CN106370667A - Visual detection apparatus and method for quality of corn kernel - Google Patents

Visual detection apparatus and method for quality of corn kernel Download PDF

Info

Publication number
CN106370667A
CN106370667A CN201610632800.4A CN201610632800A CN106370667A CN 106370667 A CN106370667 A CN 106370667A CN 201610632800 A CN201610632800 A CN 201610632800A CN 106370667 A CN106370667 A CN 106370667A
Authority
CN
China
Prior art keywords
image
corn kernel
detection
quality
module
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.)
Withdrawn
Application number
CN201610632800.4A
Other languages
Chinese (zh)
Inventor
郑振兴
梁鹏
林智勇
贾西平
蓝钊泽
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong Polytechnic Normal University
Original Assignee
Guangdong Polytechnic Normal University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guangdong Polytechnic Normal University filed Critical Guangdong Polytechnic Normal University
Priority to CN201610632800.4A priority Critical patent/CN106370667A/en
Publication of CN106370667A publication Critical patent/CN106370667A/en
Withdrawn legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/89—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles
    • G01N21/8901—Optical details; Scanning details
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/89—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles
    • G01N21/8914—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles characterised by the material examined
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/89—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles
    • G01N21/8901—Optical details; Scanning details
    • G01N2021/8908—Strip illuminator, e.g. light tube

Landscapes

  • Engineering & Computer Science (AREA)
  • Textile Engineering (AREA)
  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)

Abstract

The invention discloses a visual detection apparatus and method for the quality of a corn kernel. The visual detection apparatus comprises a detection platform and a detection system, wherein the detection platform comprises a workbench used for placing the corn kernel and a conveying mechanism; the detection system comprises a control center module, a separating and discharging module, an image analysis module, an illumination module and a motion control module; and the workbench comprises a shaking source, a scraping plate, a shading case, a photoelectric sensor, an illumination light source and a CCD camera. According to the invention, a visual sense-based detection scheme is employed for visual detection and analysis of the quality of the corn kernel so as to realize quality detection of corn kernels of different types and impurity selection, and high-speed automatic detection of high-quality corn kernels is guaranteed.

Description

A kind of corn kernel quality vision inspection apparatus and method
Technical field
The present invention relates to machine vision device and method are and in particular to corn kernel quality vision inspection apparatus and method.
Background technology
Agricultural and food industry and people's life are closely coupled, and the applied research related to these means of livelihood be not in recent years Count its number.And in agriculture field, China is Maize Production and consumption big country, gross annual output amount up to 1.2 hundred million tons, Semen Maydiss simultaneously In addition to edible, also serve as feedstuff and the raw material of industry, the corn kernel of high-quality be improve yield in agricultural and industrial processes, Efficiency and the guarantee of quality.Therefore, actively improve corn quality, obtain its quality quick detection, differentiate, classification, preferably former Reason and method, are the needs in preferably development China's Grain Development and commerce and trade market, are also that agriculture reproduction is very crucial Factor.
At present, China's existing corn quality evaluation criteria is all using unit weight, impurity and defect grain etc. as important Discrimination standard.Detection method main at present is to detect corn kernel quality using hand picking method.Hand picking method needs Human eye constantly stares at thickly dotted corn kernel on streamline, and high labor intensive, because artificial range estimation is inevitably subject to The impact of the factors such as the vision of individual, emotion, light, with very big personal subjectivity, so examination criteria is difficult to unified, inspection Survey efficiency is low, sorting difference is big, adds that site environment noise and corn kernel itself are reflective, workman is easily tired.This detection Method is both time-consuming to take a lot of work, and is difficult to ensure that detection quality again, needs badly and realizes corn kernel using efficient, automatic machine vision technique Quality automatic detection.
Another kind of main stream approach is weight test method, distinguishes qualified and substandard product by detecting product weight, but It is production line balance that corn kernel produces, and the method is easily subject to production line machinery vibrations interference to cause flase drop and missing inspection, and cannot Detect normal niblet and impurity, therefore, it is difficult to being widely popularized.
Machine vision is also a kind of technological means that corn kernel Quality Detection is commonly used, but yet suffers from deficiency: (1) adopts Carried out with machine vision corn kernel detection when it is often necessary to manually put or on hardware device adopt fixation holes method Lai Realize single grain region recognition, but manually put and take time and effort, fixation holes method also is difficult to ensure that single grain is accurately put, If hole design is improper results even in seed breakage；(2) often extract a large amount of characteristic parameters such as area, most advanced and sophisticated point, centre of form etc. to make For evaluation index, but when photographic head shooting distance changes, these features can change, thus leading to accuracy of detection Reduce.
Effective effect of technical solution of the present invention is: (1) makes sample tile separately by jitter sources and material scraping plate, that is, Make still there is adhesion situation, only need to do partial segmentation and process, hardware device so just can be avoided to separate the damage causing during sample； (2) extract the feature unrelated with shooting distance, improve vision-based detection efficiency.
Inquired about by Patents, find there is a following open source literature:
Patent " household cereal quality inspection device and detection method " [application number cn201310563704.5] discloses one Plant device and the detection method carrying out Quality Detection using corn escaping gas, this device sucks detection gas by detection probe Body, is processed using cascade bistable-state random resonance system, is mated to judge whether corn goes mouldy with the threshold value setting. The method is high for environmental requirement, and can not detect the corn of appearance defect.
Patent " automatic checkout system of solid grain " [application number cn201520438544.6] discloses a kind of solid seed The automatic checkout system of grain, system includes: seed sample introduction tank, gear unit, photoswitch, rotating disk, the first motor, signals collecting Unit and controller, the program by between each part cooperation so as to solid grain detection full-automation.The program makes With spectral technique, information processing is carried out to seed, and non-usage machine vision technique；Additionally, the program is made using rotating disk and baffle plate Seed becomes string and is processed, a seed can only be processed every time, processing speed is slow, and the present invention program is regarded using machine Feel, disposably multiple seeds can be processed, and employ jitter sources and material scraping plate so that seed tiling is not overlapping.
A kind of patent " grain unsound grain detection plate with secondary light source " [application number cn201520246581.7] profit The backlight being provided with mainboard and the front light of desk lamp offer, provide various visual angles and omnibearing light conditions, so that in grain seed Outward appearance and the details of grain unsound grain are clearly displayed on grain detection plate, is configured with auxiliary in grain seed detection plate simultaneously Equipment and grain unsound grain collection of illustrative plates and the panel of parsing data, for reviewer nearby using and with reference to comparing.The program does not have Use automatic transmission mechanism, and automatic detection is not carried out using machine vision technique.
Content of the invention
Present invention aim to overcome that the deficiencies in the prior art, especially solve existing corn kernel Quality Detection means and lack Weary automatic detection device, be difficult to automatically tile sub-material the problems such as.A kind of corn kernel quality vision inspection apparatus are provided, should Device makes sample tile separately by jitter sources and material scraping plate, even if still there being adhesion situation, only need to do partial segmentation and processing, Hardware device so just can be avoided to separate the damage causing during sample, realize efficient corn kernel quality automatic detection.
Further object is that providing a kind of machine vision method applying above-mentioned corn kernel Quality Detection, The method is passed through to extract the feature unrelated with shooting distance so that testing result is more quick, accurate, reliable.
For solving above-mentioned technical problem, the present invention adopts the following technical scheme that:
A kind of corn kernel quality vision inspection apparatus, including transmission flitch, jitter sources, material scraping plate, shading case, photoelectric transfer Sensor, lighting source, ccd camera and drive mechanism are it is characterised in that wherein:
Described jitter sources are located at transmission flitch front end；Described material scraping plate is located at transmission flitch middle-end；Described transmission flitch position Above drive mechanism；Described drive mechanism is by positioned at the Timing Belt transmitting flitch bottom and the motor being connected with Timing Belt Composition；Described transmission flitch, jitter sources, material scraping plate, photoelectric sensor are respectively positioned on inside shading case；Described photoelectric sensor is by position Form in the photoemitter of shading case both sides and photelectric receiver；Described lighting source is located at the shading case left and right sides；Described Ccd camera is located at directly over shading case, and ccd camera mirror plane is towards the table top of described transmission flitch.
Further, described drive mechanism includes motor and Timing Belt, in the main shaft of motor and Timing Belt Driving pulley connect；Motor main shaft drives Timing Belt motion, and Timing Belt drives the Semen Maydiss to be detected on transmission flitch Seed moves in the straight direction so that ccd camera is capable of image acquisition.
A kind of corn kernel quality vision detection system, including control centre's module, image analysis module, lighting module with And motion-control module is it is characterised in that wherein:
Control centre's module includes arm embedded platform and control centre's program, and control centre's program runs on arm and embeds On formula platform, arm embedded platform passes through serial ports and is connected respectively at image analysis module, lighting module, motion-control module；
Image analysis module includes dsp mainboard, ccd camera, analog monitor and image analysis system, image analysis system Run on dsp mainboard, ccd camera, analog monitor are connected dsp mainboard respectively；
Lighting module includes lighting source, illumination driver and illumination driver, and illumination driver runs on illumination Driver, illumination driver is connected with lighting source；
Motion-control module includes motion controller, motion control program, photoelectric sensor and motor, motor control Program runs on motion controller, and motion controller is connected with photoelectric sensor and motor respectively.
The corn kernel quality visible detection method that the present invention is realized is it is characterised in that comprise the following steps:
(1) image acquisition:
When corn kernel is through photoelectric sensor, the dsp being connected with photoelectric sensor is driven ccd camera to carry out image and adopts Collection, the image collecting is carried out image procossing by dsp again；
(2) image procossing: early stage process is carried out to the corn kernel image collecting, to extract in corn kernel image Characteristic information, as the input of image recognition；
(3) image recognition: according to the characteristic information of image procossing, corn kernel image is identified, identifies Semen Maydiss The quality of seed and the defect of presence.
Brief description
Fig. 1 and Fig. 2 is the structural representation of a specific embodiment of corn kernel quality vision inspection apparatus of the present invention Figure, Fig. 1 is front view, and Fig. 2 is lateral plan.
Fig. 3 is the schematic diagram of system module in a specific embodiment of the present invention.
Fig. 4 is the schematic diagram of image analysis system in a specific embodiment of the present invention.
Fig. 5 is the schematic diagram on chain code border in image procossing in a specific embodiment of the present invention.
Fig. 6 is the support vector cassification schematic diagram of picture recognition module in a specific embodiment of the present invention.
Specific embodiment
Below in conjunction with the accompanying drawings and specific embodiment the present invention is carried out in further detail with complete explanation.May be appreciated It is that specific embodiment described herein is only used for explaining the present invention, rather than limitation of the invention.
Referring to Fig. 1, a kind of corn kernel quality vision inspection apparatus of the present invention, including transmission flitch 1, jitter sources 2, scrape Flitch 3, shading case 4, photoelectric sensor 5, lighting source 6, ccd camera 7 and drive mechanism, wherein:
Described jitter sources 2 are located at transmission flitch 1 front end；Described material scraping plate 3 is located at transmission flitch 1 middle-end；Described transmission material Plate 1 is located above drive mechanism；Described drive mechanism is by positioned at the Timing Belt 8 transmitting flitch bottom and the drive being connected with Timing Belt Galvanic electricity machine 9 forms；Described transmission flitch 1, jitter sources 2, material scraping plate 3, photoelectric sensor 5 are respectively positioned on inside shading case 4；Described light Electric transducer 5 is made up of the photoemitter positioned at shading case 4 both sides and photelectric receiver；Described lighting source 6 is located at shading Case 4 left and right sides；Described ccd camera 7 is located at directly over shading case 4, and ccd camera 7 minute surface is towards the platform of described transmission flitch 1 Face.
Referring to Fig. 1 and Fig. 2, described drive mechanism includes motor 9 and Timing Belt 8, the main shaft of motor 9 with Step connects with the driving pulley in 8；The main shaft of motor 9 drives Timing Belt 8 to move, and Timing Belt 8 drives on transmission flitch 1 Corn kernel to be detected move in the straight direction so that ccd camera 7 is capable of image acquisition.
Vibration source adopts high-frequency electric vibrator；Using Japanese tk-c1381eg, 1/2 inch, 220v's ccd camera powers, 0.95lux 470 line；Light source adopts the full radiant of 10w/12v, is powered using DC source, can change lamp by adjusting voltage Light power is to adapt to different light conditions；Light source is arranged on top about shading case, can be obtained soft by the diffuse-reflectance of light The low-intensity scattered light of sum is to obtain soft image；Photoelectric sensor uses qs18 photoelectric sensor；
Referring to Fig. 3, a kind of corn kernel quality vision detection system, including control centre's module, image analysis module, light Lighting module and motion-control module, wherein:
Control centre's module includes arm embedded platform and control centre's program, and control centre's program runs on arm and embeds On formula platform；Arm embedded platform is connected with image analysis module by serial ports, execution graphical analyses relevant parameter during start Initial configuration；Arm embedded platform is connected with lighting module by serial ports, the preheating of execution light source and initial chemical industry during start Make so that brightness of taking pictures meets specifies requirement；Arm embedded platform is connected with motion-control module, during start by pci bus Actuating motor initializes；
Image analysis module includes dsp mainboard, ccd camera, analog monitor and image analysis system, image analysis system Run on dsp mainboard, ccd camera, analog monitor are connected dsp mainboard respectively, and dsp mainboard receives the number from ccd camera Word signal, and using image analysis program, digital signal is processed, output result is converted into analogue signal, sends to mould Intend monitor；
Lighting module includes lighting source, illumination driver and illumination driver, and illumination driver runs on illumination Driver, illumination driver is connected with lighting source；
Motion-control module includes motion controller, motion control program, photoelectric sensor and motor, motor control Program runs on motion controller, controls operation, the stopping of motor；Motion controller respectively with photoelectric sensor and drive Galvanic electricity machine is connected, using motec α mld intelligent DC servo-driver as motion controller.
Referring to Fig. 4, image analysis system is made up of five parts: image acquisition, image procossing, image recognition, center to center communications, Digital-to-analogue converts, wherein:
(1) image acquisition: by dsp mainboard receive from ccd camera image, and according in systems soft ware according to software phase The macrodefinition closed or configuration carry out real-time storage to the corn kernel sample image collecting.
(2) image procossing: comprehensive use various image processing algorithms, to ccd collected by camera to corn kernel image enter Row analysis and calculating, obtain the coherent detection characteristic parameter of corn kernel sample, and export the characteristic parameter of sample in real time.
(3) image recognition: the coherent detection characteristic parameter of the corn kernel sample being obtained according to image procossing, using identification The quality of method detection sample and the species of defect.
(4) center to center communications: process the signal being sent to dsp development board from transceiver terminal (arm embedded board).According to soft Part communication protocol, completes the communication of each module and other external modules on dsp development board.
(5) digital-to-analogue conversion: by running corresponding program on dsp development board, by the Rice Samples image collecting and every Intermediate result during one step analysis and image, export to communication terminal or analog video output interface, in order to directly see Examine, debug or sundry item application and development.
A kind of image processing process of corn kernel quality visible detection method of the present invention, specifically includes following steps:
(1) image gray processing: the image that video camera obtains is coloured image, and comprise contains much information, image processing speed Slower.In view of industrial production automation, real-time is had high demands, and corn kernel Quality Detection does not need using colored Information, it is necessary for coloured image is carried out with gray processing processing.Gray processing makes r, g, b component value of colour element equal Process, the gray value in gray level image is equal to the rgb meansigma methodss in original color image, that is,
Gray=(r+g+b)/3 (1)
(3) image filtering: contrast inadequate situation often occurs in the picture, this is likely due to image record dress It is also possible to due to caused by the originally under-exposure in image acquisition process caused by the dynamic range size put.In order to protect Card can obtain the feature of corn kernel in image, in order to analysis further and process.Figure can be made using histogram equalization The gray scale spacing of picture is pulled open or is made intensity profile uniformly, thus increasing contrast, making image detail clear, reaching the mesh of image enhaucament 's.Its concrete grammar is:
Provide all gray levels s of original image firstk(k=0,1 ..., l-1)；Then statistics original image is each The pixel count n of gray levelk；Use the accumulation of (3) formula calculating original image straight calculate the rectangular histogram of original image using formula (2) after again Fang Tu:
p(sk)=nk/ n, k=0,1 ..., l-1 (2)
t k = σ i = 0 k p ( s j ) , k = 0 , 1 ... l - 1 - - - ( 3 )
p(tk)=nk/n (4)
Wherein, n is total number of image pixels.To gray value tkRound, determine sk→tkMapping relations after count new histogram The pixel count n of each gray levelk；Formula (4) is finally utilized to calculate new rectangular histogram.
(4) edge extracting: image is carried out with edge extracting to realize extracting the outline of corn kernel.Generally adopt one Rank differential or second-order differential computing, try to achieve the zero crossing of maximum of gradients or second dervative, finally choose suitable threshold value to carry Take image border.
Roberts operator is the operator that a kind of utilization local difference operator finds edge, is given by following formula (5):
g ( x , y ) = [ f ( x , y ) - f ( x + 1 , y + 1 ) ] 2 + [ f ( x + 1 , y ) - f ( x , y + 1 ) ] 2 - - - ( 5 )
(5) closing operation: the factor due to affecting image image quality is a lot, is carrying out successive image feature extraction behaviour The situation that when making in fact it could happen that obscurity boundary, edge disconnects, for such situation, should be using closing fortune in morphological method Calculate, to eliminate peripheral fault.Closed operation first expands post-etching, if f (x, y) is input picture, b (x, y) is structural element, uses Input picture f is expanded structural element b and erosion operation is respectively defined as:
( f &circleplus; b ) ( x , y ) = m a x { f ( s - x , t - y ) + b ( x , y ) | ( s - x , t - y ) &element; d f , ( x , y ) &element; d b } - - - ( 6 )
(f b) (x, y)=min f (s-x, t-y)+b (x, y) | (s-x, t-y) ∈ df, (x, y) ∈ db} (7)
A kind of image recognition processes of corn kernel quality visible detection method of the present invention, specifically include following steps:
(1) Extraction of Geometrical Features: extract the geometric properties of corn kernel on the basis of image procossing, such as girth, area, Most advanced and sophisticated point, major axis, short axle, as follows respectively:
1. girth: girth refers to the boundary length around corn kernel region, it is possible to use chain representation boundary information, institute The length sum having code section is referred to as girth.In order to closer to objective contour, typically describe border using 8 chain codes, as Fig. 5 institute Show, if each yard of segment length is δ li, now all yards of segment length sums can be expressed as
σ i = 1 n δl i = n 2 + ( n 1 - n 2 ) 2 - - - ( 8 )
Wherein, n1Represent chain code line segment sum, n2Represent even number chain code line segment sum, n represents chain code line segment sum.
2. area: replace real area with the elemental area in corn kernel region as the attribute of target.If setting unit The size of pixel is 1, then i-th labelling target riAll pixels point area sum be single seed area a, permissible It is formulated into:
Wherein, (x, y) represents the coordinate of a pixel.
3. the centre of form: the positioning of the centre of form and the distribution of shapes of object have very big relation, can react seed to a certain extent Shape information, for the two dimensional image of collection, can be regarded as a plane lamina, be solved thin slice centre of form coordinate (ox, oy) public affairs Formula is:
o x = 1 a σ x i &element; d x i o y = 1 a σ y i &element; d y i - - - ( 10 )
Wherein, (xi, yi) representing the coordinate of pixel i, d represents the regional extent of certain corn kernel in image.
4. major axis: i.e. the major axis of corn kernel is as major axis, centered on the centre of form of seed, with 1 degree as step-length, 360 Degree rotation one time, chooses wherein long line segment as major axis.
5. short axle: line segment between 2 points that the centre of form and the straight line vertical with major axis and profile border are intersected will be crossed as short Axle.
6. maximum inscribed circle radius: calculate the distance between each pixel and centre of form o on border, find wherein with o away from From pixel the shortest, and beeline is expressed as dmin.Then, find with centre of form o point as the center of circle, a length of d of radiusmin/ 2 All pixels point coordinates in circle, and obtain the beeline d ' between each pixel and boundary point in circlemin, can try to achieve accordingly Maximum max (d 'min), it is the maximum inscribed circle radius r of required solutionn.
7. minimum circumscribed circle radius: centroid point o is as the center of circle for pickup, thinks dmin/ 2 radiuses are made to justify, and obtain all pictures in circle Ultimate range d ' between vegetarian refreshments and boundary pointmin, and find minima min (d ' thereinmin), it is the single of required solution Corn kernel border minimum circumscribed circle radius rw.
(2) bending moment does not extract: image geometry feature can be affected by acquisition system itself, such as resolution and focal length Change, simultaneously moving, when the pretreatment operation such as scaling, data message also can change, in addition can complete failure, Thus affecting the effectiveness of extracted feature.Need to extract and do not change with the operation such as the translation of image, rotation, scaling Invariant moment features information, these information include 5 hu not bending moments altogether, specific as follows shown:
Corn kernel graphical representation is distributed function f (x, y), zeroth order square m00Represent the quality of gray level, single order Square (m01, m10) represent, (xc, yc) central coordinate of circle is initial point, centralized moments of image is expressed as:
mpq=∫ ∫ [(x-xc)p]×[(y-yc)q] f (x, y) dxdy (11)
Wherein, p and q represents the exponent number of square, then hu not bending moment { i1, i2, i3, i4, i5By multiple high-orders not bending moment combine and Become, each High Order Moment is calculated by central moment and gets, and computing formula is as follows:
i1=m20+m02(12)
i2=(m20-m02)2+(2m11)2(13)
i3=(m30-3m12)2+(3m21-m03)2(14)
i4=(m30+m12)2+(m21+m03)2(15)
i 5 = ( m 30 - 3 m 12 ) ( m 30 + m 12 ) [ ( m 30 + m 12 ) 2 - 3 ( m 03 + m 21 ) 2 ] + ( 3 m 21 - m 03 ) ( m 03 + m 21 ) [ 3 ( m 30 + m 12 ) 2 - ( m 03 + m 21 ) 2 ] - - - ( 16 )
(3) support vector cassification: as shown in fig. 6, being realized to corn kernel quality and multiple using support vector machine The detection of defect, is described in detail below:
1. using the geometric properties of training corn kernel sample used and invariant moment features as input, it is expressed as x1, x2..., xd；
2. adopt Radial basis kernel functionFeature is carried out inner product meter Calculate, by low dimensional Feature Mapping to high-dimensional, realize high-dimensional linear separability；
3. calculate the cumulative of feature inner product and mapped using sgn function, obtain last classification results f (x), represent As follows:
f ( x ) = sgn { σ i = 1 y i α i k ( x i , x ) + b } - - - ( 17 )
sgn ( x ) = 1 x > 0 0 x = 0 - 1 x < 0 - - - ( 18 )
Wherein, yiRepresent character pair xiSpecies, αiFor Lagrangian, b compensates for real number, by all seed figures Feature as getting sequentially inputs, you can judge the species of this seed.
It is above-mentioned that but embodiments of the present invention are not limited by the above for the present invention preferably embodiment, its His any spirit without departing from the present invention and the change made under principle, modification, replacement, combine, simplify, all should be The substitute mode of effect, is included within protection scope of the present invention.

Claims (4)

1. a kind of corn kernel quality vision inspection apparatus, including transmission flitch, jitter sources, material scraping plate, shading case, photoelectric sensing Device, lighting source, ccd camera and drive mechanism are it is characterised in that wherein:
Described jitter sources are located at transmission flitch front end；Described material scraping plate is located at transmission flitch middle-end；Described transmission flitch is located at and passes Above motivation structure；Described drive mechanism is by positioned at the Timing Belt transmitting flitch bottom and the motor group being connected with Timing Belt Become；Described transmission flitch, jitter sources, material scraping plate, photoelectric sensor are respectively positioned on inside shading case；Described photoelectric sensor is by being located at The photoemitter of shading case both sides and photelectric receiver composition；Described lighting source is located at the shading case left and right sides；Described ccd Camera is located at directly over shading case, and ccd camera mirror plane is towards the table top of described transmission flitch.
2. corn kernel quality vision inspection apparatus according to claim 1 are it is characterised in that described drive mechanism bag Include motor and Timing Belt, the main shaft of motor is connected with the driving pulley in Timing Belt；Motor main shaft drives same Step band motion, Timing Belt drives the corn kernel to be detected on transmission flitch to move so that ccd phase function in the straight direction Enough realize image acquisition.
3. a kind of corn kernel quality vision detection system, including control centre's module, image analysis module, lighting module and Motion-control module is it is characterised in that wherein:
Control centre's module includes arm embedded platform and control centre's program, and it is embedded flat that control centre's program runs on arm On platform, arm embedded platform passes through serial ports and is connected respectively at image analysis module, lighting module, motion-control module；
Image analysis module includes dsp mainboard, ccd camera, analog monitor and image analysis system, and image analysis system runs On dsp mainboard, ccd camera, analog monitor are connected dsp mainboard respectively；
Lighting module includes lighting source, illumination driver and illumination driver, and illumination driver runs on illumination and drives Device, illumination driver is connected with lighting source；
Motion-control module includes motion controller, motion control program, photoelectric sensor and motor, motion control program Run on motion controller, motion controller is connected with photoelectric sensor and motor respectively.
4. the Semen Maydiss of the corn kernel quality vision inspection apparatus described in a kind of application any one of claim 1-3 and system realization Grain quality visible detection method is it is characterised in that comprise the following steps:
(1) image acquisition:
When corn kernel is through photoelectric sensor, the dsp being connected with photoelectric sensor drives ccd camera to carry out image acquisition, The image collecting is carried out image procossing by dsp again；
(2) image procossing: early stage process is carried out to the corn kernel image collecting, to extract the feature in corn kernel image Information, as the input of image recognition；
(3) image recognition: according to the characteristic information of image procossing, corn kernel image is identified, identifies corn kernel Quality and presence defect.
CN201610632800.4A 2016-07-28 2016-07-28 Visual detection apparatus and method for quality of corn kernel Withdrawn CN106370667A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610632800.4A CN106370667A (en) 2016-07-28 2016-07-28 Visual detection apparatus and method for quality of corn kernel

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610632800.4A CN106370667A (en) 2016-07-28 2016-07-28 Visual detection apparatus and method for quality of corn kernel

Publications (1)

Publication Number Publication Date
CN106370667A true CN106370667A (en) 2017-02-01

Family

ID=57878629

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610632800.4A Withdrawn CN106370667A (en) 2016-07-28 2016-07-28 Visual detection apparatus and method for quality of corn kernel

Country Status (1)

Country Link
CN (1) CN106370667A (en)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106705638A (en) * 2017-02-28 2017-05-24 安徽广深机电设备有限公司 Grain drying system
CN106895699A (en) * 2017-02-28 2017-06-27 安徽广深机电设备有限公司 A kind of grain drier control system
CN108204977A (en) * 2016-12-19 2018-06-26 广东技术师范学院 Corn kernel quality automatic detection device based on machine vision
CN108875747A (en) * 2018-06-15 2018-11-23 四川大学 A kind of wheat unsound grain recognition methods based on machine vision
CN109948405A (en) * 2017-12-21 2019-06-28 中玉金标记（北京）生物技术股份有限公司 Identification seed direction method based on artificial intelligence
CN110108715A (en) * 2019-05-06 2019-08-09 哈尔滨理工大学 A kind of defect inspection method of Plane-parallel Transparent Materiel
CN111127238A (en) * 2020-01-07 2020-05-08 北京印刷学院 Agricultural product quality detection platform
CN111579427A (en) * 2020-05-22 2020-08-25 山东农业大学 Method and system for measuring density of internal components of corn grains
CN113390801A (en) * 2021-04-28 2021-09-14 中国农业科学院农产品加工研究所 On-line detection system and method for optical nondestructive evaluation of quality of irregular meat
CN117808806A (en) * 2024-02-29 2024-04-02 德睦熙睿生物科技(天津)有限公司 Feed production quality refinement detection method based on image feature analysis

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1891354A (en) * 2006-04-29 2007-01-10 江西农业大学 Intelligent grading device for small-particle agricultural product materials
CN101701916A (en) * 2009-12-01 2010-05-05 中国农业大学 Method for quickly identifying and distinguishing variety of corn
CN102095733A (en) * 2011-03-02 2011-06-15 上海大学 Machine vision-based intelligent detection method for surface defect of bottle cap
JP4747602B2 (en) * 2005-02-17 2011-08-17 セントラル硝子株式会社 Glass substrate inspection apparatus and inspection method
CN204746905U (en) * 2015-06-24 2015-11-11 中国农业大学 Automatic check out system of solid seed grain
CN106238337A (en) * 2016-07-28 2016-12-21 广东技术师范学院 A kind of corn kernel quality vision inspection apparatus

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4747602B2 (en) * 2005-02-17 2011-08-17 セントラル硝子株式会社 Glass substrate inspection apparatus and inspection method
CN1891354A (en) * 2006-04-29 2007-01-10 江西农业大学 Intelligent grading device for small-particle agricultural product materials
CN101701916A (en) * 2009-12-01 2010-05-05 中国农业大学 Method for quickly identifying and distinguishing variety of corn
CN102095733A (en) * 2011-03-02 2011-06-15 上海大学 Machine vision-based intelligent detection method for surface defect of bottle cap
CN204746905U (en) * 2015-06-24 2015-11-11 中国农业大学 Automatic check out system of solid seed grain
CN106238337A (en) * 2016-07-28 2016-12-21 广东技术师范学院 A kind of corn kernel quality vision inspection apparatus

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
赵敏: "基于机器视觉的玉米品质检测", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *
齐新: "基于机器视觉的稻米品质检测研究及其DSP实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108204977A (en) * 2016-12-19 2018-06-26 广东技术师范学院 Corn kernel quality automatic detection device based on machine vision
CN106705638A (en) * 2017-02-28 2017-05-24 安徽广深机电设备有限公司 Grain drying system
CN106895699A (en) * 2017-02-28 2017-06-27 安徽广深机电设备有限公司 A kind of grain drier control system
CN109948405A (en) * 2017-12-21 2019-06-28 中玉金标记（北京）生物技术股份有限公司 Identification seed direction method based on artificial intelligence
CN108875747A (en) * 2018-06-15 2018-11-23 四川大学 A kind of wheat unsound grain recognition methods based on machine vision
CN110108715A (en) * 2019-05-06 2019-08-09 哈尔滨理工大学 A kind of defect inspection method of Plane-parallel Transparent Materiel
CN111127238A (en) * 2020-01-07 2020-05-08 北京印刷学院 Agricultural product quality detection platform
CN111579427A (en) * 2020-05-22 2020-08-25 山东农业大学 Method and system for measuring density of internal components of corn grains
CN113390801A (en) * 2021-04-28 2021-09-14 中国农业科学院农产品加工研究所 On-line detection system and method for optical nondestructive evaluation of quality of irregular meat
CN113390801B (en) * 2021-04-28 2023-03-14 中国农业科学院农产品加工研究所 On-line detection system and method for optical nondestructive evaluation of quality of irregular meat
CN117808806A (en) * 2024-02-29 2024-04-02 德睦熙睿生物科技(天津)有限公司 Feed production quality refinement detection method based on image feature analysis
CN117808806B (en) * 2024-02-29 2024-05-03 德睦熙睿生物科技(天津)有限公司 Feed production quality refinement detection method based on image feature analysis

Similar Documents

Publication Publication Date Title
CN106370667A (en) Visual detection apparatus and method for quality of corn kernel
CN106733701B (en) Potato multi-stage combination dysnusia detecting system and detection method
CN102445456B (en) Device and method for nondestructively detecting egg quality based on DSP (Dynamic Signal Processing) machine vision
CN102495069B (en) Method for detecting defects of chain belts of zipper on basis of digital image processing
CN1485616A (en) Fowl eggs quality non-destruction automatic detection grading apparatus and process
CN111709333B (en) Tracing early warning system based on abnormal excrement of cage-raised chickens and health monitoring method
CN111805541B (en) Deep learning-based traditional Chinese medicine decoction piece cleaning and selecting device and cleaning and selecting method
CN103645573A (en) Liquid crystal display (LCD) detection method and system based on machine vision
CN113145492A (en) Visual grading method and grading production line for pear appearance quality
CN114354637A (en) Fruit quality comprehensive grading method and device based on machine vision and X-ray
CN108318494B (en) The red online vision detection and classification devices and methods therefor for proposing fruit powder
CN101916366B (en) Brooding-before clear egg distinguishing method and device thereof
CN210071686U (en) Fruit grading plant based on orthogonal binocular machine vision
CN111220544A (en) Lens quality detection device and detection method
CN109509173A (en) A kind of method of counting and counting device of bacterial clump
CN107036542A (en) A kind of ring gear internal-and external diameter appearance detecting method and device
CN108204977A (en) Corn kernel quality automatic detection device based on machine vision
CN111239142A (en) Paste appearance defect detection device and method
CN105118028B (en) A kind of grafting seam identification device of full-automatic grafting machine and seam recognition methods
CN106238337A (en) A kind of corn kernel quality vision inspection apparatus
CN105195438A (en) Embedded type automatic pearl sorting device and method based on image recognition
CN110575973B (en) Crop seed quality detection and screening system
Zhao et al. A backlight and deep learning based method for calculating the number of seeds per silique
CN114577805A (en) MiniLED backlight panel defect detection method and device
CN109115775A (en) A kind of betel nut level detection method based on machine vision

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WW01 Invention patent application withdrawn after publication
WW01 Invention patent application withdrawn after publication

Application publication date: 20170201