CN115228595B - Intelligent mining area segmentation method based on target detection - Google Patents
Intelligent mining area segmentation method based on target detection Download PDFInfo
- Publication number
- CN115228595B CN115228595B CN202210858722.5A CN202210858722A CN115228595B CN 115228595 B CN115228595 B CN 115228595B CN 202210858722 A CN202210858722 A CN 202210858722A CN 115228595 B CN115228595 B CN 115228595B
- Authority
- CN
- China
- Prior art keywords
- ore
- model
- bed surface
- belt
- target detection
- 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
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000005065 mining Methods 0.000 title claims abstract description 26
- 238000001514 detection method Methods 0.000 title claims abstract description 21
- 230000011218 segmentation Effects 0.000 title claims abstract description 19
- 238000004364 calculation method Methods 0.000 claims abstract description 17
- 238000012549 training Methods 0.000 claims abstract description 17
- 238000012545 processing Methods 0.000 claims abstract description 12
- 230000033001 locomotion Effects 0.000 claims abstract description 7
- 238000011156 evaluation Methods 0.000 claims abstract description 5
- 238000009826 distribution Methods 0.000 claims description 23
- 238000005457 optimization Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 4
- 229910052500 inorganic mineral Inorganic materials 0.000 description 38
- 239000011707 mineral Substances 0.000 description 38
- 239000012141 concentrate Substances 0.000 description 15
- 238000000926 separation method Methods 0.000 description 14
- 230000008859 change Effects 0.000 description 10
- 238000005070 sampling Methods 0.000 description 7
- 238000012360 testing method Methods 0.000 description 7
- 238000011161 development Methods 0.000 description 6
- 230000018109 developmental process Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 238000007689 inspection Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- ATJFFYVFTNAWJD-UHFFFAOYSA-N Tin Chemical compound [Sn] ATJFFYVFTNAWJD-UHFFFAOYSA-N 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 4
- UCKMPCXJQFINFW-UHFFFAOYSA-N Sulphide Chemical compound [S-2] UCKMPCXJQFINFW-UHFFFAOYSA-N 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 230000005484 gravity Effects 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000011010 flushing procedure Methods 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 238000005461 lubrication Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000002035 prolonged effect Effects 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005188 flotation Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 210000001503 joint Anatomy 0.000 description 1
- 238000007885 magnetic separation Methods 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000003032 molecular docking Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000002000 scavenging effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- AFNRRBXCCXDRPS-UHFFFAOYSA-N tin(ii) sulfide Chemical compound [Sn]=S AFNRRBXCCXDRPS-UHFFFAOYSA-N 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B03—SEPARATION OF SOLID MATERIALS USING LIQUIDS OR USING PNEUMATIC TABLES OR JIGS; MAGNETIC OR ELECTROSTATIC SEPARATION OF SOLID MATERIALS FROM SOLID MATERIALS OR FLUIDS; SEPARATION BY HIGH-VOLTAGE ELECTRIC FIELDS
- B03B—SEPARATING SOLID MATERIALS USING LIQUIDS OR USING PNEUMATIC TABLES OR JIGS
- B03B5/00—Washing granular, powdered or lumpy materials; Wet separating
- B03B5/02—Washing granular, powdered or lumpy materials; Wet separating using shaken, pulsated or stirred beds as the principal means of separation
- B03B5/04—Washing granular, powdered or lumpy materials; Wet separating using shaken, pulsated or stirred beds as the principal means of separation on shaking tables
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an intelligent mining belt segmentation method based on target detection, which comprises the following steps: acquiring data of a cradle bed surface ore belt, performing feature calibration on the acquired bed surface ore belt information, and performing model training and evaluation on the data to obtain a target detection model algorithm; the camera collects data of the ore belt of the cradle bed surface in real time and transmits the collected data to the software central platform; the software center platform performs preliminary clear processing on the image; transmitting the pictures processed by the algorithm to a trained model, evaluating the picture information by the model, and outputting the pictures calibrated by the system; the algorithm obtains the moving direction and the moving distance through accurate calculation, the software center platform sends a moving direction and moving distance instruction to the motion control system, and the control system drives the motor to drive the ore receiving plate to move to the designated position. The intelligent shaking table ore dressing belt identification and segmentation method is a brand new algorithm and has a good effect.
Description
Technical Field
The invention belongs to the technical field of gravity separation and particularly relates to an intelligent ore belt segmentation method based on target detection.
Background
China has abundant mineral resources, but whenever the mineral resources are important basic resources which are indispensable in the fields of economic development and military of China. Along with the continuous progress of scientific technology, industrial automation and intellectualization are the development trend which is not blocked in the future, most industries in China face industrial upgrading, mineral industries inevitably face problems of adjustment of industrial structures, equipment upgrading and the like, and under the large trend of the development of the times, only the pace of the development of the times is followed, so that the mineral industries can provide powerful support for mineral resource requirements in China.
Mineral separation modes are mainly divided into four major categories of gravity separation, magnetic separation, flotation and electric separation according to the composition components of minerals and the properties of the minerals. The shaking bed mineral separation is one of important equipment in the reselection, and is widely applied at home and abroad. Minerals are sorted on the bed surface of the shaking table in the form of ore pulp, and are mainly the result of the combination of the type of the bed bars, the asymmetric movement of the bed surface and the transverse flushing on the bed surface. The bed surface presents a plurality of fan-shaped ore zones such as a concentrate zone, a secondary concentrate zone, a middling zone, a tailing zone and the like, so that minerals with different grades are separated.
At present, the technology of cradle mineral separation is mature, but the automation level of mineral separation is still at a relatively low level, and in the process of cradle mineral separation, the position of a mineral strip of a bed surface, the width of the mineral strip and the color of the mineral strip are affected to different degrees due to the influence of the change of the field mineral feeding amount, the mineral feeding concentration, the granularity of the mineral feeding and the grade of raw mineral, so that the position of the mineral strip is changed in real time, an operator on site inspection must observe the characteristic change of the mineral strip of the bed surface through naked eyes, and then the position of a butt joint ore plate is correspondingly adjusted according to the working experience of the operator, so that the sorting purpose of different mineral strips is achieved, and the concentrate grade required by a factory is achieved. The traditional mode of adjusting the ore receiving plate has high adjustment frequency of the position of the receiving plate, and the labor intensity of inspection workers is high; because of the working difference of different personnel, the mineral separation index is unstable, the smooth development of mineral separation work is difficult to ensure, and the working quality and the working efficiency are affected.
The present invention has been made in view of this.
Disclosure of Invention
The invention aims to solve the technical problem of overcoming the defects of the prior art and providing an intelligent mining area segmentation method based on target detection. In order to solve the technical problems, the invention adopts the basic conception of the technical scheme that:
An intelligent mining area segmentation method based on target detection comprises the following steps:
Step 1, collecting data of a cradle bed surface ore belt, carrying out feature calibration on the collected data of the cradle bed surface ore belt, and carrying out model training and evaluation on the data to obtain a target detection model algorithm, wherein the model training comprises the following steps:
step 101, building a model, inputting x, and obtaining distribution parameters of a plurality of hidden variables q (t) through a first cnn; inputting a group of t values obtained according to q (t) distribution, and obtaining a distribution parameter of a final picture p (x) through a second cnn;
step 102, proving that the calculation of the hidden variable model is unbiased;
Step 103, model optimization, inputting x, and obtaining distribution parameters sd and mean of a plurality of hidden variables q (t) through a first cnn;
Obtaining a set of z values by a standard Gaussian distribution p (z); then a group of t values are obtained through z×sd+mean;
inputting a t value, and obtaining a distribution parameter of a final picture p (x) through a second cnn;
Step 2, the software central platform sends an instruction to the field camera, acquires the data of the ore belt of the cradle bed surface in real time, and transmits the acquired data to the software central platform through the 4G gateway;
Step 3, the software central platform carries out corresponding algorithm processing on the transmitted bed surface ore belt image so as to obtain a clear bed surface ore belt image;
Step 4, transmitting the pictures processed by the algorithm to a trained model, evaluating the picture information by the model, and outputting the pictures calibrated by the system;
step 5, judging whether the ore belt is changed relative to the previous round after the model algorithm is calculated, and if so, carrying out step 6; if no, performing step 7;
step 6, the algorithm obtains the moving direction and the moving distance through accurate calculation, the software central platform sends a moving direction and a moving distance instruction to the motion control system, and the control system drives the motor to drive the ore receiving plate to move to a designated position;
and 7, maintaining the original state of the equipment.
Further, the formula adopted in the step 101 model construction is that
qi(ti)＝N(m(xi,φ),diag(s2(xi,φ)))
Further, the step 102 proves that the calculation of the hidden variable model is unbiased by adopting the formula as
Further, the model optimization in step 103 adopts the formula of
p(εi)＝N(0,I)
Further, the information collected in the step 2 comprises the mining band width, the mining band color and the mining band position distribution characteristics.
Further, the camera is installed above the cradle bed through a camera support, and the camera support is installed on the base.
Further, the step 4 of evaluating the picture information includes comparing and identifying the on-site bed surface mine belt image transmitted to the platform by the camera according to the model training data, and calibrating the identification points.
Further, the model training further comprises the step of carrying out multipoint connection on the identification points, and an extension line is made towards the direction of the cradle bed tail according to the connection trend, and is prolonged to the bed tail.
Further, the software central platform sets for 2min to perform primary image acquisition on the on-site cradle bed surface ore belt, and performs corresponding ore receiving plate adjustment.
By adopting the technical scheme, compared with the prior art, the invention has the following beneficial effects.
According to the invention, through a model of target detection and identification, image information of the mine belt surface obtained in real time is transmitted to a software central platform through a camera system, and the information is calculated and judged through an algorithm to judge whether the position of the ore receiving plate is adjusted so as to achieve good ore selection work. The intelligent shaking table ore dressing belt identification and segmentation method is a brand new algorithm and has a good effect; the method can accurately divide the ore belt of the ore concentrate, secondary ore concentrate, middling and tailing of the shaking table surface, and cannot be influenced by site environment factors such as light rays.
The invention utilizes the rapid transmission of the network and the rapid processing of the algorithm to the image information, has high calculation speed, can almost achieve the synchronization with the information of the ore belt of the on-site cradle bed surface, and meets the on-site mineral separation requirement. The error that leads to the fact because the manpower judgement is reduced, improves shaking table work efficiency, and job quality, guarantees that ore dressing work goes on smoothly.
The following describes the embodiments of the present invention in further detail with reference to the accompanying drawings.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application. It is evident that the drawings in the following description are only examples, from which other drawings can be obtained by a person skilled in the art without the inventive effort. In the drawings:
FIG. 1 is a schematic diagram of a system logic framework of the present invention;
FIG. 2 is a schematic of the workflow of the present invention;
FIG. 3 is a schematic front view of a shaker of the present invention;
FIG. 4 is a schematic top view of the shaker of the present invention;
FIG. 5 is a schematic diagram of an algorithm model set-up of the present invention;
FIG. 6 is a schematic diagram showing comparison of the grade of the first-day machine and the artificial ore receiving according to the embodiment of the invention;
FIG. 7 is a comparison of the grade of the next-day machine and the grade of the artificial ore according to the embodiment of the invention;
FIG. 8 is a comparison of machine and artificial ore receiving grades for the third day according to the embodiment of the invention.
In the figure: 1-bed head; 2-a mineral feeding tank; 3-bed surface; 4-a water supply tank; 5-a slope adjusting machine; 6-a lubrication system; 7-a bed bar; 8-a motor; 9-a base; 10-camera mount; 11-video camera.
It should be noted that these drawings and the written description are not intended to limit the scope of the inventive concept in any way, but to illustrate the inventive concept to those skilled in the art by referring to the specific embodiments.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, which are used for illustrating the present invention but are not intended to limit the scope of the present invention.
In the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper", "lower", "front", "rear", "left", "right", "vertical", "inner", "outer", etc. are based on the directions or positional relationships shown in the drawings, are merely for convenience of describing the present invention and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and thus should not be construed as limiting the present invention.
In the description of the present invention, it should be noted that, unless explicitly specified and limited otherwise, the terms "mounted," "connected," and "connected" are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium. The specific meaning of the above terms in the present invention will be understood in specific cases by those of ordinary skill in the art.
As shown in fig. 1 to 8, the intelligent mining area segmentation method based on target detection provided by the invention comprises the following steps:
Step 1, collecting data of a cradle bed surface ore belt, carrying out feature calibration on the collected data of the cradle bed surface ore belt, and carrying out model training and evaluation on the data to obtain a target detection model algorithm, wherein the model training comprises the following steps:
step 101, building a model, inputting x, and obtaining distribution parameters of a plurality of hidden variables q (t) through a first cnn; inputting a group of t values obtained according to q (t) distribution, and obtaining a distribution parameter of a final picture p (x) through a second cnn;
qi(ti)＝N(m(xi,φ),diag(s2(xi,φ)))
If data is transferred between two cnn models by parameters only, the neural network model is basically similar to the common autocoder neural network model, and the formula is as follows:
qi(ti)＝N(m(xi,φ),diag(s2(xi,φ)))
Ifs(x)＝0then
step 102, prove that the calculation of the hidden variable model is an unbiased process:
Step 103, model optimization, inputting x, and obtaining distribution parameters sd and mean of a plurality of hidden variables q (t) through a first cnn;
Obtaining a set of z values by a standard Gaussian distribution p (z); then a group of t values are obtained through z×sd+mean;
inputting a t value, and obtaining a distribution parameter of a final picture p (x) through a second cnn;
model optimization has the advantage that standard distribution p (z) without training is stripped out; sd, mean, to be trained are only in the linear expression:
p(εi)＝N(0,I)
utilizing a first cnn model to provide hidden variables, and judging whether abnormal values exist or not based on the hidden variables; and obtaining a new picture by utilizing a second cnn model through transforming and inputting the hidden variable parameter t.
Step 2, the software central platform sends an instruction to the field camera, acquires the ore belt data of the cradle bed surface in real time, comprises the characteristics of the ore belt bandwidth, the ore belt color and the ore belt position distribution, and transmits the acquired data to the software central platform through the 4G gateway; wherein the camera is arranged above the shaking table bed through a camera bracket which is arranged on the base.
Step 3, the software central platform carries out corresponding algorithm processing on the transmitted bed surface ore belt image so as to obtain a clear bed surface ore belt image;
And 4, transmitting the pictures processed by the algorithm to a trained model, comparing and identifying the field bed surface mine belt images transmitted to the platform by the camera according to model training data, outputting identification images, and calibrating identification points. Because the ore dressing of the shaking table is that ore pulp is supplied to the bed surface and water is added, the condition of light reflection is unavoidable, algorithm improvement is carried out on the problem, the identification points are connected in a multi-point mode, and an extension line is made to the direction of the tail of the shaking table according to the trend of the connection, and the extension line is prolonged to the tail of the shaking table.
Step 5, judging whether the ore belt is changed relative to the previous round after the model algorithm is calculated, and if so, carrying out step 6; if no, performing step 7;
and 6, the algorithm gives a moving direction and a moving distance after accurate calculation, the moving distance is accurate to millimeter, the pivot platform in software sends an instruction to the motion control system, and the control system drives the motor to drive the ore receiving plate to move to a designated position. Through the understanding of the on-site ore dressing condition and the ore belt change frequency, the system software central platform sets 2min to control the camera to collect the pictures of the ore belt on the surface of the on-site cradle bed once, the algorithm is calculated once, the ore belt can send corresponding instructions when the upper wheel has the change software central platform, and then the direction and the distance of the ore plate are adjusted.
And 7, maintaining the original state of the equipment.
The intelligent mining belt segmentation method is the core of the patent, the image processing plays an important role in the whole intelligent segmentation method, mainly processes picture information transmitted back on site, extracts mining belt information of a shaking table surface, extracts and analyzes characteristics such as mining belt width, mining belt color, mining belt position distribution and the like in real time, and directly influences the mining belt segmentation precision on the processing precision of the mining belt image of the bed surface. Because shaking table is in cyclic reciprocating's jolt motion always in normal course of working, the camera is installed in shaking table tail's top, but the camera is fixed, will lead to even the ore deposit area does not change like this, but the bed surface is reciprocating motion always, and the position of the ore deposit area boundary photo that the camera draws also has corresponding change, causes certain error for the testing result.
Aiming at the problem that a camera cannot be static relative to a bed surface, a machine learning method is utilized to solve the problem, and based on training data, a machine learning algorithm is developed, wherein the algorithm applies learned information to invisible data to make predictions or other types of decisions. The pattern that interprets data better than the original data itself is called a feature in the data. And carrying out mining area data acquisition, model construction and initialization on the bed surface in the early stage. And (3) performing feature calibration on the collected bed surface ore belt information, performing model training and evaluation on the data to obtain a target detection model, and finally applying the target detection model.
Before the image segmentation is in a normal working state, the on-site shaking table ore belt information is required to be acquired, then the acquired shaking table ore belt information is subjected to characteristic calibration, the calibrated image is formed into a data set, then a model is built for model training, and the trained model is subjected to test and then is subjected to certain adjustment, so that the using effect of the model is more perfect. The computer and the 4G gateway can be connected through a network cable, and can also be communicated by using a SIM card, and the 4G gateway is connected with the camera through the network cable.
When the intelligent system starts to work, firstly, the system sends an instruction to a wireless gateway, the gateway sends a command signal to a camera, the camera collects images of the ore strips on the cradle bed surface, the collected images are transmitted to a system control platform through a 4G gateway, the system is provided with an image preprocessing module, firstly, the collected information images of the cradle strips are preprocessed, then the preprocessed images of the cradle strips are transmitted to a trained data model, the model carries out data comparison, analysis, calibration and line drawing on the image information, concentrate strips, secondary concentrate strips, middling strips and tailing strips are divided one by one, then an output result image is output, and the system sends an instruction to a motor control system according to the change condition of the ore strips, so that corresponding adjustment is carried out on a docking plate to finish the index set by a factory. The system can set the image acquisition frequency according to the requirement so as to achieve the purpose of real-time monitoring.
The cradle for mineral separation is a common device for separating fine ore. Is usually composed of a bed surface, a frame and a transmission mechanism. In addition, a water flushing tank, a mineral feeding tank and the like are arranged, the whole bed surface is supported or lifted by a frame, the bed surface is arranged on the frame, and a slope adjusting machine is arranged at the bottom of the frame. The inside of the bed head is provided with a lubrication system and a motor as a transmission mechanism, and the bed surface is provided with a bed bar. The bed head is connected with one end of the bed head. The utility model discloses a mineral shaking table is common equipment, and the current shaking table equipment that adopts this time also does not carry out technical improvement to it, and its concrete connection structure and theory of operation are not repeated here.
Example 1
As shown in fig. 1 to 8, according to the method for intelligently dividing the ore belt based on target detection in the embodiment, intelligent table application research is implemented in a mineral processing center of cloud tin large bin tin ore in Yunnan, the daily treatment capacity of oxidized ore in a workshop is 2000t/d, and a plurality of tables 200 are used; treating 4000t/d of sulfide ores on a daily basis and shaking 288 sheets; the method is used for rough concentration, scavenging and fine concentration treatment of the metal ore tin ore. Because the quantity of the shaking tables is quite large, the change frequency of the ore belt is high, a large number of inspection workers are required to conduct continuous inspection to adjust the position of the ore receiving plate so as to achieve the desired index of the factory, and therefore the labor force of the inspection workers and the running cost of enterprises are greatly increased.
The industrial test site is selected in a large-sized Tun cloud tin sulfide ore working section, 3-section 3-18 shaking tables are used as test shaking tables, 220v power supplies are selected to supply power to the equipment, a camera is arranged above the tail of the shaking tables, the installation height is 2m, real-time picture acquisition and image transmission and processing are carried out on the shaking table surfaces, then a drive motor drives an ore receiving plate to move to a target position, and full-automatic ore receiving of the 3-18 shaking tables is achieved.
1. Real-time monitoring efficiency
Through the acquisition discovery of the real-time data of the on-site shaking table in the earlier stage, the system is set to calculate the time interval for 3min each time according to the change condition of the on-site ore belt, so that a camera automatically acquires real-time pictures at each interval for 3min and transmits the pictures to a software platform, and then the positions of the butt-joint ore plates are adjusted through calculation so as to meet the ore receiving requirement.
2. Ore strip image processing
The shaking table for installing the equipment is the shaking table for the rough concentration section of the sulphide ore, so that only secondary concentrate and middling are required to be separated according to the field requirement, concentrate and secondary concentrate are connected together and then enter the next step for recleaning, the secondary concentrate ore zone and middling ore zone have great differences in chromatic aberration, the middling ore zone is colored white, the secondary concentrate ore zone is colored black, since ten thousands of pictures are collected in the early stage for characteristic calibration, four points are calibrated in the middle of each picture, the model training is carried out, the image acquisition is carried out on the field shaking table according to the setting, the image is simply processed and input into the model, and the segmentation map of the secondary concentrate ore zone and the middling ore zone is output.
3. Mineral separation index comparison
In order to compare the difference of mineral concentration index under intelligent shaking table equipment and the manual operation condition, experimental two adjacent shaking tables of roughing operation workshop section of selecting, 3-17 shaking tables are manual operation shaking tables, and 3-18 are intelligent equipment installation shaking tables, and two shaking tables are given the ore by same rod mill, and raw ore taste is the same, and give the ore granularity the same, and the ore pulp concentration is the same. Under the condition of ensuring that all conditions are the same, 3 days of industrial test sampling are carried out, and each sampling interval is 1 hour, and 13 groups of sampling are carried out. Sample comparison data are shown in table 1. The test results are shown in fig. 6, 7 and 8.
Comparison of concentrate grades in tables 1 3-17 and 3-18
As can be seen from the Sn grade line plot of 3 data samples: ① According to the three line diagrams, under the same condition, the Sn grade fluctuation range of the mechanical ore receiving is much smaller than that of the artificial ore receiving, and the data fluctuation phase difference is most obvious in the first day; ② The three broken line diagrams show that the Sn grade of the machine ore receiving is more stable than that of the artificial ore receiving; ③ As can be seen from the whole of the three line diagrams, the Sn grade of the machine ore-receiving is higher than that of the artificial ore-receiving.
In order to further prove that the Sn grade of the machine-connected ore is higher than that of the artificial-connected ore and is stable than that of the artificial-connected ore, three-day sampling data are calculated through the average value and the variance, and the calculation formula is as follows:
the results of the data calculations for the 3-17 manually operated shaker and the 3-18 intelligent shaker are shown in Table 2:
comparison of the results of the calculation of the mean and variance of the sampled data in tables 2 3-17 and 3-18
The mean and variance calculations for three days of sampled data can be seen: ① The first day and the third day of the calculation result of the sampling data average value are equal, and the second day and the third day are the artificial ore collection with 3-18 machine ore collection more than 3-17; ② The calculation result of the variance of the three-day sampling data is that the 3-18 machine ore collection is smaller than the 3-17 artificial ore collection. Therefore, the calculation result shows that the machine ore receiving grade is high and stable compared with the artificial ore receiving grade.
4. Conclusion(s)
According to the industrial test sampling for three continuous days, the 3-18 machine ore receiving is stable compared with the 3-17 artificial ore receiving under the same condition, and the grade of the machine ore receiving Sn is higher than that of the artificial ore receiving Sn, so that the stability and the Sn grade are higher than that of the artificial ore receiving Sn, and the production index of a factory can be achieved. The intelligent monitoring equipment for the shaking table can replace manual work, monitor the change condition of the ore strip of the shaking table surface in real time and realize the unattended mode. The boundary points of the mine belt are accurately identified through machine learning, and the identification accuracy meets the use requirement.
In the process of the rough concentration section table of the sulphide ores in the mineral processing center of the cloud tin of Yunnan, the industrial experimental study of the application of the intelligent table is implemented, and the concentrate product received by the automatic ore receiving plate of the intelligent table can meet the process requirement of the factory selection. The industrial application of intelligent shaking table mineral processing system can liberate the manual labor, realizes shaking table mineral processing intellectuality, improves shaking table mineral processing's production level and enterprise benefit, has powerful promotion effect to the development of gravity separation equipment.
The above description is only of the preferred embodiments of the present invention, and is not intended to limit the present invention in any way, although the present invention has been described in the preferred embodiments, it is not intended to limit the present invention, and any person skilled in the art will not depart from the scope of the present invention, while the technical content mentioned above can be utilized to make some changes or modifications to equivalent embodiments, any simple modification, equivalent changes and modification made to the above embodiments according to the technical substance of the present invention will still fall within the scope of the present invention.
Claims (6)
1. The intelligent mining area segmentation method based on target detection is characterized by comprising the following steps of:
Step 1, collecting data of a cradle bed surface ore belt, carrying out feature calibration on the collected data of the cradle bed surface ore belt, and carrying out model training and evaluation on the data to obtain a target detection model algorithm, wherein the model training comprises the following steps:
Step 101, building a model, inputting x, and obtaining distribution parameters of a plurality of hidden variables q (t) through a first cnn; inputting a group of t values obtained according to q (t) distribution, and obtaining a distribution parameter of a final picture p (x) through a second cnn;
The formula adopted by the model construction is that
；
；
；
Step 102, proving that the calculation of the hidden variable model is unbiased, and adopting the formula as
；
Step 103, model optimization, inputting x, and obtaining distribution parameters sd and mean of a plurality of hidden variables q (t) through a first cnn;
Obtaining a set of z values by a standard Gaussian distribution p (z); then a group of t values are obtained through z×sd+mean;
inputting a t value, and obtaining a distribution parameter of a final picture p (x) through a second cnn;
The formula adopted by the model optimization is that
；
；
；
Step2, the software central platform sends an instruction to the field camera, acquires the data of the ore belt of the cradle bed surface in real time, and transmits the acquired data to the software central platform through the 4G gateway;
Step 3, the software central platform carries out corresponding algorithm processing on the transmitted bed surface ore belt image so as to obtain a clear bed surface ore belt image;
step 4, transmitting the pictures processed by the algorithm to a trained model, evaluating the picture information by the model, and then outputting the pictures calibrated by the system;
step 5, judging whether the ore belt is changed relative to the previous round after the model algorithm is calculated, and if so, carrying out step 6; if no, performing step 7;
Step 6, the algorithm obtains the moving direction and the moving distance through accurate calculation, the software central platform sends a moving direction and a moving distance instruction to the motion control system, and the control system drives the motor to drive the ore receiving plate to move to a designated position;
and 7, maintaining the original state of the equipment.
2. The intelligent mining strip segmentation method based on target detection according to claim 1, wherein the method comprises the following steps: the information acquired in the step 2 comprises the mining band width, the mining band color and the mining band position distribution characteristics.
3. The intelligent mining strip segmentation method based on target detection according to claim 1, wherein the method comprises the following steps: the camera is installed above the shaking table bed through a camera support, and the camera support is installed on the base.
4. The intelligent mining strip segmentation method based on target detection according to claim 1, wherein the method comprises the following steps: and 4, evaluating the picture information, namely comparing and identifying the on-site bed surface mine belt image transmitted to the platform by the camera according to the model training data, and calibrating the identification points.
5. The intelligent mining strip segmentation method based on target detection according to claim 4, wherein the method comprises the following steps: the model training further comprises the step of carrying out multipoint connection on the identification points, and an extension line is made towards the direction of the cradle bed tail according to the connection trend, and is extended to the bed tail.
6. The intelligent mining strip segmentation method based on target detection according to claim 4, wherein the method comprises the following steps: and setting the software center platform for 2min, carrying out primary image acquisition on the ore belt of the on-site cradle bed surface, and carrying out corresponding ore receiving plate adjustment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210858722.5A CN115228595B (en) | 2022-07-20 | 2022-07-20 | Intelligent mining area segmentation method based on target detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210858722.5A CN115228595B (en) | 2022-07-20 | 2022-07-20 | Intelligent mining area segmentation method based on target detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115228595A CN115228595A (en) | 2022-10-25 |
| CN115228595B true CN115228595B (en) | 2024-10-25 |
Family
ID=83672517
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210858722.5A Active CN115228595B (en) | 2022-07-20 | 2022-07-20 | Intelligent mining area segmentation method based on target detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115228595B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117409009A (en) * | 2023-12-15 | 2024-01-16 | 长沙矿冶研究院有限责任公司 | Real-time sorting method for dry magnetic separation particles based on UNet |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103785532A (en) * | 2014-02-18 | 2014-05-14 | 云南锡业集团有限责任公司研究设计院 | Method for automatically monitoring tin ore table beneficiation |
| CN103810500A (en) * | 2014-02-25 | 2014-05-21 | 北京工业大学 | Place image recognition method based on supervised learning probability topic model |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108681743B (en) * | 2018-04-16 | 2019-12-06 | 腾讯科技（深圳）有限公司 | Image object recognition method and device and storage medium |
| CN113269675B (en) * | 2021-05-18 | 2022-05-13 | 东北师范大学 | Time-variant data time super-resolution visualization method based on deep learning model |
-
2022
- 2022-07-20 CN CN202210858722.5A patent/CN115228595B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103785532A (en) * | 2014-02-18 | 2014-05-14 | 云南锡业集团有限责任公司研究设计院 | Method for automatically monitoring tin ore table beneficiation |
| CN103810500A (en) * | 2014-02-25 | 2014-05-21 | 北京工业大学 | Place image recognition method based on supervised learning probability topic model |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115228595A (en) | 2022-10-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110567974A (en) | Cloud artificial intelligence based surface defect detection system | |
| Kaartinen et al. | Machine-vision-based control of zinc flotation—a case study | |
| CN115228595B (en) | Intelligent mining area segmentation method based on target detection | |
| CN109115785A (en) | A kind of casting grinding quality determining method, device and its application method | |
| CN111753912A (en) | Coal slime flotation clean coal ash content prediction method based on deep learning | |
| CN110070526A (en) | Defect inspection method based on the prediction of deep neural network temperature figure | |
| CN102284428A (en) | Corn real-time detection and grading machine vision system | |
| CN111157537A (en) | Strip steel defect detection device and system based on bionic machine vision | |
| CN110394307B (en) | Tobacco leaf automated inspection letter sorting system | |
| CN117764443B (en) | Internet of things-based full life cycle green quality control method for big data concrete | |
| CN106979952A (en) | A kind of algae detection and Statistics of Density devices and methods therefor based on machine vision technique | |
| CN110243732B (en) | Whole-course ore pulp grinding granularity online detection system with granularity range of 2-500um | |
| CN114858806A (en) | Cigarette quality inspection data analysis system and method | |
| CN109726777A (en) | PCB appearance detection system and detection method Internet-based | |
| CN112485257A (en) | Method and device for detecting roller surface abrasion of roller press based on machine vision | |
| CN202114018U (en) | Machine vision system for grain real-time detection and grading | |
| CN207222339U (en) | A kind of intelligent controller of stack high-frequency vibration undersize sieve | |
| CN112801211A (en) | Power grid defect accurate identification method based on image identification technology | |
| CN108380525A (en) | Bastard coal device for visual identification and method | |
| CN114378002A (en) | Hericium erinaceus grading system and grading method based on machine vision | |
| CN112580185A (en) | Prediction method for intermediate grade content of ore grinding product of vertical stirring mill | |
| CN113304869B (en) | Automatic identification and access method and device for shaking table ore belt | |
| CN214894918U (en) | Belted steel defect detecting device based on bionical machine vision | |
| CN116740449A (en) | Shaving form detection method and system based on AI (advanced technology attachment) computer vision technology | |
| CN117034538A (en) | Design method for structural counterweight iron slag concrete mix proportion |
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 | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Zhang Wenkang Inventor after: Wang Chunjing Inventor after: Yuan Linxun Inventor after: Liu Dan Inventor after: Yu Longzhou Inventor before: Zhang Wenkang Inventor before: Wang Chunjing Inventor before: Yuan Linxun Inventor before: Liu Dan Inventor before: Zhou Gongqiang Inventor before: He Qixue Inventor before: Lei Chao Inventor before: Yu Longzhou |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |