[go: up one dir, main page]

CN110738674B - Texture feature measurement method based on particle density - Google Patents

Texture feature measurement method based on particle density Download PDF

Info

Publication number
CN110738674B
CN110738674B CN201911005627.5A CN201911005627A CN110738674B CN 110738674 B CN110738674 B CN 110738674B CN 201911005627 A CN201911005627 A CN 201911005627A CN 110738674 B CN110738674 B CN 110738674B
Authority
CN
China
Prior art keywords
bubble
particle
gray
point
degrees
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
Application number
CN201911005627.5A
Other languages
Chinese (zh)
Other versions
CN110738674A (en
Inventor
唐朝晖
罗金
张国勇
李涛
范影
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Central South University
Original Assignee
Central South University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Central South University filed Critical Central South University
Publication of CN110738674A publication Critical patent/CN110738674A/en
Application granted granted Critical
Publication of CN110738674B publication Critical patent/CN110738674B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/40—Analysis of texture
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/90—Determination of colour characteristics
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Mathematical Analysis (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Computing Systems (AREA)
  • Multimedia (AREA)
  • Algebra (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Analysing Materials By The Use Of Radiation (AREA)

Abstract

一种基于颗粒密集度的纹理特征度量方法，在泡泡浮选领域，本发明提出一种基于颗粒密集度的纹理特征度量方法，此方法基于现场设置的数字图像采集系统提取泡泡图像，提出了颗粒区域的概念，并对泡泡表面的颗粒区域进行准确提取，将所提取的颗粒区域按中心点位置进行分级，继而对颗粒区域之间的密集程度进行度量，定义了纹理特征颗粒密集度，用以反映整幅图像的纹理特征。有效弥补了传统的纹理特征提取方法没有考虑泡泡表面颗粒的缺陷，从而可以更准确的判断工况并有效指导加药。

Figure 201911005627

A texture feature measurement method based on particle density, in the field of bubble flotation, the present invention proposes a texture feature measurement method based on particle density, the method extracts bubble images based on a digital image acquisition system set up on site, and proposes The concept of particle area is introduced, and the particle area on the surface of the bubble is accurately extracted, and the extracted particle area is classified according to the position of the center point, and then the density between the particle areas is measured, and the texture feature particle density is defined. , to reflect the texture features of the whole image. It effectively makes up for the defect that the traditional texture feature extraction method does not consider the particles on the surface of the bubble, so that it can more accurately judge the working conditions and effectively guide the dosing.

Figure 201911005627

Description

Texture feature measurement method based on particle density
Technical Field
The invention belongs to the technical field of froth flotation, and particularly relates to a texture feature measurement method in a zinc flotation process.
Background
The froth flotation is a mineral separation method widely used at home and abroad, and the method can effectively separate target minerals according to the difference between the hydrophilicity and the hydrophobicity of the surfaces of the minerals. In the froth flotation process, target minerals and gangue symbiotic with the target minerals are ground into particles with proper sizes and then sent into a flotation tank, different mineral particle surface properties are adjusted by adding medicaments, and meanwhile, the particles are continuously stirred and blown in the flotation process, so that a large number of bubbles with characteristic information such as different sizes, forms and textures are formed in ore pulp, useful mineral particles are adhered to the surfaces of the bubbles, the bubbles carry the mineral particles to rise to the surfaces of the flotation tank to form bubble layers, and gangue minerals are left in the ore pulp, and therefore mineral separation is achieved. Because the flotation process has long flow, an undefined internal mechanism, a plurality of influence factors, a plurality of related variables, severe nonlinearity, incapability of on-line detection of process indexes and the like, the flotation process mainly depends on manual visual observation of the bubble state on the surface of the flotation tank to complete on-site operation, the production mode has strong subjectivity, objective evaluation and cognition of the flotation bubble state are difficult to realize, the situations of frequent fluctuation of flotation production indexes, severe loss of mineral raw materials, large medicament consumption, low resource recovery rate and the like are caused, particularly in the present day that high-grade mineral resources are increasingly deficient, the flotation mineral source components are complex, the mineral taste is low, and the manual operation of flotation production is difficult to effectively carry out. Machine vision is applied to the flotation process, the flotation bubble image is analyzed by using a digital image processing technology, objective description of the bubble state can be realized, and then the relation between the bubble characteristic parameters and the process indexes is further searched and analyzed, so that the production automation of the flotation process is promoted. Flotation bubbles present a special texture state along with the difference of the flotation state, the texture of a bubble image is the comprehensive reflection of the roughness, the contrast and the viscosity of the bubble surface, the method is closely related to flotation production operation variables such as dosage, ventilation quantity and the like, the concentrate grade, the tailing content and other flotation production indexes, the current bubble texture information extraction method mainly extracts local features, the problems of insufficient extraction precision, no consideration of bubble surface particles in the texture extraction process and the like exist, the working condition is difficult to accurately reflect, in fact, a plurality of ores or magazine small particles are often attached to the bubble surface to cause the roughness of the bubble surface, the quantity and the distribution density of the small particles are closely related to the zinc concentrate grade, the problem of the bubble surface particles is not considered in the previous research, and a new texture feature measurement method based on the particle density is provided, according to the method, the bubble image is extracted based on the digital image acquisition system arranged on the site, then the particle areas on the surface of the bubbles are accurately extracted, the density among the particle areas is quantitatively measured, the new texture feature particle density is defined to reflect the texture features of the whole image, the limitation of the traditional texture feature extraction method is effectively avoided, and therefore the working condition is more accurately judged and the medicine adding is effectively guided.
Disclosure of Invention
The invention aims to provide a texture feature measuring method based on particle concentration, in flotation production, the surface texture of flotation bubbles is important visual information reflecting the ore grade, is closely related to the flotation working condition, and directly reflects the mineralization degree of a bubble layer. Aiming at the problem that the grain on the surface of bubbles is not considered in the extraction of the texture features of the existing flotation bubble images, the grain density of the texture feature measurement method is defined. According to the method, firstly, bubble images are segmented, interesting bubbles are extracted, then particle areas on the surfaces of the bubbles are extracted, and analysis and explanation are carried out on how the particle density reflects the working condition, so that the method can more accurately regulate and control the mineral grade and guide flotation production.
The adopted technical scheme comprises the following steps:
the method comprises the following steps: collecting bubble videos of zinc flotation by using a flotation field image acquisition system, converting the bubble videos into continuous images, and performing data preprocessing on the acquired zinc flotation image data as follows:
1) rejecting erroneous data that exceeds a normal variation threshold;
2) removing incomplete data;
step two: converting the bubble image from an RGB color image into a gray image to obtain a gray matrix A of the image:
Figure GDA0002755139770000021
egfthe gray value corresponding to each pixel point in the gray image is represented, wherein g is equal to N, f is equal to N, N is equal to (400, 800).
Step three: in the bubble image, the conventional bubble shape is smooth in surface, the highlight point is positioned at the top end of the convex curved surface of a single bubble, the highlight point area presents the minimum gray value, the gray value gradually increases downwards by taking the highlight point as the center, and the maximum gray value is reached when the highlight point reaches the bubble boundary; in fact, the bubbles are often attached with some ore or magazine small particles, which cause the surface of the bubbles to be rough and uneven, and the quantity and distribution density of the small particles are related to the dosage and the zinc concentrate grade;
firstly, dividing bubbles, dividing the bubbles by a watershed method to obtain h individual bubbles, storing a gray matrix of each bubble to obtain a gray matrix set B ═ B of each bubble1,b2,b3,...,bλ,...,bh}，bλIs the lambda-th bubble gray value matrix, h is the total number of the bubbles after segmentation, in the single bubble image set after the segmentation, screens out the bubble that the bubble size is greater than 1200 pixel values, be the region of interest promptly, for the bubble after the screening, replaces the grey scale value of this bubble highlight point part with the grey scale mean value of single bubble, obtains to detect bubble gray matrix set C ═ { C ═ of bubble gray matrix set C ═ of { C ═ of detecting1,c2,c3,...,cε,...,cKK is the number of bubbles meeting the bubble size requirement.
Step four: detection of particle area:
1. in the non-particle area, the bubble surface is smooth, and the change range of the gray value is in the gradual change range;
2. the change of the gray value in the particle area exceeds the gradual change range;
the particle region was extracted using the following steps:
s1: defining a searching mode of a particle area for the bubbles after segmentation and screening: for bubble cεTaking eight directions of 0 degrees, 45 degrees, 90 degrees, 135 degrees, 180 degrees, 225 degrees, 270 degrees and 315 degrees, forming an array by gray values of all pixel points in any direction in the bubble, wherein the maximum width of a single bubble is a finite value, and the row number of the leftmost pixel point of the bubble in the gray matrix is HmThe column number of the rightmost pixel point of the bubble in the gray matrix is HnH from bubble gray matrix at search timemThe columns start searching from left to right and from top to bottom from the bubble upper half boundary, the initial search direction is 270 °;
s2: marking a trip point:
(1) setting the threshold value of gray gradation to [0,8 ]]Firstly, starting to search from the 270-degree direction by taking the leftmost pixel point of the boundary of the upper half part of the bubble as a starting point, taking the step length as 1, defining the modulus of the difference value of the next gray value and the current gray value as a gray gradient value in the 270-degree direction, comparing the gray gradient value with a threshold value, and marking the position of the current gray value as a jump point d if the gray gradient value exceeds the threshold value1The position is expressed in cartesian coordinates as: (x)1,y1) The column H where the first trip point is located0Is denoted as the left boundary of the particle region, where Hm<...<H0<H1<H2<...<Hk<Hk+1<Hk+2<...<Hn，H1Is H0The first column, and so on, in this search from H0Column search to the right through Hk+2When the column is not searched for a trip point, Hk+1Column is the right boundary of this granular area;
(2) from H1Is listed to HkBetween columns, each column is searched from top to bottom in 270 deg. direction to obtain two jump points, and in H th search1When the column is arranged, the jumping points of the ray directions of 90 degrees and 270 degrees are recorded as d2And d3Taking the midpoint of the positions of the two jump points as a position center divergence point o1(ii) a Diverging from the central point o by position1Starting the search to both sides in the ray directions of 0 DEG and 180 DEG as starting points, and marking the position d of the nearest jump point4，d5(ii) a Starting to search from the position divergence central point to two sides in the ray directions of 45 degrees and 225 degrees, and marking the position d of the nearest jump point6，d7(ii) a Starting to search from the position divergence center point to two sides in the ray directions of 135 degrees and 315 degrees, and marking the position d of the nearest jump point8，d9；
S3: extraction of particle region:
for bubble cεFrom H1The position of the column diverges from the center point o18 trip points, H, are obtained2,H3,...,HkThe position of the column diverges from the center point o2,o3,...,okRespectively obtaining 8 jumping points, the particle area obtains 8k jumping points in total, if 8k<24, the region is noise, and is not considered, when 8k is more than or equal to 24, the 8k transition points are connected in sequence to obtain a particle region which is marked as Dr(r∈a)。
Step five: determination of the position of the center point of the particle region:
for the particle region DrThe number of the transition points of the grain region is recorded as trI.e. the zone boundary has trA vertex with coordinates of (x)i,yi),i＝1,2,...,trVertex, point
Figure GDA0002755139770000031
And vertex (x)1,y1) Similarly, the area of the particle region is shown by the following formula:
Figure GDA0002755139770000032
center point coordinates of particle region
Figure GDA0002755139770000033
As shown in the following formula:
Figure GDA0002755139770000034
Figure GDA0002755139770000035
repeating the step four to search out all particle areas in the gray value matrix of the single bubble in the bubble set to be detected, and regarding the bubble cεCounting the number L of particle regionsεThen, the coordinates of the center point of each particle region are determined
Figure GDA0002755139770000036
Defining primary neighborhood region, secondary neighborhood region and other neighborhood regions by the straight line distance between the central points of different particle regions, and aligning the particle region PuU ∈ a, defined as follows:
Figure GDA0002755139770000041
wherein v belongs to a and v is not equal to u;
marking particle region PuThe number of the first-level neighborhood region is quThe number of the second-level neighborhood region is suAnd the number of other neighborhood regions is wuThen there is qu+su+wu＝Lε-1; the weights of the number of the first-level neighborhood regions, the number of the second-level neighborhood regions and the number of the other neighborhood regions are respectively set to be 0.6, 0.3 and 0.1, so that the bubble c is treatedεDefining the grain density Z of the texture feature quantityεThe expression is as follows:
Figure GDA0002755139770000042
the particle concentration G of the whole image is defined as shown in the following formula:
Figure GDA0002755139770000043
step six: the working condition is judged according to the particle concentration:
Figure GDA0002755139770000044
when the bubble is in the state I, the surface texture of the bubble is fine, the medicament is excessive, and the mineral particles carried in the bubble exceed the carrying capacity of the bubble to break the bubble in large quantity, so that the medicament waste is serious and the concentrate grade is low;
when the flotation device is in the state II, a proper amount of flotation reagents are used, the flotation performance is good, and the flotation production efficiency is highest;
when the pulp is in the state III, the pulp viscosity is low, the addition amount of the medicament is insufficient, the ore content of bubbles is less, the water content is high, and the concentrate grade is low.
In the second step, the bubble image is converted into a gray level image from an RGB color image, and a gray level matrix A of the image is obtained, wherein N belongs to (400, 800).
The invention defines new texture feature particle density to reflect the texture feature of the whole image, effectively avoids the limitation of the traditional texture feature extraction method, and effectively overcomes the influence of the uneven illumination phenomenon of the flotation field on the texture feature extraction, thereby more accurately judging the working condition and effectively guiding the dosing.
Drawings
FIG. 1 is a flow chart of a texture feature measurement method based on particle concentration.
Fig. 2 is a schematic diagram of the region of the particles extracted at S3 in step four.
Detailed Description
FIG. 1 is a flow chart of the present invention.
The method comprises the following steps: collecting bubble videos of zinc flotation by using a flotation field image acquisition system, converting the bubble videos into continuous images, and performing data preprocessing on the acquired zinc flotation image data as follows:
1) rejecting erroneous data that exceeds a normal variation threshold;
2) removing incomplete data;
step two: converting the bubble image from an RGB color image into a gray image to obtain a gray matrix A of the image:
Figure GDA0002755139770000051
egfthe gray value corresponding to each pixel point in the gray image is represented, wherein g is equal to N, f is equal to N, N is equal to (400, 800).
Step three: in the bubble image, the conventional bubble shape is smooth in surface, the highlight point is positioned at the top end of the convex curved surface of a single bubble, the highlight point area presents the minimum gray value, the gray value gradually increases downwards by taking the highlight point as the center, and the maximum gray value is reached when the highlight point reaches the bubble boundary; in fact, the bubbles are often attached with some ore or magazine small particles, which cause the surface of the bubbles to be rough and uneven, and the quantity and distribution density of the small particles are related to the dosage and the zinc concentrate grade;
firstly, dividing bubbles, dividing the bubbles by a watershed method to obtain h individual bubbles, storing a gray matrix of each bubble to obtain a gray matrix set B ═ B of each bubble1,b2,b3,...,bλ,...,bh}，bλIs the lambda-th bubble gray value matrix, h is the total number of the bubbles after segmentation, in the single bubble image set after the segmentation, screens out the bubble that the bubble size is greater than 1200 pixel values, be the region of interest promptly, for the bubble after the screening, replaces the grey scale value of this bubble highlight point part with the grey scale mean value of single bubble, obtains to detect bubble gray matrix set C ═ { C ═ of bubble gray matrix set C ═ of { C ═ of detecting1,c2,c3,...,cε,...,cKK is the number of bubbles meeting the bubble size requirement.
Step four: detection of particle area:
1. in the non-particle area, the bubble surface is smooth, and the change range of the gray value is in the gradual change range;
2. the change of the gray value in the particle area exceeds the gradual change range;
the particle region was extracted using the following steps:
s1: defining a searching mode of a particle area for the bubbles after segmentation and screening: for bubble cεTaking eight directions of 0 degrees, 45 degrees, 90 degrees, 135 degrees, 180 degrees, 225 degrees, 270 degrees and 315 degrees, forming an array by gray values of all pixel points in any direction in the bubble, wherein the maximum width of a single bubble is a finite value, and the row number of the leftmost pixel point of the bubble in the gray matrix is HmThe column number of the rightmost pixel point of the bubble in the gray matrix is HnH from bubble gray matrix at search timemThe column starts searching from left to right and from top to bottom starting from the bubble top half boundary, the initial search direction is 270 °.
S2: marking a trip point:
(1) setting the threshold value of gray gradation to [0,8 ]]Firstly, starting to search from the 270-degree direction by taking the leftmost pixel point of the boundary of the upper half part of the bubble as a starting point, taking the step length as 1, defining the modulus of the difference value of the next gray value and the current gray value as a gray gradient value in the 270-degree direction, comparing the gray gradient value with a threshold value, and marking the position of the current gray value as a jump point d if the gray gradient value exceeds the threshold value1The position is expressed in cartesian coordinates as: (x)1,y1) The column H where the first trip point is located0Is denoted as the left boundary of the particle region, where Hm<...<H0<H1<H2<...<Hk<Hk+1<Hk+2<...<Hn，H1Is H0The first column, and so on, in this search from H0Column search to the right through Hk+2When the column is not searched for a trip point, Hk+1Column is the right boundary of this granular area;
(2) from H1Is listed to HkBetween columns, each column from top to bottomSearching in 270 deg direction to obtain two jump points, and searching in the H th1When the column is arranged, the jumping points of the ray directions of 90 degrees and 270 degrees are recorded as d2And d3Taking the midpoint of the positions of the two jump points as a position center divergence point o1(ii) a Diverging from the central point o by position1Starting the search to both sides in the ray directions of 0 DEG and 180 DEG as starting points, and marking the position d of the nearest jump point4，d5(ii) a Starting to search from the position divergence central point to two sides in the ray directions of 45 degrees and 225 degrees, and marking the position d of the nearest jump point6，d7(ii) a Starting to search from the position divergence center point to two sides in the ray directions of 135 degrees and 315 degrees, and marking the position d of the nearest jump point8，d9；
S3: extraction of particle region:
for bubble cεFrom H1The position of the column diverges from the center point o18 trip points, H, are obtained2,H3,...,HkThe position of the column diverges from the center point o2,o3,...,okRespectively obtaining 8 jumping points, the particle area obtains 8k jumping points in total, if 8k<24, the region is noise, and is not considered, when 8k is more than or equal to 24, the 8k transition points are connected in sequence to obtain a particle region which is marked as Dr(r ∈ a), FIG. 2 is a schematic diagram of the extracted particle region.
Step five: determination of the position of the center point of the particle region:
for the particle region DrThe number of the transition points of the grain region is recorded as trI.e. the zone boundary has trA vertex with coordinates of (x)i,yi),i＝1,2,...,trVertex, point
Figure GDA0002755139770000061
And vertex (x)1,y1) Similarly, the area of the particle region is shown by the following formula:
Figure GDA0002755139770000062
center point coordinates of particle region
Figure GDA0002755139770000063
As shown in the following formula:
Figure GDA0002755139770000064
Figure GDA0002755139770000065
repeating the step four to search out all particle areas in the gray value matrix of the single bubble in the bubble set to be detected, and regarding the bubble cεCounting the number L of particle regionsεThen, the coordinates of the center point of each particle region are determined
Figure GDA0002755139770000066
Defining primary neighborhood region, secondary neighborhood region and other neighborhood regions by the straight line distance between the central points of different particle regions, and aligning the particle region PuU ∈ a, defined as follows:
Figure GDA0002755139770000071
wherein v belongs to a and v is not equal to u;
marking particle region PuThe number of the first-level neighborhood region is quThe number of the second-level neighborhood region is suAnd the number of other neighborhood regions is wuThen there is qu+su+wu＝Lε-1; the weights of the number of the first-level neighborhood regions, the number of the second-level neighborhood regions and the number of the other neighborhood regions are respectively set to be 0.6, 0.3 and 0.1, so that the bubble c is treatedεDefining the grain density Z of the texture feature quantityεThe expression is as follows:
Figure GDA0002755139770000072
the particle concentration G of the whole image is defined as shown in the following formula:
Figure GDA0002755139770000073
step six: the working condition is judged according to the particle concentration:
Figure GDA0002755139770000074
when the bubble is in the state I, the surface texture of the bubble is fine, the medicament is excessive, and the mineral particles carried in the bubble exceed the carrying capacity of the bubble to break the bubble in large quantity, so that the medicament waste is serious and the concentrate grade is low;
when the flotation device is in the state II, a proper amount of flotation reagents are used, the flotation performance is good, and the flotation production efficiency is highest;
when the pulp is in the state III, the pulp viscosity is low, the addition amount of the medicament is insufficient, the ore content of bubbles is less, the water content is high, and the concentrate grade is low.
In the second step, the bubble image is converted into a gray level image from an RGB color image, and a gray level matrix A of the image is obtained, wherein N belongs to (400, 800).
The invention defines new texture feature particle density to reflect the texture feature of the whole image, effectively avoids the limitation of the traditional texture feature extraction method, and effectively overcomes the influence of the uneven illumination phenomenon of the flotation field on the texture feature extraction, thereby more accurately judging the working condition and effectively guiding the dosing.

Claims (6)

1. A texture feature measurement method based on particle density is characterized by comprising the following steps:
the method comprises the following steps: collecting bubble videos of zinc flotation at historical moments by using a flotation field image collecting system, converting the bubble videos into multi-frame continuous images, and performing data preprocessing on collected zinc flotation image data;
step two: converting the bubble image from RGB color image to gray image to obtain gray matrix A of image
Figure FDA0002755139760000011
egfExpressing the gray value corresponding to each pixel point in the gray image, wherein g belongs to N, and f belongs to N;
step three: dividing the bubbles, dividing the bubbles by a watershed method to obtain h individual bubbles, storing the gray matrix of each bubble to obtain a gray matrix set B ═ B of each bubble1,b2,b3,...,bλ,...,bh}，bλIs the lambda-th bubble gray value matrix, and in the divided single bubble image set, the bubble with the size larger than 1200 pixel values is screened out, and is marked as C ═ C1,c2,c3,...,cε,...,cKK is the number of single bubbles meeting the size requirement of the bubbles;
step four: detection of particle regions
The surface of the bubble in the non-particle area is smooth, the change range of the gray value is in the gradual change range, and the change of the gray value in the particle area exceeds the gradual change range, and the particle area is extracted by adopting the following steps:
s1: defining a searching mode of a particle area for the bubbles after segmentation and screening;
s2: marking a jumping point in the searching process of the particle area;
s3: after marking the jumping points, extracting particle areas;
step five: repeating the step four to search out all particle areas in the gray value matrix of the single bubble in the bubble set to be detected, and for the bubbles c subjected to segmentation and screeningεCounting the number L of particle regionsεThen, the coordinates of the center point of each particle region are determined
Figure FDA0002755139760000013
Defining primary neighborhood region, secondary neighborhood region and other neighborhood regions by the straight line distance between the central points of different particle regions, and aligning the particle region PuU ∈ a, defined as follows:
Figure FDA0002755139760000012
wherein v belongs to a and v is not equal to u;
marking particle region PuThe number of the first-level neighborhood region is quThe number of the second-level neighborhood region is suAnd the number of other neighborhood regions is wuThen there is qu+su+wu＝Lε-1; the weights of the number of the first-level neighborhood regions, the number of the second-level neighborhood regions and the number of the other neighborhood regions are respectively set to be 0.6, 0.3 and 0.1, so that the bubble c is treatedεDefining the grain density Z of the texture feature quantityεThe expression is as follows:
Figure FDA0002755139760000021
the particle concentration G of the whole image is defined as shown in the following formula:
Figure FDA0002755139760000022
step six: the working condition is judged according to the particle concentration:
Figure FDA0002755139760000023
when the bubble is in the state I, the surface texture of the bubble is fine, the medicament is excessive, and the mineral particles carried in the bubble exceed the carrying capacity of the bubble, so that the bubble is greatly crushed, the medicament waste is serious, and the concentrate grade is low;
when the flotation agent is in the state II, the flotation agent is proper, the flotation performance is good, and the flotation production efficiency is high;
when the pulp is in the state III, the pulp viscosity is low, the addition amount of the medicament is insufficient, the ore content of bubbles is less, the water content is high, and the concentrate grade is low.
2. The method of claim 1, wherein the texture feature measurement method based on particle density comprises: the second step comprises the following steps: converting the bubble image from the RGB color image into a gray image to obtain a gray matrix A of the image, wherein N belongs to (400,800).
3. The method of claim 1, wherein the texture feature measurement method based on particle density comprises: the step four S1 includes: defining a searching mode of a particle area for the bubbles after segmentation and screening:
for bubble cεTaking eight directions of 0 degree, 45 degrees, 90 degrees, 135 degrees, 180 degrees, 225 degrees, 270 degrees and 315 degrees, at the bubble cεThe gray values of all the pixel points in any direction form an array, the maximum width of a single bubble is a finite value, and the column number of the pixel point on the leftmost side of the bubble in the gray matrix is HmThe column number of the rightmost pixel point of the bubble in the gray matrix is HnH from bubble gray matrix at searchmThe column starts searching from left to right and from top to bottom starting from the bubble top half boundary, the initial search direction is 270 °.
4. The method of claim 3, wherein the texture feature measurement method based on particle concentration comprises: in step S2, for the search method for dividing the sorted bubble definition particle area, the process of marking the jumping point is as follows:
1) setting the threshold value of gray gradation to [0,8 ]]Firstly, starting to search from the 270-degree direction by taking the leftmost pixel point of the boundary of the upper half part of the bubble as a starting point, taking the step length as 1, defining the modulus of the difference value between the next gray value and the current gray value as a gray gradient value in the 270-degree direction, and comparing the gray gradient value with a threshold valueIf the gray level gradient value exceeds the threshold value, the current gray level position is marked as a jumping point d1The position is expressed in cartesian coordinates as: (x)1,y1) The column H where the first trip point is located0Is denoted as the left boundary of the particle region, where Hm<...<H0<H1<H2<...<Hk<Hk+1<Hk+2<...<Hn，H1Is H0The first column, and so on, in this search from H0Column search to the right through Hk+2When the column is not searched for a trip point, Hk+1Column is the right boundary of this granular area;
2) from H1Is listed to HkBetween columns, each column is searched from top to bottom in 270 deg. direction to obtain two jump points, and in H th search1When the column is arranged, the jumping points of the ray directions of 90 degrees and 270 degrees are recorded as d2And d3Taking the midpoint of the positions of the two jump points as a position center divergence point o1(ii) a Diverging from the central point o by position1Starting the search to both sides in the ray directions of 0 DEG and 180 DEG as starting points, and marking the position d of the nearest jump point4，d5(ii) a Starting to search from the position divergence central point to two sides in the ray directions of 45 degrees and 225 degrees, and marking the position d of the nearest jump point6，d7(ii) a Starting to search from the position divergence center point to two sides in the ray directions of 135 degrees and 315 degrees, and marking the position d of the nearest jump point8，d9。
5. The method of claim 4, wherein the texture feature measurement method based on particle concentration comprises: in step four S3, the extraction process of the particle region is as follows:
for bubble cεFrom H1The position of the column diverges from the center point o18 trip points, H, are obtained2,H3,...,HkThe position of the column diverges from the center point o2,o3,...,okIf 8 jumping points are obtained respectively, the particle area can obtain 8k jumping points in total, if 8k<24 then the area is noiseTaking the points out of consideration, when 8k is more than or equal to 24, the 8k transition points are connected in sequence to obtain a particle area which is marked as Dr,r∈a。
6. The method of claim 5, wherein the texture feature measurement method based on particle concentration comprises: in the fifth step, the determining process of the position of the center point of the particle region is as follows:
for the particle region DrThe number of the transition points of the grain region is recorded as trI.e. the zone boundary has trA vertex with coordinates of (x)i,yi),i＝1,2,...,trVertex, point
Figure FDA0002755139760000031
And vertex (x)1,y1) Similarly, the area of the particle region is shown by the following formula:
Figure FDA0002755139760000032
center point coordinates of particle region
Figure FDA0002755139760000035
As shown in the following formula:
Figure FDA0002755139760000033
Figure FDA0002755139760000034
CN201911005627.5A 2019-07-22 2019-10-22 Texture feature measurement method based on particle density Active CN110738674B (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910661088 2019-07-22
CN2019106610884 2019-07-22

Publications (2)

Publication Number Publication Date
CN110738674A CN110738674A (en) 2020-01-31
CN110738674B true CN110738674B (en) 2021-03-02

Family

ID=69270787

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911005627.5A Active CN110738674B (en) 2019-07-22 2019-10-22 Texture feature measurement method based on particle density

Country Status (1)

Country Link
CN (1) CN110738674B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113112432A (en) * 2021-05-13 2021-07-13 广州道一科学技术有限公司 Method for automatically identifying image strips

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000068672A1 (en) * 1999-05-05 2000-11-16 Antti Niemi Method and apparatus for monitoring and analyzing the surface of floated material
CN101339659A (en) * 2008-08-22 2009-01-07 北京矿冶研究总院 A Rule-Based Region Growing Image Segmentation Method
CN102855492A (en) * 2012-07-27 2013-01-02 中南大学 Classification method based on mineral flotation foam image
CN103632156A (en) * 2013-12-23 2014-03-12 中南大学 Multi-scale neighboring dependence matrix-based method for extracting froth image texture characteristics
CN105405149A (en) * 2015-11-20 2016-03-16 中南大学 Composite texture feature extraction method for flotation froth image
CN106257498A (en) * 2016-07-27 2016-12-28 中南大学 Zinc flotation work condition state division methods based on isomery textural characteristics
US9652841B2 (en) * 2015-07-06 2017-05-16 International Business Machines Corporation System and method for characterizing NANO/MICRO bubbles for particle recovery
CN107705283A (en) * 2017-06-14 2018-02-16 华北理工大学 Particle and bubble hit detection method based on Otsu image segmentation
CN108647722A (en) * 2018-05-11 2018-10-12 中南大学 A kind of zinc ore grade flexible measurement method of Kernel-based methods size characteristic
CN108986077A (en) * 2018-06-19 2018-12-11 东北大学 Flotation froth operating mode's switch method based on dual-tree complex wavelet domain symbiosis augmented matrix
CN110288591A (en) * 2019-07-02 2019-09-27 中南大学 Discrimination Method of Zinc Flotation Conditions Based on Improved Adaptive Multi-Population Genetic Algorithm

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2567055C (en) * 2004-05-20 2013-11-19 Mcmaster University Method for controlling the appearance of products and process performance by image analysis

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2000068672A1 (en) * 1999-05-05 2000-11-16 Antti Niemi Method and apparatus for monitoring and analyzing the surface of floated material
CN101339659A (en) * 2008-08-22 2009-01-07 北京矿冶研究总院 A Rule-Based Region Growing Image Segmentation Method
CN102855492A (en) * 2012-07-27 2013-01-02 中南大学 Classification method based on mineral flotation foam image
CN103632156A (en) * 2013-12-23 2014-03-12 中南大学 Multi-scale neighboring dependence matrix-based method for extracting froth image texture characteristics
US9652841B2 (en) * 2015-07-06 2017-05-16 International Business Machines Corporation System and method for characterizing NANO/MICRO bubbles for particle recovery
CN105405149A (en) * 2015-11-20 2016-03-16 中南大学 Composite texture feature extraction method for flotation froth image
CN106257498A (en) * 2016-07-27 2016-12-28 中南大学 Zinc flotation work condition state division methods based on isomery textural characteristics
CN107705283A (en) * 2017-06-14 2018-02-16 华北理工大学 Particle and bubble hit detection method based on Otsu image segmentation
CN108647722A (en) * 2018-05-11 2018-10-12 中南大学 A kind of zinc ore grade flexible measurement method of Kernel-based methods size characteristic
CN108986077A (en) * 2018-06-19 2018-12-11 东北大学 Flotation froth operating mode's switch method based on dual-tree complex wavelet domain symbiosis augmented matrix
CN110288591A (en) * 2019-07-02 2019-09-27 中南大学 Discrimination Method of Zinc Flotation Conditions Based on Improved Adaptive Multi-Population Genetic Algorithm

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Process working condition recognition based on the fusion of morphological and pixel set features of froth for froth flotation;Xiaoli Wang 等;《Minerals Engineering》;20181130;第128卷;第17-26页 *
一种新的浮选泡沫图像纹理特征提取方法;桂卫华 等;《中国科技论文》;20120415(第4期);第277-281页 *

Also Published As

Publication number Publication date
CN110738674A (en) 2020-01-31

Similar Documents

Publication Publication Date Title
CA3006240C (en) A stepwise refinement detection method for pavement cracks
Duchesne et al. Multivariate image analysis in the process industries: A review
CN113435460B (en) A recognition method for bright crystal granular limestone images
Chen et al. Bottom-up image detection of water channel slope damages based on superpixel segmentation and support vector machine
CN102902974B (en) Image based method for identifying railway overhead-contact system bolt support identifying information
CN102750703A (en) Remote sensing image airport automatic detecting method based on linear cutting
CN105260805B (en) A kind of antimony ore grade flexible measurement method selectively merged based on isomery grader
CN112964712A (en) Method for rapidly detecting state of asphalt pavement
CN101996405A (en) Method and device for rapidly detecting and classifying defects of glass image
Köse et al. Statistical methods for segmentation and quantification of minerals in ore microscopy
CN104850854A (en) Talc ore product sorting processing method and talc ore product sorting system
YANG et al. Bubble size estimation using interfacial morphological information for mineral flotation process monitoring
Galdames et al. Classification of rock lithology by laser range 3D and color images
CN102004916A (en) Image characteristic extraction system and method
CN110728676B (en) A Texture Feature Measurement Method Based on Sliding Window Algorithm
CN110728253B (en) A Texture Feature Measurement Method Based on Grain Roughness
CN104751474A (en) Cascade quick image defect segmentation method
CN110738674B (en) Texture feature measurement method based on particle density
CN116523865A (en) A method for predicting workpiece surface roughness
CN110728677B (en) A Texture Roughness Definition Method Based on Sliding Window Algorithm
CN108986077A (en) Flotation froth operating mode&#39;s switch method based on dual-tree complex wavelet domain symbiosis augmented matrix
CN108647722A (en) A kind of zinc ore grade flexible measurement method of Kernel-based methods size characteristic
Xu et al. Multi-model soft measurement method of the froth layer thickness based on visual features
CN107516315A (en) A monitoring method for slag discharge of roadheader based on machine vision
CN117557548A (en) A sand gradation detection method based on machine vision

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