particle flow distribution uniformity analysis.
In the image preprocessing, if the collected image is not a gray image, image decomposition is firstly carried out, an RGB image is decomposed into three component images of R, G and B, and then a component image with obvious gray difference is selected for gray conversion or is converted into an HSV color space; if the collected image is a gray image, directly filtering small noise which is difficult to be perceived by naked eyes in the background by adopting a discrete Gaussian filter function, namely scanning each pixel point in the image ROI through convolution check, replacing the value of a central pixel point corresponding to the convolution kernel by pixel weighted average gray in a convolution neighborhood of the convolution kernel, and performing better smoothing processing on the image to remove the noise; the sharp edge is enhanced to perform linear transformation on the gray value of the image so as to improve the contrast between light and shade, so that the bright position is brighter, the dark position is darker, and the segmentation precision during image segmentation is facilitated;
the image segmentation is to acquire an ROI region, that is, a region where particles are distributed, and the image with the enhanced sharpened edge is subjected to threshold segmentation, so that a binary image of the particle region can be acquired.
The feature processing method specifically comprises the following steps:
(1) Determining the size of the shape area and the position of the center of the area, wherein the size of the area is mainly determined by the number of pixel points of the binary image of the particle area, and the position of the center of the area mainly refers to the row-column coordinate index value of the center of the particle area;
(2) The number of particles flowing to the image acquisition area is represented by the area of the particle area, and the particle area and the image acquisition area are in positive correlation;
(3) The index value of the center coordinates of the particle area is used for representing the smoothness of the particle flow;
(4) Distributing the weight of three influence factors of the area S of the particle area, the index value m of the row coordinate center of the particle area and the index value m of the column coordinate center of the particle area by using a variation coefficient method, and normalizing the three;
(5) The standard particle flow coefficient is P, and the particle flow distribution uniformity is obtained through three influencing factors.
In the distribution of the variation coefficient method, the influence weight of the area S of the particle area, the index value m of the row coordinate center of the particle area and the index value m of the column coordinate center of the particle area on the uniformity of the particle flow distribution is specifically defined as W 1 ，W 2 、W 3 ：
Wherein V 1 、V 2 、V 3 The coefficient of variation, sigma, of the area of the particle region, the index value of the line coordinate center of the particle region, and the index value of the column coordinate center of the particle region 1 、σ 2 、σ 3 Respectively the data standard deviations of the area of the particle area, the index value of the row coordinate center of the particle area and the index value of the column coordinate center of the particle area,
the average value of the area of the particle area, the index value of the row coordinate center of the particle area and the index value of the column coordinate center of the particle area is the data;
normalizing the area S of the particle area, the index value m of the line coordinate center of the particle area and the index value m of the line coordinate center of the particle area by using a z-score function, and recording the processed data as S * 、m * 、n *
In the step (5), the particle flow coefficient is defined as P, and the formula is shown as (7)
P＝|S * |×W 1 +|m * |×W 2 +|n * |xW 3 #(7)
Due to S * 、m * 、n * All conform to the standard normal distribution, soWhen | S * |、|m * |、|n * The larger the value of | is, the more abnormal the flow distribution value is, and it can be considered that the flow state of the particles shows a non-uniform fluctuation tendency at this time, and the larger the particle flow coefficient P is at this time; when | S * |、|m * |、|n * When the value of | is smaller and is close to 0, the flow distribution value is close to a constant value, the flow state of the particles shows a uniform and smooth trend, and the flow coefficient P of the particles is smaller.
The invention has the beneficial effects that:
1) By adopting the cooperation of a high-speed industrial camera and software, the shooting quality can be improved by adjusting the exposure time and the exposure gain, the particle state can be clearly captured, and the experimental efficiency and the data precision can be obviously improved.
2) The weight of different characteristics is distributed by adopting a coefficient of variation method, and the flowing state of the particles can be quickly and accurately obtained. .
3) The starting and stopping aperture control structure is simple and easy to control, and can be used for mechanically controlling the starting and stopping state of the hopper and the aperture of the granule outflow, thereby avoiding inconvenience caused by manual operation and contingency of possible improper operation; hopper of different bottom apertures of needs design probably need carry out redesign and preparation to hopper structure wholly, but adopt this device and need not redesign hopper also can change the aperture, reduce 3D like this and print manufacturing design, preparation material, preparation time, the cost of manufacture of hopper by a wide margin, reduce the quantity of hopper.
Drawings
FIG. 1 is a particle flowability detection system of the present invention;
FIG. 2 is a working apparatus for particle flowability detection according to the present invention;
FIG. 3 is a raw image of a collected particle flow distribution;
FIG. 4 is a Gaussian filtered comparison of a static image of particle flow;
FIG. 5 is a diagram showing the influence of a sharpened edge on segmentation (left: original image, middle: original image segmentation, right: original image segmentation after enhancement);
FIG. 6 is a uniform flow definition schematic;
FIG. 7 is dataTreated S * The data distribution and the fitting curve of (1);
FIG. 8 is a graph plotting the derived particle flow coefficient P;
fig. 9 shows the abnormal data value No. 848 found from the particle flow coefficient P curve.
Fig. 10 is a schematic diagram of a start-stop aperture control architecture.
Detailed Description
1. The invention provides a system for detecting particle flow distribution uniformity based on an image, which mainly comprises a particle flow system, an image acquisition system and an image processing system. A particle fluidity detection working device is set up, and a particle flow distribution uniformity detection method is provided.
2. Among the granule mobility detection equipment, metal support is used for the fixed stay to deposit the hopper of granule, and top granule hopper is used for placing granule powder object, and the granule flows from hopper bottom round hole department, has placed the collecting box bottom the support for collect deposit the granule.
3. The background can adopt black light absorption flocking cloth, or the surface of the background plate is sprayed with black matt paint, so that the image acquisition quality can be improved, and the environmental influence is reduced.
4. The illumination system adopts a single light source, the light adopts a white light LED lamp with the power of about 40W, the environment of the whole flowing device can be obviously illuminated under the conventional working condition, the working voltage of the light source lamp is 220V, the lamp holder clamp adopts a universal lamp holder clamp with a long hose of about 400mm, the light source lamp can be conveniently bent and twisted at will, and meanwhile, the metal hose also has good flexibility and flexibility, and can be well adjusted and fixed.
5. The high-speed camera for carrying out image acquisition on particles in the flow adopts an industrial high-speed camera of a Lingyun photon 4M180-CL model, the lens is matched with an industrial lens of a Lingyun photon neutron series high resolution, the resolution can reach 2048 multiplied by 2048, the highest frame rate under the resolution can reach 179fps, the position and the angle of the torsional translation support can adjust the shooting position of the camera, the camera can improve the shooting quality by adjusting the exposure time and the exposure gain, the particle state can be clearly captured, and the particle state can be monitored and stored in real time through software.
6. The acquisition card adopts an OR-X4C0-XPF00 series acquisition card of the linguistics photon, can realize image acquisition and storage, and simultaneously meets the high standard requirements of high-speed data import, compression, fidelity and transmission.
7. The software for image data acquisition adopts the cooperation of Sapera Cam Expert and Stream 7. The former provides an interface for acquisition card configuration and camera setting, can also realize real-time interactive parameter adjustment for image acquisition configuration, and has other setting and diagnosis functions; the latter can collect and store the motion image in real time, record the image under each frame according to the local computer time of the working PC, and can export to various formats according to the demand, and can do the lossless compression, save and transmit to the image, and can select the proper ROI according to the self demand is also an important advantage of the software.
8. And Halcon machine vision software is used at the PC end to perform image processing on the acquired particle flow images, so that Halcon has the advantages of powerful and complete functions, strong compatibility and great advantage in vision application. The main processes for carrying out image processing on the collected particle flow images comprise filtering denoising pretreatment, image enhancement sharpening edge, image threshold segmentation to convert into binary images, characteristic parameter extraction and the like.
9. Since the image acquired by the cloud photon 4M180CL is a gray image, steps such as image decomposition, gray conversion and the like are not required, if other cameras are adopted, the image decomposition can be performed firstly, the RGB image is decomposed into three component images of R, G and B, then the component image with obvious gray difference is selected for gray conversion, and if necessary, the component image can be converted into a color space such as HSV and the like and then judged.
10. After image acquisition, the gray image is preprocessed firstly, and noises which are tiny in the background and difficult to be perceived by naked eyes are filtered by adopting a discrete Gaussian filter function, each pixel point in the image ROI is scanned by convolution check substantially, the value of a central pixel point corresponding to the convolution check is replaced by pixel weighted average gray in a convolution kernel convolution neighborhood, and the image can be well smoothed to eliminate the noises.
11. The method is characterized in that image enhancement and edge sharpening are carried out before image segmentation, and the essence is that the image is subjected to linear transformation of gray values to improve the contrast between light and dark, so that the bright position is brighter and the dark position is darker, which is more beneficial to the segmentation precision of the image during image segmentation, so that a computer can better find a light and dark boundary, and the possibility that a part of a black background area is divided into an ROI (namely a particle area) by the computer is greatly reduced.
12. The image segmentation is to acquire an ROI (region of interest), namely a region of particle distribution, and a threshold segmentation mode with simple selection and high speed and efficiency is adopted, so that a binary image of a particle region can be very easily and intuitively acquired due to less image noise and obvious contrast difference after filtering processing and image enhancement.
13. The segmented particle region is still a whole, the size of the shape area and the central position of the region are extracted from the binary image of the whole particle connected domain, the size of the area is mainly determined by the number of pixel points of the binary image of the particle region, and the central position of the region mainly refers to the central row-column coordinate index value of the particle region.
14. The particle area represents the number of particles flowing to the image acquisition area, and the two are in positive correlation, as shown in fig. five, the flow area is regarded as a section of pipeline, and if the number of particles flowing in from the particle hopper is approximately equal to the number of particles flowing out, the particles are considered to be in a uniform flow distribution trend in the period of time. The two-dimensional image shows that the area of the particle region in the image has small change. If partial particles to flow out of the particle hopper are agglomerated due to overhigh water content, the bottom of the particle hopper can be blocked even due to unsmooth flow, so that the particle flow fluctuates, the discharging rate is reduced, and the information reflected in the image is that the area of the particle area has obvious mutation within a period of time.
15. The index value of the center coordinate of the particle area is also used for representing the smoothness of the particle flow, and when the particles continuously and smoothly flow, the index value of the center coordinate of the particle area basically does not change greatly. If the particles in the particle hopper suddenly arch or the flowing particles incline due to the influence of wind direction and other factors, the change reflected in the image is the large change of the index value of the center coordinate of the image particle area.
16. Because the sizes of the area of the particle region and the index value of the row-column coordinate center of the particle region are different, the direct comparison of the sizes of the particle region and the index value of the row-column coordinate center of the particle region obviously causes unreasonable influence on the flow distribution uniformity of the particles, and therefore the variation coefficient method is adopted to distribute the weights of the particle region, the row-column coordinate center of the particle region and the index value of the row-column coordinate center of the particle region. The influence weight of the area of the particle region, the index value of the row coordinate center of the particle region and the index value of the column coordinate center of the particle region on the uniformity of the particle flow distribution is as follows 1 ，W 2 、W 3 ：
Wherein V 1 、V 2 、V 3 The coefficient of variation, σ, of the area of the particle region, the index value of the row coordinate center of the particle region, and the index value of the column coordinate center of the particle region 1 、σ 2 、σ 3 Respectively the data standard deviations of the area of the particle area, the index value of the row coordinate center of the particle area and the index value of the column coordinate center of the particle area,
the average value of the area of the particle area, the index value of the coordinate center of the row of the particle area and the index value of the coordinate center of the column of the particle area is obtained.
17. From this, three factors affecting the particle flow distribution are determined, denoted as S, m, n, and the weights W of the three factors are also obtained 1 、W 2 、W 3 . But is composed ofSince the index value data of the area of the particle region and the center coordinate of the particle region are often large and difficult to compare, normalization processing is performed on S, m, and n. When the z-score function is used for standardization processing, the influence of dimensions and a data range can be avoided, meanwhile, the mean value of the processed data is 0, the standard deviation is 1, the standard normal distribution is basically met, and the condition of particle flow distribution can be reflected very intuitively. Recording the processed data as S * 、m * 、n *
18. Defining the particle flow coefficient as P, the formula is (7)
P＝|S * |×W 1 +|m * |×W 2 +|n * |×W 3 #(7)
Due to S * 、m * 、n * All conform to the standard normal distribution, so when | S * |、|m * |、|n * The larger the value of | is, the more abnormal the flow distribution value is, and it can be considered that the flow state of the particles shows a non-uniform fluctuation tendency at this time, and the larger the particle flow coefficient P is at this time; when | S * |、|m * |、|n * The smaller the value of | is, the closer to 0, the flow distribution value is, and it is considered that the flow state of the particles shows a uniform and smooth tendency, and the smaller the particle flow coefficient P is.
Example 1
The technical scheme of the invention is further explained by combining the attached drawings
1. The whole particle flow distribution detection system based on the image mainly comprises three subsystems, namely a particle flow system, an image acquisition system and an image processing system as shown in figure 1. The present example further illustrates that the sugar granule sample is the research object, and the granule sample used in the present invention is the conventional edible sugar granule sold in the supermarket, but the object aimed by the present method is not limited to sugar granules, and should be all the objects such as granules, powder, etc.
2. The particle flowing system is mainly used for realizing the flowing and collecting of particles. As shown in FIG. 2, in the granule flow working apparatus, a metal holder is used for fixedly supporting a hopper for storing granules, a top granule hopper is used for placing a granule powder object, the granules flow out from a round hole at the bottom of the hopper, and a collecting box is placed at the bottom of the holder and used for collecting and storing the granules. The background can be pasted on the background plate by adopting black light absorption flocking cloth, and also can be directly sprayed with black matt paint on the surface of the background plate, thereby improving the image acquisition quality and reducing the environmental influence. The illumination system adopts a 40W white light LED single light source lamp, the universal lamp holder clamp of the long hose can realize the random bending and torsion of the light source lamp, and the good flexibility and flexibility of the metal hose allow the light source to be well adjusted and fixed.
3. The image acquisition system mainly comprises two parts, namely hardware and software, and realizes real-time lossless acquisition, storage and transmission of particle images. In the hardware part, a high-speed 4M180-CL high-speed Lingyun photon camera is adopted as the camera, a Myutron series lens is adopted as the lens, and the position and the angle of the torsional translation support can be adjusted to adjust the shooting position of the camera. The acquisition card adopts a Lingyun photon OR-X4C0-XPF00 series acquisition card. The software part adopts the cooperation of Sapera Cam Expert and Stream 7. The former provides an interface for acquisition card configuration and camera setting, and can also realize real-time interactive parameter adjustment for image acquisition configuration; the latter can carry out real-time acquisition and storage to the motion image.
4. The image processing system mainly realizes the processing of images, obtains interested characteristic parameters and obtains an evaluation method of particle flow distribution uniformity, and the utilized software is Halcon machine vision software.
5. Before the particles flow, firstly, placing and aligning the device, connecting a camera, a light source lamp and a PC workstation with a power supply and connecting, aligning the camera with the working device and enabling the camera to be close to the device as far as possible, adjusting the aperture to be large and maximum, opening camera parameter adjusting software, setting the window resolution to be 400X 300 and the whole center, setting an exposure time parameter to be 150, setting the transmission frame number per second to be 30, keeping the rest parameters in a default state, starting a real-time monitoring function, if the whole image is still dark and the particles and the background are difficult to distinguish, properly adjusting an exposure gain parameter, setting Analog in image gain to be 1.4X, setting Digital to be 1.25X, and keeping the rest parameters in a default state. Then the focus of adjustment camera can be put the scale at granule flow region center, can regard as the focus to adjust to the suitable position when the scale and the figure that the camera can catch the scale clearly, screws the staple on the camera afterwards, makes the diaphragm and focus size can not change because of the mistake touches.
6. After the preparation is finished, sugar particles are poured into a particle hopper, the particles flow smoothly, after the particles flow for about 5 seconds, a recording function is started in software, a particle flow video of a period of time is recorded, the particle flow video is exported and stored as a TIF file, a 16-bit gray level is selected for a storage format, an original pixel value option is selected and reserved, and 1400 continuous images in the flow process are taken and stored in a PC workstation as shown in figure 3.
7. And (3) reading an image by using Halcon software, and then processing the image, wherein the main flow comprises filtering and denoising pretreatment, image enhancement and edge sharpening, image threshold segmentation and conversion into a binary image, characteristic parameter extraction and the like.
8. The image is preprocessed, the noise which is tiny and difficult to be perceived by naked eyes in the background is filtered by adopting a discrete Gaussian filter function, the size of an input parameter convolution kernel is selected to be 5, and the sigma value of the corresponding Gaussian function is 1.075, and the noise in the image is greatly reduced to an acceptable degree through simple verification of a program as shown in figures 4-5. FIG. 4 is a left image of an original image, which is segmented by the original image and right-enhanced; noise in the original image is artificially visualized, and noise in the actual image still exists but is not as obvious as the noise in the image.
9. And performing linear gray scale transformation on the image subjected to Gaussian filtering by using a gray histogram to enhance the light and shade contrast of the image, so that the particle edge is easier to be segmented by a computer. The input parameters include a multiplier factor set to 0.010625 and an addend factor set to-64.
10. The image segmentation obtains the particle distribution area, the segmentation mode is a threshold segmentation mode which is simple and convenient to select and has high speed and efficiency, and due to the fact that the noise of the image after filtering processing and image enhancement is small and the contrast difference of the light and the shade is obvious, the binary image of the particle area can be obtained very easily and intuitively. The input parameters Mingray and maxggray for the threshold segmentation are set to 100 and 620.
11. The divided particle regions are still a whole, the size of the shape area and the central position of the region are extracted from the binary image of the whole particle connected domain, the size of the area is mainly determined by the number of pixels of the binary image of the particle regions, and the central position of the region mainly refers to the central row-column coordinate index value of the particle regions.
12. Calculating the influence weight of the area S of the particle region, the index value m of the row coordinate center of the particle region and the index value n of the column coordinate center of the particle region on the uniformity of the particle flow distribution by using a coefficient of variation method as follows 1 ，W 2 、W 3 The calculation formulas are respectively shown as (1), (2) and (3).
13. Performing standard normalization on the extracted features S, m, n by using a z-score function, wherein m is * 、n * And are not drawn again similarly. As can be seen from fig. 7, the mean value of the processed data is 0, the standard deviation is 1, the processed data substantially conforms to the standard normal distribution, and the flow distribution of the particles can be very intuitively reflected. Recording the processed data as S * 、m * 、n * The data processing equations are shown in (4), (5), (6) and fig. 7.
14. And (3) calculating the particle flow coefficient to be P and deriving data, wherein the whole calculation and data processing process is still carried out in Halcon as shown in formula (7), and the final output parameter of the Halcon program is P as shown in figure 8. The uniformity of the particle flow distribution was evaluated based on the calculated particle flow coefficient P, and the flow distribution state showed a tendency of fluctuation and abnormality if P is larger, and a tendency of smoothness and uniformity if P is smaller, as shown in fig. 6.
15. As shown in fig. 9, the result of P from the derived particle flow coefficient indicates that there is an anomaly in the 848 frames of image data, and the comparison image can find that there is really a sudden change in flow around this time period, which results in an anomaly in which the flow is not uniform.
16. The overall particle flow distribution evaluation and particle flow coefficient P data recording time takes an average time of around 2.78 seconds in total, with a maximum test time of no more than 5 seconds.
17. As shown in fig. 10, the start-stop aperture control structure includes: (1) a singlechip controller; (2) DuPont wire; (3) a steering engine; and (4) an aperture rotation control mechanism. The main function of the structure is to control the pore size of the granules flowing out of the bottom of the hopper, and since the flow performance of the granule hopper is significantly influenced by the pore size of the bottom of the hopper, the structure is very meaningful to research factors influencing the flow performance of the granules by controlling variables.
Should open and stop each component part of control structure and have glue to connect, the back and the hopper support of steering wheel and singlechip are fixed mutually, and the dupont line makes it keep motionless by fixed buckle. Firstly, connecting a singlechip to a computer end, burning a designed steering engine control program, then connecting the singlechip to a power supply, connecting three pins of the singlechip and three fixed pins at a steering engine end by DuPont wires, wherein the pins have the corresponding relation that VCC is connected with a positive electrode of the power supply, a signal wire is connected with an IO pin of the singlechip, and the power ground is connected with a negative electrode of the steering engine according to the specified decision of the burning program. The steering wheel receives the control of singlechip signal, can take place fixed angle and rotate, and aperture rotary control joint will let the hopper bottom mouth in appointed aperture be connected bottom the hopper like this, has just so played the effect in control granule outflow aperture, adopts and does not have closed bottom mouth then to plug up granule hopper bottom, plays the effect of stopping the granule flow.
Specifically, the single chip microcomputer can be controlled by arduino, the operation is convenient and flexible, the operation is easy, the steering engine can be an SG 90s 9g steering engine, the precision is high, the response is fast, and the structure is firm and not prone to damage and deformation.
Claims (6)
1. A method of testing an apparatus for testing particle flow distribution, comprising the steps of:
(1) Image preprocessing: carrying out image decomposition, gray level transformation, noise filtration and edge sharpening enhancement on the acquired image;
(2) Image segmentation, namely obtaining a binary image of a particle region through threshold segmentation;
(3) Characteristic processing: detecting the characteristic parameters of the particle image to obtain the particle flow distribution uniformity analysis;
the feature processing method specifically comprises the following steps:
(1) Determining the size of the shape area and the position of the center of the area, wherein the size of the area is mainly determined by the number of pixel points of the binary image of the particle area, and the position of the center of the area mainly refers to the row-column coordinate index value of the center of the particle area;
(2) The number of particles flowing to the image acquisition region is represented by the area of the particle region, and the particle region and the image acquisition region are in positive correlation;
(3) The index value of the center coordinates of the particle area is used for representing the smoothness of particle flow;
(4) Distributing the weight of three influence factors of the area S of the particle area, the index value m of the row coordinate center of the particle area and the index value m of the column coordinate center of the particle area by using a variation coefficient method, and normalizing the three;
(5) The standard particle flow coefficient is P, and the particle flow distribution uniformity is obtained through three influencing factors;
in the distribution of the variation coefficient method, the influence weight of the area S of the particle area, the index value m of the row coordinate center of the particle area and the index value m of the column coordinate center of the particle area on the uniformity of the particle flow distribution is specifically defined as W 1 ，W 2 、W 3 ：
Wherein V 1 、V 2 、V 3 The coefficient of variation, σ, of the area of the particle region, the index value of the row coordinate center of the particle region, and the index value of the column coordinate center of the particle region 1 、σ 2 、σ 3 Respectively the data standard deviations of the area of the particle area, the index value of the row coordinate center of the particle area and the index value of the column coordinate center of the particle area,
the average value of the area of the particle area, the index value of the row coordinate center of the particle area and the index value of the column coordinate center of the particle area is used as the data average value;
normalizing the area S of the particle area, the index value m of the line coordinate center of the particle area and the index value m of the line coordinate center of the particle area by using a z-score function, and recording the processed data as S * 、m * 、n *
In the step (5), the particle flow coefficient is defined as P, and the formula is shown as (7)
P＝|S * |×W 1 +|m * |×W 2 +|n * |×W 3 #(7)
Due to S * 、m * 、n * All conform to the standard normal distribution, so when | S * |、|m * |、|n * The larger the value of | is, the more abnormal the flow distribution value is, and it can be considered that the flow state of the particles shows a non-uniform fluctuation tendency at this time, and the larger the particle flow coefficient P is at this time; when | S * |、|m * |、|n * When the value of | is smaller and close to 0, the flow distribution value is close to a constant value, the flowing state of the particles shows a uniform and stable trend, and the flowing coefficient P of the particles is smaller;
the structure of the device adopted by the detection method comprises a particle flow system, an image acquisition system and an image processing system, wherein the particle flow system comprises a particle storage hopper, a strong light source and a particle collecting box; the particle storage hopper is used for placing particle powder objects, the particle powder objects are fixed through the support, particles flow out of the bottom of the hopper and fall into a particle collection box placed at the bottom of the support, the background of a particle flow system is black, a strong light source is adopted for irradiation in the flow process, the particle flow state is collected through an industrial high-speed camera and a collection card, the real-time monitoring and storage are realized through image data collection software, and finally, the image processing is carried out on the collected particle flow images through image processing software.
2. The method according to claim 1, wherein the background of the particle flow system is black light absorbing flocked cloth or a background plate coated with black matt paint; the strong light source is a white light LED lamp with power being more than or equal to 40W, and the LED lamp is fixed through a long hose universal lamp holder clamp and can adjust the position and the angle of the light source; the industrial high-speed camera is fixed through the movable support, and the shooting position and angle can be adjusted.
3. The detection method of the device for detecting the particle flow distribution according to claim 1, wherein the particle flow system further comprises a start-stop aperture control structure, the structure of the start-stop aperture control structure comprises a single chip microcomputer controller, a steering engine and an aperture rotation control mechanism, the aperture rotation control mechanism is provided with a plurality of circular rings with different inner diameters, the outer diameters of the circular rings are equal to the outer diameter of the hopper outlet, the aperture rotation control mechanism is rotatably mounted on the steering engine through a rotating shaft, and the steering engine is fixedly mounted at the bottom of the support, so that the circular rings with different apertures can be in close contact with the hopper outlet of the particle storage bin; the control signal input end of the steering engine is connected with the control signal output end of the single chip microcomputer controller, and the single chip microcomputer controller controls the steering engine to rotate at a fixed angle to switch the circular ring, so that the diameter of the outlet of the particle storage hopper is controlled.
4. The method as claimed in claim 3, wherein three pins of the single chip microcomputer are connected with three fixed pins at the end of the steering engine through DuPont wires, the DuPont wires are fixed and kept still by a buckle, VCC is connected with the positive electrode of the power supply, a signal wire is connected with the IO pin of the single chip microcomputer, and the ground of the power supply is connected with the negative electrode of the steering engine.
5. The method according to claim 1, wherein the industrial high-speed camera is a 4M180-CL type Lingyun photon industrial high-speed camera, the lens is a high-resolution Lingyun photon Myutron series industrial lens, and the acquisition card is a Lingyun photon OR-X4C0-XPF00 series acquisition card; the image data acquisition software adopts the cooperation of Sapera Cam Expert and Stream7, and the image processing software is Halcon machine vision software.
6. The method as claimed in claim 1, wherein in the preprocessing, if the captured image is not a gray image, the image decomposition is performed to decompose the RGB image into three component images of R, G and B, and then the component image with obvious gray difference is selected for gray conversion or is converted into HSV color space; if the collected image is a gray image, directly filtering small noise which is difficult to be perceived by naked eyes in the background by adopting a discrete Gaussian filter function, namely scanning each pixel point in the image ROI through convolution check, replacing the value of a central pixel point corresponding to the convolution kernel by pixel weighted average gray in a convolution neighborhood of the convolution kernel, and performing better smoothing processing on the image to remove the noise; the sharp edge is enhanced to perform linear transformation on the gray value of the image so as to improve the contrast between light and shade, so that the bright position is brighter, the dark position is darker, and the segmentation precision during image segmentation is facilitated;
the image segmentation is to acquire an ROI region, that is, a region where particles are distributed, and the image after enhancing the sharpened edge is subjected to threshold segmentation, so that a binary image of the particle region can be obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110810094.9A CN113624645B (en) | 2021-07-18 | 2021-07-18 | Device and method for detecting particle flow distribution |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110810094.9A CN113624645B (en) | 2021-07-18 | 2021-07-18 | Device and method for detecting particle flow distribution |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113624645A CN113624645A (en) | 2021-11-09 |
| CN113624645B true CN113624645B (en) | 2023-01-31 |
Family
ID=78380142
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110810094.9A Active CN113624645B (en) | 2021-07-18 | 2021-07-18 | Device and method for detecting particle flow distribution |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113624645B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115239736B (en) * | 2022-09-23 | 2022-12-13 | 海安玻克超硬材料有限公司 | Method for monitoring quality of mixed material of abrasive layer for production of diamond-impregnated wheel |
| CN115294106B (en) * | 2022-09-28 | 2022-12-20 | 加乐新材料(南通)有限公司 | Method for detecting uniformity of particles in latex solution |
| CN117129413B (en) * | 2023-04-07 | 2024-04-05 | 南京工业大学 | Particle adhesion and surface energy measuring device and using method |
| CN116908051B (en) * | 2023-06-21 | 2023-12-29 | 广州市嘉雁粘合剂有限公司 | Hot melt adhesive flow property detection method |
| CN117594151A (en) * | 2023-11-27 | 2024-02-23 | 北京航空航天大学 | Method for realizing qualitative and quantitative TEP (test equipment) by utilizing image recognition |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104596429A (en) * | 2013-10-31 | 2015-05-06 | 西南科技大学 | Particle flattening thickness on-line measuring device |
| CN106447669A (en) * | 2016-04-08 | 2017-02-22 | 潍坊学院 | Circular masking-out area rate determination-based adhesive particle image concave point segmentation method |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017063128A1 (en) * | 2015-10-12 | 2017-04-20 | 深圳市大疆创新科技有限公司 | Ejection quality testing system, method and sampling auxiliary device |
| CN107870138B (en) * | 2016-09-23 | 2020-04-28 | 浙江大学 | Device for on-line detection of particle properties in fluidized bed granulation process |
| CN106447690B (en) * | 2016-09-30 | 2018-11-20 | 南京航空航天大学 | The coarse aggregate characteristic analysis method of Cold Recycled Mixture with Emulsified Asphalt |
| CN107063946B (en) * | 2016-12-23 | 2023-04-28 | 天津市三特电子有限公司 | Video green ball particle diameter instrument and green ball particle diameter state detection method thereof |
| CN108230319A (en) * | 2018-01-17 | 2018-06-29 | 北京农业智能装备技术研究中心 | For evaluating the method and device that granulated fertilizer sheds effect |
| CN108613989A (en) * | 2018-04-24 | 2018-10-02 | 南京财经大学 | A kind of detection method of paddy unsound grain |
| CN109612888B (en) * | 2018-12-12 | 2021-06-22 | 太原理工大学 | Powder mixing uniformity detection method based on image technology |
-
2021
- 2021-07-18 CN CN202110810094.9A patent/CN113624645B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104596429A (en) * | 2013-10-31 | 2015-05-06 | 西南科技大学 | Particle flattening thickness on-line measuring device |
| CN106447669A (en) * | 2016-04-08 | 2017-02-22 | 潍坊学院 | Circular masking-out area rate determination-based adhesive particle image concave point segmentation method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113624645A (en) | 2021-11-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113624645B (en) | Device and method for detecting particle flow distribution | |
| CN111507426B (en) | Non-reference image quality grading evaluation method and device based on visual fusion characteristics | |
| CN106918594B (en) | Method for on-line analysis of seed cotton color characteristics and impurities | |
| CN102253052B (en) | Grain quality on-line detection apparatus based on field programmable gate array (FPGA) | |
| CN111260640B (en) | Tree generator network gear pitting image measuring method and device based on cyclean | |
| CN105931225B (en) | A method of based on realtime graphic detection technique analyzing crystal growth form and Size Distribution | |
| CN112907519A (en) | Metal curved surface defect analysis system and method based on deep learning | |
| CN114705686A (en) | Method and system for detecting surface defects of cylindrical pot, controller and industrial personal computer | |
| CN108613989A (en) | A kind of detection method of paddy unsound grain | |
| CN114119526A (en) | Steel plate surface defect detection and identification system and method based on machine vision | |
| CN112862744B (en) | Intelligent detection method for internal defects of capacitor based on ultrasonic image | |
| CN116246174B (en) | Sweet potato variety identification method based on image processing | |
| CN116563799B (en) | Dust online monitoring method based on video monitoring | |
| CN114926407A (en) | Steel surface defect detection system based on deep learning | |
| CN114494305A (en) | On-line detection method and system for particle size of green pellets | |
| CN108596176B (en) | Method and device for identifying diatom types of extracted diatom areas | |
| CN111179275A (en) | Medical ultrasonic image segmentation method | |
| CN110378887A (en) | Screen defect inspection method, apparatus and system, computer equipment and medium | |
| CN116559111A (en) | Sorghum variety identification method based on hyperspectral imaging technology | |
| CN202083645U (en) | FPGA (field programmable gate array)-based food quality on-line detector | |
| CN108765448B (en) | Shrimp larvae counting analysis method based on improved TV-L1 model | |
| Wu et al. | Experimental analysis of image dehazing algorithms for pelletization process images | |
| CN116500052A (en) | Edible oil impurity visual detection system and application method thereof | |
| Quan et al. | Design and testing of an on-line omnidirectional inspection and sorting system for soybean seeds | |
| CN111489351A (en) | Video analysis method for measuring surface segregation degree based on stack image |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |