CN107876429A - A kind of waste non-ferrous metals automatic sorting system based on machine vision - Google Patents
A kind of waste non-ferrous metals automatic sorting system based on machine vision Download PDFInfo
- Publication number
- CN107876429A CN107876429A CN201711268404.9A CN201711268404A CN107876429A CN 107876429 A CN107876429 A CN 107876429A CN 201711268404 A CN201711268404 A CN 201711268404A CN 107876429 A CN107876429 A CN 107876429A
- Authority
- CN
- China
- Prior art keywords
- unit
- visual identity
- transmission
- transmission range
- color
- 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.)
- Granted
Links
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B07—SEPARATING SOLIDS FROM SOLIDS; SORTING
- B07C—POSTAL SORTING; SORTING INDIVIDUAL ARTICLES, OR BULK MATERIAL FIT TO BE SORTED PIECE-MEAL, e.g. BY PICKING
- B07C5/00—Sorting according to a characteristic or feature of the articles or material being sorted, e.g. by control effected by devices which detect or measure such characteristic or feature; Sorting by manually actuated devices, e.g. switches
- B07C5/34—Sorting according to other particular properties
- B07C5/342—Sorting according to other particular properties according to optical properties, e.g. colour
- B07C5/3422—Sorting according to other particular properties according to optical properties, e.g. colour using video scanning devices, e.g. TV-cameras
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B07—SEPARATING SOLIDS FROM SOLIDS; SORTING
- B07C—POSTAL SORTING; SORTING INDIVIDUAL ARTICLES, OR BULK MATERIAL FIT TO BE SORTED PIECE-MEAL, e.g. BY PICKING
- B07C5/00—Sorting according to a characteristic or feature of the articles or material being sorted, e.g. by control effected by devices which detect or measure such characteristic or feature; Sorting by manually actuated devices, e.g. switches
- B07C5/36—Sorting apparatus characterised by the means used for distribution
- B07C5/363—Sorting apparatus characterised by the means used for distribution by means of air
- B07C5/365—Sorting apparatus characterised by the means used for distribution by means of air using a single separation means
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B07—SEPARATING SOLIDS FROM SOLIDS; SORTING
- B07C—POSTAL SORTING; SORTING INDIVIDUAL ARTICLES, OR BULK MATERIAL FIT TO BE SORTED PIECE-MEAL, e.g. BY PICKING
- B07C2501/00—Sorting according to a characteristic or feature of the articles or material to be sorted
- B07C2501/0036—Sorting out metallic particles
Abstract
The present invention relates to a kind of automatic sorting system, including feed unit, transmission unit, visual identity unit, sensing unit, feed separation unit, central control unit, transmission unit includes the first transmission range and the second transmission range, second transmission range is provided with visual identity region, feed unit is by material transmission to the first transmission range, sensing unit includes the first sensor for being arranged on the second transmission range, visual identity unit carries out visual identity to the material for entering visual identity region, sensing unit, visual identity unit, feed separation unit is electrically connected with central control unit, feed separation unit includes separator and storage device, separator includes jet module, storage device is provided with multiple stock chests along material transferring direction, jet module sorts the material of identification into corresponding stock chest.The present invention, which improves the accuracy rate of old metal identification and sorting capability, reduces hand labor, operation and cost of serving, reduces environmental pollution.
Description
Technical field
The invention belongs to old metal sorting technology field, and in particular to a kind of waste non-ferrous metals based on machine vision are automatic
Separation system.
Background technology
In this way, people are increasing to the consumption figure of metals resources, and Domestic Resources consumption is rare, and dependence on external supply is strong.
China there are about 5,000,000 tons of scrap iron and steel every year, more than 20 ten thousand tons of non-ferrous metal does not recycle.Therefore, old metal material returns
Receive to utilize and there is great economic benefit, social benefit and environmental benefit.But metal is often with element in old metal material
Or the form of alloy is present, the mode such as generally use melting carries out secondary resource recovery.Old metal material is carried out before melting
Corresponding sorting process, the difficulty of secondary metal purifying technique can be reduced, improve organic efficiency and profit.
In the prior art, domestic separation system is mainly the physical characteristic according to metal, using selection by winnowing, magnetic separation, whirlpool electricity
Classification etc., as the ratio increase of non-ferrous metal in old metal material to waste non-ferrous metals such as copper, aluminium, zinc, it is necessary to carry out thin
Change sorting, above-mentioned common method for separating can not adaption demand, it is necessary to new automation sorting solution.
In summary, cost of labor can be reduced by needing offer one kind badly, improve the accuracy rate and sorting capability of old metal identification
Automatic sorting device.
The content of the invention
It is an object of the invention to provide one kind can reduce cost of labor, improve the accuracy rate and sorting capability of old metal identification
Automatic sorting device.
Above-mentioned purpose is to be achieved through the following technical solutions：A kind of waste non-ferrous metals automatic sorting system based on machine vision
System, including feed unit, transmission unit, visual identity unit, sensing unit, feed separation unit and central control unit,
The sensing module, visual identity are electrically connected with unit and feed separation unit with central control unit, the transmission unit edge
Material transferring direction includes the first transmission range and the second transmission range, and the transmission speed of first transmission range is less than the second transmission range
Transmission speed, second transmission range is preset with visual identity region, and the feed separation unit includes jet-propelled separation dress
Put and storage device, the jet-propelled separator include being arranged on the jet module of the second transmission range end, the storing
Device is provided with multiple stock chests along material transferring direction, and the feed unit transmits the first of material transmission to transmission unit
Area, the sensing unit comprise at least the first sensor for being arranged on the second transmission range, and the first sensor is used for inductor
Material passes to central control unit into visual identity region and by induced signal, and the central control unit controls visual identity
Unit carries out visual identity to the material for entering visual identity region, and identification information is passed to center by the visual identity unit
Control unit, the central control unit identification information of reception is handled after compared with the visual information to prestore
Match somebody with somebody, and jet direction and/or the whiff pressure of jet module, institute are adjusted according to the jet-propelled separator of comparison match output control
Jet module is stated to sort the material of identification into corresponding stock chest.
Present invention is mainly used for old metal material of the sorting of waste non-ferrous metals after process upstream is handled to pass through feed
The first transmission range of transmission unit is sent to after being separated after unit, the second transmission range is then transmit to, due to the second transmission range
Certain speed difference be present with the first transmission range, material can be pulled open further between the material of rear after entering the second transmission range
Induced signal is passed to central control unit by distance, sensing unit after visual identity region detection to material signal, center
Control unit controls visual identity unit to carry out visual identity to the material for entering visual identity region, and visual identity unit will be known
Other information transmission carries out image preprocessing, color to central control unit, the central control unit to the identification information of reception
Matched after feature extraction compared with the visual information to prestore, and according to comparison match output control and to feed separation unit
Related adjustment is carried out, according to adjustment result, separation module sorts the material of identification into corresponding stock chest.The present invention can have
Effect completes the sorting of waste non-ferrous metals, and improving the accuracy rate of old metal identification and sorting capability reduces hand labor, operation
And cost of serving, reduce environmental pollution, effectively facilitate the development of recycling economy.
Preferably, further technical scheme is：The central control unit is handled the identification information of reception
Process the step of comprising at least image preprocessing and color feature extracted, the central control unit is by the color characteristic of extraction
Matched afterwards compared with the color characteristic of the visual information to prestore.Old metal surface of material generally has abrasion, scratches, to figure
As needing to be pre-processed before feature extraction, it is contemplated that the actual conditions of engineering, select 5*5 circular shuttering adding window intermediate value herein
Wave filter is handled waste material image.Image of the image that video camera is gathered in industry spot except including waste material
Outside, also it is mingled with the background image of complexity, can be interfered to the feature extraction of target image, preferably carrying out image segmentation will be useless
Image outside metalliferous material is rejected.
Preferably, further technical scheme is：Described image pre-treatment step includes image filtering and handles and scheme
As dividing processing, described image dividing processing realizes the mutually segmentation of target image and background image.
Preferably, further technical scheme is：Image filtering processing, described image point are carried out using median filter
Processing is cut using the color image segmentation method under YCrCb color spaces.Because the image that visual identity unit collects is RGB
True color image, preferably using Color image filtering method, median filtering method is not only in terms of impulse disturbances and swept noise are filtered out
Significant effect, and the influence that the image edge detailss that linear filter can be overcome to bring well obscure.
Preferably, further technical scheme is：The step of described image dividing processing, is as follows：
(1) picture of visual identity unit photographs is transformed into YCrCb color spaces；
(2) color feature value of pixel is extracted；
(3) the color feature value selected seed of the pixel of extraction in step (2) is utilized；
(4) seed region is grown up；
(5) judge whether that all pixels point is processed, be, the seed region in step (4) is merged, otherwise
It is back to the step of step (4) continues seed region growth.
Preferably, further technical scheme is：YCrCb color spaces are utilized in the step of described image dividing processing
The redundancy of RGB color color channel is reduced, the hue and luminance after splitting using image represents colouring information, and Y is brightness
Value, is obtained, calculation formula is by color RGB color R, G, B value weighting：Y=irR+igG+ibB, wherein, ir, ig, ibIt is to add
Weight factor, tone can be represented that calculation formula is as follows by given Y and three different aberration Cr, Cg, Cb：
Cr=R-Y；Cg=G-Y；Cb=B-Y.
Preferably, further technical scheme is：Useless based on machine vision according to claim 2 has coloured gold
Belong to automatic sorting system, it is characterised in that the selection of the color feature extracted more stable tone H and saturation degree S is carried out
Color characteristic analyze, wherein, be first HSI models by RGB model conversations by the picture of visual identity unit photographs, wherein, color
Adjust H calculation formula be： Saturation degree S
Calculated by following formula：
Preferably, further technical scheme is：Also include multilayer sense after the step process of the color feature extracted
Know the training process of device neutral net, the multilayer perceptron neural network algorithm structure includes the data processing of Multilevel method device
Structure, including input layer, output layer and hidden layer positioned there between, the training of multilayer perceptron neutral net
In journey using extraction it is multigroup including tone H and saturation degree S color characteristics as input.Wherein, the number of multilayer perceptron nerve net
Learning model is：
yi=f (xi) i=1,2 ... m
In formula, wijFor weight matrix, sjIt is input matrix, θiIt is threshold function table, f (xi) it is activation primitive, preferably select
Softmax functions are as activation primitive.The training of certain multilayer perceptron neutral net is completed before normal work.
Preferably, further technical scheme is：Input layer nerve in the training process of multilayer perceptron neutral net
First number is 2, and it is 2 to imply the number of plies, and every layer of hidden layer neuron number is 4, and output layer neuron number has according to the useless of sorting
The species setting that non-ferrous metal is.
Preferably, further technical scheme is：First transmission range, which is provided with, is used to sense material into the first biography
The second sensor of the predeterminated position in defeated area.It is in time in the purpose of the second sensor of the predeterminated position of the first transmission range
Material signal passes to central processing module, and central processing module control visual identity unit carries out visual identity preparation, works as thing
Operation is identified after entering the visual identity region of the second transmission range in material in time, improves its reaction speed；It is preferred that described first
Sensor and second sensor are fibre optical sensor.
Preferably, further technical scheme is：The jet-propelled sorting unit also includes air compressor, the spray
Gas module is connected with the air compressor by pipeline, and the jet direction of the jet module is adjustable.It is arranged such, in
The material information of Central Processing Unit feedback, can control the folding of jet module puff prot, by the use of spraying air-flow as being before used as power,
Movement locus, speed and the acceleration of material is changed, so that material falls on default different stock chests, realize automatic
The purpose of sorting.Because air is inexhaustible, pump-down process is simple, free from environmental pollution, and only needs rationally
The position of arrangement nozzle and airflow direction and pressure can realize the separation of various metals waste material, good separation effect.
Preferably, further technical scheme is：Second transmission range is separated into multiple transmission channels by dividing plate, often
The end of individual transmission channel is equipped with jet module, and the jet module is provided with multiple puff prots, the jet side of the puff prot
To adjustable.It is arranged such, on the one hand the second transmission range is separated into multiple transmission channels by dividing plate, so can as far as possible second
The material of transmission range, which separates, to be set, and is avoided being superimposed with each other, is improved visual identity effect, while can then be effectively increased the second transmission range
Width, the same time can blow to multiple materials of the same width position in end of the second transmission range, improve sorting effect
Rate；On the other hand each jet module is equipped with multiple puff prots, is so controlled by the synergic adjustment of multiple puff prots, jet
Angular adjustment effect is more preferable.
Preferably, further technical scheme is：The jet module is provided with the stepping electricity for being used to adjust jet angle
Machine and the high speed pulse valve for controlling gas channel size, the stepper motor are connected with the jet module by connecting rod.Such as
This is set, and the stepper motor, which rotates, drives linkage, adjusts the angle of the puff prot of jet module, and control air-flow
Direction and pressure are realized is blown into corresponding stock chest by different types of material.
Preferably, further technical scheme is：The feed unit includes vibratory sieve and reason material slide rail, the reason material
Slide rail is in predetermined inclination angle, and the vibratory sieve is connected with the first transmission range by managing material slide rail, and the loading of the vibratory sieve is put down
Face is higher than the transport plane of the first transmission range.In this way, by the vibration of vibratory sieve, old metal material is separated by shake, and progressively
The first transmission range is included in by managing material slide rail, can so pull open the distance between material, and material is entered into preliminary arrangement, avoids weight
Stack, visual identity effect can be improved, and then improve separating effect and efficiency.
Preferably, further technical scheme is：The visual identity unit includes kilomega network industrial camera and light source.
Using kilomega network industrial camera, can long range high speed transmission data, while can quick and precisely catch image, the setting master of light source
Ensure the stability of light in target visual identification process.
Preferably, further technical scheme is：The central control unit includes switch board and industrial computer, the work
Control machine is provided with touch-screen.Industrial computer is the core of control system, is responsible for data storage, image procossing and characteristic matching, according to figure
As result to slave computer sends control signal etc., the working condition and fortune of each part can be monitored in real time additionally by touch-screen
Market condition, sorting information is checked, set sorting pattern, while ensureing stability and durability, improve the side of man-machine interaction
Just property.
Preferably, further technical scheme is：The length of first transmission range is 1.5~3m, and transmission speed is
0.3~0.5m/s.The setting of first transmission range of scheduled transmission speed and length, the main old metal thing to after vibrator supply
Expect certain buffering, reduce influence of the vibration to institute's material visual identity as far as possible, improve the quality of shooting image, the first transmission
The transmission speed in area can not be too fast, and length is unsuitable too short, and otherwise buffering effect is bad, and material still suffers from certain shake during shooting
It is dynamic, while its transmission speed can not be too fast, length is unsuitable long, can so reduce the efficiency of separation.
Preferably, further technical scheme is：The transmission speed of second transmission range is less than 7m/s.
Preferably, further technical scheme is：The transmission unit is the transmission belt of motor driven.
Brief description of the drawings
The accompanying drawing for forming the part of the present invention is used for providing a further understanding of the present invention, schematic reality of the invention
Apply example and its illustrate to be used to explain the present invention, do not form inappropriate limitation of the present invention.
Fig. 1 and the automatic sorting device involved by one embodiment of the present invention structural representation；
Fig. 2 is a kind of figure of old metal material of the visual identity unit photographs involved by one embodiment of the present invention
Picture；
Fig. 3 is the image of old metal material involved in Fig. 2 through the pretreated image of central control unit.
In figure：
The visual identity region of 1 vibratory sieve 2 reason material slide rail, 3 transmission unit 4
The central control unit of 5 first sensor, 6 second sensor, 7 visual identity unit 8
The air compressor of 9 jet module, 10 storage device, 11 stock chest 12
The motor of 13 dividing plate 14
Embodiment
The present invention will be described in detail below in conjunction with the accompanying drawings, and the description of this part is only exemplary and explanatory, should not
There is any restriction effect to protection scope of the present invention.In addition, description of the those skilled in the art according to this document, can be right
Feature in this document in embodiment and in different embodiments carries out respective combination.
The embodiment of the present invention is as follows, such as Fig. 1, a kind of waste non-ferrous metals automatic sorting system based on machine vision, including
Feed unit, transmission unit 3, visual identity unit 7, sensing unit, feed separation unit and central control unit 8, it is described
Sensing module, visual identity are electrically connected with unit and feed separation unit with central control unit 8, and the transmission unit 3 is along thing
Material transmission direction includes the first transmission range and the second transmission range, and the transmission speed of first transmission range is less than the second transmission range
Transmission speed, second transmission range are preset with visual identity region 4, and the feed separation unit includes jet-propelled separator
And storage device 10, the jet-propelled separator include being arranged on the jet module 9 of the second transmission range end, the storing
Device 10 is provided with multiple stock chests 11 along material transferring direction, the feed unit by material transmission to transmission unit 3 first
Transmission range, the sensing unit comprise at least the first sensor 5 for being arranged on the second transmission range, and the first sensor 5 is used for
Sensing material enters visual identity region 4 and induced signal is passed into central control unit 8, and the central control unit 8 is controlled
Visual identity unit 7 processed carries out visual identity to the material for entering visual identity region 4, and the visual identity unit 7 will identify
To central control unit 8, the central control unit 8 regards information transmission after handling the identification information of reception with what is prestored
Feel that information is compared matching, and the jet side of jet module 9 is adjusted according to the jet-propelled separator of comparison match output control
To and/or whiff pressure, the jet module 9 material of identification is sorted into corresponding stock chest 11.
Present invention is mainly used for old metal material of the sorting of waste non-ferrous metals after process upstream is handled to pass through feed
The first transmission range of transmission unit 3 is sent to after being separated after unit, the second transmission range is then transmit to, due to the second transmission range
Certain speed difference be present with the first transmission range, material can be pulled open further between the material of rear after entering the second transmission range
Distance, sensing unit detects in visual identity region 4 passes to central control unit 8 after material signal by induced signal, in
Control unit 8 is entreated to control visual identity unit 7 to carry out visual identity, visual identity list to the material for entering visual identity region 4
Identification information is passed to central control unit 8 by member 7, and it is pre- that the central control unit 8 carries out image to the identification information of reception
Matched after processing, color feature extracted compared with the visual information to prestore, and according to comparison match output control and to thing
Expect that separative element carries out related adjustment, according to adjustment result, separation module sorts the material of identification to corresponding stock chest 11
In.The present invention can efficiently accomplish the sorting of waste non-ferrous metals, and the accuracy rate and sorting capability for improving old metal identification reduce
Hand labor, operation and cost of serving, reduce environmental pollution, have effectively facilitated the development of recycling economy.
On the basis of above-described embodiment, in another embodiment of the present invention, identification of the central control unit 8 to reception
The process that information is handled comprises at least the step of image preprocessing and color feature extracted, and the central control unit 8 will
Matched after the color characteristic of extraction compared with the color characteristic of the visual information to prestore.Old metal surface of material generally has mill
Damage, scratch, need to pre-process before to image characteristics extraction, it is contemplated that the actual conditions of engineering, select 5*5 circle herein
Shape template adding window median filter is handled waste material image.The image that video camera is gathered in industry spot is except including
Outside the image of waste material, also it is mingled with the background image of complexity, can interferes to the feature extraction of target image, preferably carry out
The image outside old metal material is rejected in image segmentation.
On the basis of above-described embodiment, in another embodiment of the present invention, described image pre-treatment step is filtered including image
Ripple processing and image dividing processing, described image dividing processing realize the mutually segmentation of target image and background image.
On the basis of above-described embodiment, in another embodiment of the present invention, carried out using median filter at image filtering
Reason, described image dividing processing use the color image segmentation method under YCrCb color spaces.Because visual identity unit 7 is adopted
The image integrated is preferably not only filtering out impulse disturbances as RGB true color images using Color image filtering method, median filtering method
With significant effect in terms of swept noise, and the fuzzy shadow of the image edge detailss that linear filter brings can be overcome well
Ring.
It is as follows the step of described image dividing processing in another embodiment of the present invention on the basis of above-described embodiment：
(1) picture that visual identity unit 7 is shot is transformed into YCrCb color spaces；
(2) color feature value of pixel is extracted；
(3) the color feature value selected seed of the pixel of extraction in step (2) is utilized；
(4) seed region is grown up；
(5) judge whether that all pixels point is processed, be, the seed region in step (4) is merged, otherwise
It is back to the step of step (4) continues seed region growth.
On the basis of above-described embodiment, in another embodiment of the present invention, the step of described image dividing processing in utilize
YCrCb color spaces reduce the redundancy of RGB color color channel, and the hue and luminance after being split using image represents color
Information, Y are brightness value, are obtained by color RGB color R, G, B value weighting, calculation formula is：Y=irR+igG+ibB, its
In, ir, ig, ibIt is weighted factor, tone can represent that calculation formula is such as by given Y and three different aberration Cr, Cg, Cb
Under：Cr=R-Y；Cg=G-Y；Cb=B-Y.
Under a kind of embodiment, a kind of image of old metal material of visual identity unit identification is single through center control
The front and rear comparing result of member pretreatment is as shown in Figures 2 and 3.
It is according to claim 2 to be regarded based on machine in another embodiment of the present invention on the basis of above-described embodiment
The waste non-ferrous metals automatic sorting system of feel, it is characterised in that the more stable tone H of the selection of the color feature extracted and full
Color characteristic analysis is carried out with degree S, wherein, it is first HSI moulds by RGB model conversations by the picture that visual identity unit 7 is shot
Type, wherein, tone H calculation formula be：
Saturation degree S is calculated by following formula：
On the basis of above-described embodiment, in another embodiment of the present invention, after the step process of the color feature extracted
Also include the training process of multilayer perceptron neutral net, the multilayer perceptron neural network algorithm structure includes Multilevel method
The data processing structure of device, including input layer, output layer and hidden layer positioned there between, multilayer perceptron nerve
In the training process of network using extraction it is multigroup including tone H and saturation degree S color characteristics as input.Wherein, Multilayer Perception
The mathematical modeling of device nerve net is：
yi=f (xi) i=1,2, m.i.=1,2, m...
In formula, wijFor weight matrix, sjIt is input matrix, θiIt is threshold function table, f (xi) is activation primitive, is preferably selected
Softmax functions are as activation primitive.The training of certain multilayer perceptron neutral net is completed before normal work.
On the basis of above-described embodiment, in another embodiment of the present invention, the training process of multilayer perceptron neutral net
Middle input layer number is 2, and it is 2 to imply the number of plies, and every layer of hidden layer neuron number is 4, output layer neuron number root
The species setting that waste non-ferrous metals according to sorting are.
, it is necessary to which the species of the waste non-ferrous metals material of sorting has 6 kinds, respectively early, middle and late under a kind of specific embodiment
The waste material of different shapes of each species is taken pictures, every kind of waste material, which amounts to, shoots 50 pictures, using 25 pictures as instruction
Practice sample, remaining 25 pictures are as test sample.2*25 matrix is inputted in the input layer of multilayer perceptron, stores useless gold
Belong to the characteristic information of material, output layer is 25*6 matrix.It is as a result as shown in the table after sifting sort of the present invention：
| Waste material species name | Correct number | Preparation rate (%) |
| Red copper | 24 | 96.0 |
| Brass | 24 | 96.0 |
| Aluminium | 23 | 92.0 |
| Magnesium | 23 | 92.0 |
| Lead | 24 | 96.0 |
| Zinc | 23 | 92.0 |
As seen from the above table, the present invention is high for the classification accuracy of the obvious red copper of color characteristic, brass, reachable
96%, and for magnesium, aluminium, zinc these three old metal materials, after oxidation, color is close, so the accuracy rate of classification is relative
Decrease.
On the basis of above-described embodiment, in another embodiment of the present invention, such as Fig. 1, first transmission range, which is provided with, to be used for
Sense the second sensor 6 that material enters the predeterminated position of the first transmission range.Second in the predeterminated position of the first transmission range passes
The purpose of sensor 6 is material signal is passed into central processing module, central processing module control visual identity unit 7 in time
Visual identity preparation is carried out, operation is identified in time after material enters the visual identity region 4 of the second transmission range, improves it
Reaction speed；It is preferred that the first sensor 5 and second sensor 6 are fibre optical sensor.
On the basis of above-described embodiment, in another embodiment of the present invention, such as Fig. 1, the jet-propelled sorting unit also wraps
Air compressor 12 is included, the jet module 9 is connected with the air compressor 12 by pipeline, the spray of the jet module 9
Gas direction is adjustable.It is arranged such, the material information fed back according to CPU, the controllable puff prot of jet module 9 is opened
Close, by the use of air-flow is sprayed as being before used as power, movement locus, speed and the acceleration of material is changed, so that material
Default different stock chests 11 are fallen on, realize the purpose of automatic sorting.Because air is inexhaustible, exhaust
Processing is simple, free from environmental pollution, and only need the position of reasonable Arrangement nozzle and airflow direction and pressure can realize it is a variety of
The separation of scrap metal, good separation effect.
On the basis of above-described embodiment, in another embodiment of the present invention, such as Fig. 1, second transmission range is by dividing plate 13
Multiple transmission channels are separated into, the end of each transmission channel is equipped with jet module 9, and the jet module 9 is provided with multiple sprays
Gas port, the jet direction of the puff prot is adjustable.It is arranged such, on the one hand the second transmission range is separated into multiple biographies by dividing plate 13
Defeated passage, so can the second transmission range as far as possible material separate set, avoid being superimposed with each other, improve visual identity effect,
The width of the second transmission range can be then effectively increased simultaneously, and the same time can be to the more of the same width position in end of the second transmission range
Individual material is blowed, and improves the efficiency of separation；On the other hand each jet module 9 is equipped with multiple puff prots, so by more
The synergic adjustment control of individual puff prot, jet angular adjustment effect are more preferable.
On the basis of above-described embodiment, in another embodiment of the present invention, the jet module 9, which is provided with, to be used to adjust jet
The stepper motor 14 of angle and the high speed pulse valve for controlling gas channel size, the stepper motor 14 and the jet module
9 are connected by connecting rod.It is arranged such, the stepper motor 14, which rotates, drives linkage, the puff prot of adjustment jet module 9
Angle, and control airflow direction and pressure to realize different types of material being blown into corresponding stock chest 11.
On the basis of above-described embodiment, in another embodiment of the present invention, the feed unit includes vibratory sieve 1 and reason is expected
Slide rail 2, the reason material slide rail 2 are in predetermined inclination angle, and the vibratory sieve 1 is connected with the first transmission range by managing material slide rail 2, institute
The loading plane for stating vibratory sieve 1 is higher than the transport plane of the first transmission range.In this way, pass through the vibration of vibratory sieve 1, old metal material
Separated by shake, and have stepped through reason material slide rail 2 and be included in the first transmission range, can so pull open the distance between material, and by thing
Expect, into preliminary arrangement, to avoid overlapping on together, visual identity effect can be improved, and then improve separating effect and efficiency.
On the basis of above-described embodiment, in another embodiment of the present invention, the visual identity unit 7 includes kilomega network work
Industry camera and light source.Using kilomega network industrial camera, can long range high speed transmission data, while can quick and precisely catch figure
Picture, the stability that light in principal security target visual identification process is set of light source.
On the basis of above-described embodiment, in another embodiment of the present invention, the central control unit 8 include switch board and
Industrial computer, the industrial computer are provided with touch-screen.Industrial computer is the core of control system, is responsible for data storage, image procossing and spy
Sign is matched, sends control signal etc. to slave computer according to processing result image, and each part can be monitored in real time additionally by touch-screen
Working condition and running situation, check sorting information, set sorting pattern, while ensureing stability and durability, improve
The convenience of man-machine interaction.
On the basis of above-described embodiment, in another embodiment of the present invention, the length of first transmission range for 1.5~
3m, transmission speed are 0.3~0.5m/s.The setting of first transmission range of scheduled transmission speed and length, main give supply through vibration
The certain buffering of old metal material after material, influence of the vibration to institute's material visual identity is reduced as far as possible, improves shooting image
Quality, the transmission speed of the first transmission range can not be too fast, and length is unsuitable too short, and otherwise buffering effect is bad, material during shooting
Certain vibration is still suffered from, while its transmission speed can not be too fast, length is unsuitable long, can so reduce the efficiency of separation.
On the basis of above-described embodiment, in another embodiment of the present invention, the transmission speed of second transmission range is small
In 7m/s.
On the basis of above-described embodiment, in another embodiment of the present invention, such as Fig. 1, the transmission unit 3 is the band of motor 14
Dynamic transmission belt.
Described above is only the preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications also should
It is considered as protection scope of the present invention.
Claims (10)
1. a kind of waste non-ferrous metals automatic sorting system based on machine vision, it is characterised in that single including feed unit, transmission
Member, visual identity unit, sensing unit, feed separation unit and central control unit, the sensing module, visual identity list
Member and feed separation unit electrically connect with central control unit, and the transmission unit includes the first transmission along material transferring direction
Area and the second transmission range, the transmission speed of first transmission range are less than the transmission speed of the second transmission range, second transmission
Area is preset with visual identity region, and the feed separation unit includes jet-propelled separator and storage device, the jet
Formula separator includes being arranged on the jet module of the second transmission range end, and the storage device is provided with more along material transferring direction
First transmission range of material transmission to transmission unit, the sensing unit are comprised at least and set by individual stock chest, the feed unit
The first sensor in the second transmission range is put, the first sensor is used to sense material into visual identity region and will sensed
Signal passes to central control unit, and the central control unit control visual identity unit is to the thing into visual identity region
Material carries out visual identity, and identification information is passed to central control unit, the central control unit by the visual identity unit
Matched after handling the identification information of reception compared with the visual information to prestore, and according to comparison match output control
The jet direction of jet-propelled separator adjustment jet module and/or whiff pressure, the jet module divide the material of identification
Choosing is into corresponding stock chest.
2. the waste non-ferrous metals automatic sorting system according to claim 1 based on machine vision, it is characterised in that described
The process that central control unit is handled the identification information of reception comprises at least image preprocessing and color feature extracted
Step, the central control unit by after the color characteristic of extraction compared with the color characteristic of the visual information to prestore
Match somebody with somebody.
3. the waste non-ferrous metals automatic sorting system according to claim 2 based on machine vision, it is characterised in that described
Image preprocessing step include image filtering handle and image dividing processing, described image dividing processing realize target image with
The mutually segmentation of background image.
4. the waste non-ferrous metals automatic sorting system according to claim 3 based on machine vision, it is characterised in that use
Median filter carries out image filtering processing, and described image dividing processing uses the color images under YCrCb color spaces
Method.
5. the waste non-ferrous metals automatic sorting system according to claim 4 based on machine vision, it is characterised in that described
The step of image dividing processing, is as follows：
(1) picture of visual identity unit photographs is transformed into YCrCb color spaces；
(2) color feature value of pixel is extracted；
(3) the color feature value selected seed of the pixel of extraction in step (2) is utilized；
(4) seed region is grown up；
(5) judge whether that all pixels point is processed, be, the seed region in step (4) is merged, otherwise returned
The step of continuing seed region growth to step (4).
6. the waste non-ferrous metals automatic sorting system according to claim 4 based on machine vision, it is characterised in that described
The redundancy of RGB color color channel is reduced in the step of image dividing processing using YCrCb color spaces, utilizes image point
Hue and luminance after cutting represents colouring information, and Y is brightness value, is obtained by color RGB color R, G, B value weighting, calculates
Formula is：Y=irR+igG+ibB, wherein, ir, ig, ibIt is weighted factor, tone can be by Y and three given different aberration
Cr, Cg, Cb represent that calculation formula is as follows：Cr=R-Y；Cg=G-Y；Cb=B-Y.
7. the waste non-ferrous metals automatic sorting system according to claim 2 based on machine vision, it is characterised in that described
The selection of color feature extracted more stable tone H and saturation degree S carries out color characteristic analysis, wherein, first by visual identity
The picture of unit photographs is HSI models by RGB model conversations, wherein, tone H calculation formula be：
Saturation degree S is calculated by following formula：
<mrow>
<mi>S</mi>
<mo>=</mo>
<mn>1</mn>
<mo>-</mo>
<mfrac>
<mn>3</mn>
<mrow>
<mi>R</mi>
<mo>+</mo>
<mi>G</mi>
<mo>+</mo>
<mi>B</mi>
</mrow>
</mfrac>
<mo>&lsqb;</mo>
<mi>M</mi>
<mi>I</mi>
<mi>N</mi>
<mrow>
<mo>(</mo>
<mi>R</mi>
<mo>,</mo>
<mi>G</mi>
<mo>,</mo>
<mi>B</mi>
<mo>)</mo>
</mrow>
<mo>&rsqb;</mo>
<mo>.</mo>
</mrow>
8. the waste non-ferrous metals automatic sorting system according to claim 7 based on machine vision, it is characterised in that described
Also include the training process of multilayer perceptron neutral net, the multilayer perceptron nerve after the step process of color feature extracted
Network algorithm structure includes the data processing structure of Multilevel method device, including input layer, output layer and positioned at the two it
Between hidden layer, by the multigroup special including tone H and saturation degree S colors of extraction in the training process of multilayer perceptron neutral net
Sign is as input.
9. the waste non-ferrous metals automatic sorting system according to claim 8 based on machine vision, it is characterised in that multilayer
Input layer number is 2 in the training process of perceptron neural network, and it is 2 to imply the number of plies, every layer of hidden layer neuron
Number is 4, and the species that output layer neuron number is according to the waste non-ferrous metals of sorting is set.
10. the waste non-ferrous metals automatic sorting system according to claim 8 based on machine vision, it is characterised in that institute
State the first transmission range and be provided with and be used to sense material into the second sensor of the predeterminated position of the first transmission range.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711268404.9A CN107876429B (en) | 2017-12-05 | 2017-12-05 | Waste non-ferrous metal automatic separation system based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711268404.9A CN107876429B (en) | 2017-12-05 | 2017-12-05 | Waste non-ferrous metal automatic separation system based on machine vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107876429A true CN107876429A (en) | 2018-04-06 |
| CN107876429B CN107876429B (en) | 2020-05-19 |
Family
ID=61772754
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711268404.9A Expired - Fee Related CN107876429B (en) | 2017-12-05 | 2017-12-05 | Waste non-ferrous metal automatic separation system based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107876429B (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109047004A (en) * | 2018-06-27 | 2018-12-21 | 天津大学 | Ball rejected product elimination method |
| CN109047034A (en) * | 2018-07-17 | 2018-12-21 | 中国农业大学 | A kind of double standard automatic multi-step sorting units of raisins color size |
| CN109261527A (en) * | 2018-08-24 | 2019-01-25 | 昆明理工大学 | A kind of coffee bean grader and stage division |
| CN111507391A (en) * | 2020-04-13 | 2020-08-07 | 武汉理工大学 | Intelligent identification method for nonferrous metal broken materials based on machine vision |
| CN111530778A (en) * | 2020-05-11 | 2020-08-14 | 湖南工学院 | Plastic color selection mechanism and method for plastic regeneration device and plastic regeneration device |
| TWI723604B (en) * | 2018-10-31 | 2021-04-01 | 日商Ｊｘ金屬股份有限公司 | The composition analysis device of electronic and electrical equipment parts, the processing device of electronic and electrical equipment parts and the processing methods of electronic and electrical equipment parts |
| CN112930302A (en) * | 2018-09-26 | 2021-06-08 | 聪慧公司 | Robot full-set goods picking system |
| CN113414123A (en) * | 2021-06-10 | 2021-09-21 | 云南永昌硅业股份有限公司 | Vein quartz color sorting method and device for improving quality of industrial silicon smelting raw materials |
| CN113426707A (en) * | 2021-06-10 | 2021-09-24 | 云南永昌硅业股份有限公司 | Silicon slag color separation method and device for effectively improving silicon metal recovery rate |
| CN113516637A (en) * | 2021-06-23 | 2021-10-19 | 深圳市小绿人网络信息技术有限公司 | Method and device for automatically identifying steel scrap types based on images |
| US11731792B2 (en) | 2018-09-26 | 2023-08-22 | Dexterity, Inc. | Kitting machine |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1967294A2 (en) * | 2007-03-05 | 2008-09-10 | Picvisa Machine Vision Systems, S.L. | Machine with artificial sight for automatic separation by colour of recyclable plastics, with multispectral sight |
| CN202021164U (en) * | 2011-03-18 | 2011-11-02 | 青岛农业大学 | Peanut exterior quality detecting and sorting device |
| CN102737367A (en) * | 2012-06-17 | 2012-10-17 | 四川农业大学 | Tea image enhancement and division method based on color characteristic |
| CN103473568A (en) * | 2013-09-13 | 2013-12-25 | 山东科技大学 | Separation device for coal and waste rocks and separation method thereof |
| CN205270174U (en) * | 2015-12-08 | 2016-06-01 | 徐浩军 | A look selection machine for giving up metal recycling |
| CN205289011U (en) * | 2015-12-17 | 2016-06-08 | 赵凌宇 | Building rubbish sorting device based on machine vision |
| WO2016102725A1 (en) * | 2014-12-23 | 2016-06-30 | Environmental Green Engineering, S.L. | Machine with computer vision provided for the automatic separation of black plastic waste by chemical composition |
| CN106984556A (en) * | 2017-04-28 | 2017-07-28 | 福州大学 | A kind of plastic bottle intelligent sorting system and its method of work |
| CN107103606A (en) * | 2017-02-27 | 2017-08-29 | 口碑控股有限公司 | A kind of image-recognizing method and device |
-
2017
- 2017-12-05 CN CN201711268404.9A patent/CN107876429B/en not_active Expired - Fee Related
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1967294A2 (en) * | 2007-03-05 | 2008-09-10 | Picvisa Machine Vision Systems, S.L. | Machine with artificial sight for automatic separation by colour of recyclable plastics, with multispectral sight |
| CN202021164U (en) * | 2011-03-18 | 2011-11-02 | 青岛农业大学 | Peanut exterior quality detecting and sorting device |
| CN102737367A (en) * | 2012-06-17 | 2012-10-17 | 四川农业大学 | Tea image enhancement and division method based on color characteristic |
| CN103473568A (en) * | 2013-09-13 | 2013-12-25 | 山东科技大学 | Separation device for coal and waste rocks and separation method thereof |
| WO2016102725A1 (en) * | 2014-12-23 | 2016-06-30 | Environmental Green Engineering, S.L. | Machine with computer vision provided for the automatic separation of black plastic waste by chemical composition |
| CN205270174U (en) * | 2015-12-08 | 2016-06-01 | 徐浩军 | A look selection machine for giving up metal recycling |
| CN205289011U (en) * | 2015-12-17 | 2016-06-08 | 赵凌宇 | Building rubbish sorting device based on machine vision |
| CN107103606A (en) * | 2017-02-27 | 2017-08-29 | 口碑控股有限公司 | A kind of image-recognizing method and device |
| CN106984556A (en) * | 2017-04-28 | 2017-07-28 | 福州大学 | A kind of plastic bottle intelligent sorting system and its method of work |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109047004A (en) * | 2018-06-27 | 2018-12-21 | 天津大学 | Ball rejected product elimination method |
| CN109047034A (en) * | 2018-07-17 | 2018-12-21 | 中国农业大学 | A kind of double standard automatic multi-step sorting units of raisins color size |
| CN109047034B (en) * | 2018-07-17 | 2023-12-29 | 中国农业大学 | Raisin color size double-standard automatic multi-stage sorting device |
| CN109261527A (en) * | 2018-08-24 | 2019-01-25 | 昆明理工大学 | A kind of coffee bean grader and stage division |
| CN109261527B (en) * | 2018-08-24 | 2021-06-18 | 昆明理工大学 | Coffee bean grading machine and grading method |
| US11731792B2 (en) | 2018-09-26 | 2023-08-22 | Dexterity, Inc. | Kitting machine |
| CN112930302A (en) * | 2018-09-26 | 2021-06-08 | 聪慧公司 | Robot full-set goods picking system |
| TWI723604B (en) * | 2018-10-31 | 2021-04-01 | 日商Ｊｘ金屬股份有限公司 | The composition analysis device of electronic and electrical equipment parts, the processing device of electronic and electrical equipment parts and the processing methods of electronic and electrical equipment parts |
| CN111507391A (en) * | 2020-04-13 | 2020-08-07 | 武汉理工大学 | Intelligent identification method for nonferrous metal broken materials based on machine vision |
| CN111530778B (en) * | 2020-05-11 | 2021-11-16 | 湖南工学院 | Plastic color selection mechanism and method for plastic regeneration device and plastic regeneration device |
| CN111530778A (en) * | 2020-05-11 | 2020-08-14 | 湖南工学院 | Plastic color selection mechanism and method for plastic regeneration device and plastic regeneration device |
| CN113426707A (en) * | 2021-06-10 | 2021-09-24 | 云南永昌硅业股份有限公司 | Silicon slag color separation method and device for effectively improving silicon metal recovery rate |
| CN113414123A (en) * | 2021-06-10 | 2021-09-21 | 云南永昌硅业股份有限公司 | Vein quartz color sorting method and device for improving quality of industrial silicon smelting raw materials |
| CN113516637A (en) * | 2021-06-23 | 2021-10-19 | 深圳市小绿人网络信息技术有限公司 | Method and device for automatically identifying steel scrap types based on images |
| CN113516637B (en) * | 2021-06-23 | 2022-02-15 | 深圳市小绿人网络信息技术有限公司 | Method and device for automatically identifying steel scrap types based on images |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107876429B (en) | 2020-05-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107876429A (en) | A kind of waste non-ferrous metals automatic sorting system based on machine vision | |
| WO2020187077A1 (en) | Deep neural network-based security check system and method | |
| CN108596277A (en) | A kind of testing vehicle register identification method, apparatus and storage medium | |
| CN106874840B (en) | Vehicle information recognition method and device | |
| CN107134144B (en) | A kind of vehicle checking method for traffic monitoring | |
| CN106548165B (en) | A kind of face identification method of the convolutional neural networks based on image block weighting | |
| CN104156726B (en) | A kind of workpiece identification method and device based on geometric characteristic | |
| CN108280856A (en) | The unknown object that network model is inputted based on mixed information captures position and orientation estimation method | |
| CN106216219B (en) | A kind of visual identity coal and gangue and the system detached with manipulator | |
| CN108108679B (en) | Full-automatic tungsten ore concentrating machine | |
| CN107038448A (en) | Target detection model building method | |
| CN107729818A (en) | A kind of multiple features fusion vehicle recognition methods again based on deep learning | |
| CN107486415A (en) | Thin bamboo strip defect on-line detecting system and detection method based on machine vision | |
| CN101021870A (en) | Picture inquiry method and system | |
| CN106737664A (en) | Sort the Delta robot control methods and system of multiclass workpiece | |
| CN102214306B (en) | Leaf disease spot identification method and device | |
| CN110245663A (en) | One kind knowing method for distinguishing for coil of strip information | |
| CN107341688A (en) | The acquisition method and system of a kind of customer experience | |
| CN108764115A (en) | A kind of truck danger based reminding method | |
| CN108334906A (en) | A kind of finance pawns the guaranty automatic identification appraisal procedure and device of service | |
| CN107292933A (en) | A kind of vehicle color identification method based on BP neural network | |
| CN109506628A (en) | Object distance measuring method under a kind of truck environment based on deep learning | |
| CN111359915B (en) | Material sorting method and system based on machine vision | |
| CN109215010A (en) | A kind of method and robot face identification system of picture quality judgement | |
| CN108287010A (en) | A kind of crab multi objective grading plant and method |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200519 Termination date: 20201205 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |