CN108896574A - A kind of bottled liquor method for detecting impurities and system based on machine vision - Google Patents
A kind of bottled liquor method for detecting impurities and system based on machine vision Download PDFInfo
- Publication number
- CN108896574A CN108896574A CN201810451360.1A CN201810451360A CN108896574A CN 108896574 A CN108896574 A CN 108896574A CN 201810451360 A CN201810451360 A CN 201810451360A CN 108896574 A CN108896574 A CN 108896574A
- Authority
- CN
- China
- Prior art keywords
- image
- bottled liquor
- moving target
- impurity
- bottled
- 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
-
- 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
- G01N21/90—Investigating the presence of flaws or contamination in a container or its contents
- G01N21/9018—Dirt detection in containers
- G01N21/9027—Dirt detection in containers in containers after filling
Landscapes
- 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)
- Image Processing (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention discloses a kind of bottled liquor method for detecting impurities and system based on machine vision, move wine liquid using overturning mode first, and acquire the consecutive image of wine liquid movement, pass through improved second order difference algorithm again, isolate moving target, finally according to moving target profile attributes, judge whether bottled liquor contains impurity, this detection mode not only can be to avoid a large amount of bubble of generation, interference effect brought by bubble is reduced, and solves background and body bring interference effect;Therefore the high-precision detection to bottled liquor impurity may be implemented.
Description
Technical field
The invention belongs to liquor determination technical fields, are related to a kind of bottled liquor method for detecting impurities based on machine vision
And system.
Background technique
In the generating process of bottle of liquor, due to generating technique, encapsulation technology etc., it is likely to occur in wine liquid some miscellaneous
Matter.Currently used liquid impurity detection method includes artificial lamp test, semi-automatic lamp test, laser beam detection method, optical-electronic
(resistance) detection method, machine vision detection method etc.;Machine vision detection method, it is easy to operate in view of its, that detection accuracy is high etc. is excellent
Point is widely used in beer detection, Liquid detection of alcoholic drink and pharmaceuticals industry etc..
Machine vision detection method be visual sensing, Digital Image Processing, pattern-recognition with manually can only technology combine
Non-contact detection method, principle is:It is the moving image that rotation is obtained by video camera first, then by computer vision
System handles the moving image of acquisition【Including being filtered using filtering method (such as intermediate value or mean filter) to image
Part required for handling and being extracted using edge detection method】, finally to computer vision system processing by altimetric image into
Row uh, to obtain required detection information (such as acceptance or rejection).Currently based on the detection device factory of Machine Vision Detection
Mainly there are German Seidenada, Miho, Heuft, Italian Brevetti, GF, CMP, Japanese Esia, Toshiba etc. in family.It is specific and
Speech, the defects inspecting equipment based on machine vision can be divided into intermittent visual monitoring, tracking mode visual monitoring and camera again and fix
The defects inspecting of mode.Intermittent vision-based detection refers to that equipment intermittent operation, i.e. detection main rotary table use " walking-stopping-to walk-stop "
Cycle operation mode, detected product enters after high speed rotation station takes pictures a little, the acquisition sequences of the cameras such as n millisecond of pause
It is further continued for operating after image, the type, which represents equipment, the low capacity Medicine inspection machine of Italian Brevetti, Italy's CMP lamp
Inspection machine etc. predominantly detects loading amount between 1~20mL, and speed is generally 150~300 bottles/minute, such type has acquisition
The relatively easy advantage of image stabilization, mechanical device, but it is slower to detect speed simultaneously, is easier to miss to products with larger capacity detection
Sentence or missing inspection.Tracking mode visual detection equipment refers to that detection wheel disc at the uniform velocity operates, and imaging device tracks examined product
It takes pictures, obtains sequence image, track up next batch detected product after then quickly returning.Such detection system it is main
Feature is that detection main rotary table is continuously run, and operates or uses the operating of reflective mirror with angular speed by camera, make product image
It is opposing stationary in sequence image, and impurity has relative displacement, thus judges positive substandard products.The type equipment mainly has
Filtec, Miho, Seidenada, Brevett, Pharmamech, Bosch etc., major part manufacturer is regarded using tracking mode both at home and abroad
Feel detection, due to using track up, detection speed than it is intermittent faster, generally in 200-600 bottle/minute, detection loading amount is 1
Within the scope of~500 milliliters.The advantages of tracking mode is that speed is fast, disengaging bottle bottle explosion rate it is low, but its mechanical device it is more complicated and with
Track, which is taken pictures, needs accurate stabilization.The defects inspecting of camera fixed form is fixed using video camera, shooting when detected product is passed through
Whether a small amount of photo is compared with the standard photographs of storage in a program, to judge in detected product containing obvious different
Object, this mode is fastest, can fully meet the requirement of on-line checking on high-speed automated production line, but due to using and marking
The algorithm thinking of quasi- photo comparison, therefore detection accuracy is not high, can only identify the obvious foreign matter in solution.
In addition, for transparency liquids impurity such as oral solution, injection drugs, the method for detecting impurities master based on machine vision
Sequential frame image is obtained using video camera using the method for rotation emergency stop, processing then is carried out to image and finds moving target,
Foreign matter identification is finally carried out using support vector machines or other track algorithms (such as meanshift algorithm), completes impurity
Detection.However, this coloured primarily directed to beer, health liquor, oral solution etc. based on the method for detecting impurities of machine vision
The detection of liquid, there are no the detections specifically for bottled liquor impurity;It is based particularly on rotation emergency stop mode, it is easily anxious in rotation
A large amount of bubbles can be generated during stopping, and seriously affect defects inspecting precision in white wine.
Therefore, the existing liquid impurity detection method based on machine vision is perfect not enough, and it is miscellaneous not to be able to satisfy liquor industry
The demand of quality detection.
Summary of the invention
The purpose of the present invention is intended to provide a kind of bottled liquor method for detecting impurities based on machine vision, realizes to bottled
The accurate detection of impurity in white wine.
Another object of the present invention is intended to provide a kind of bottled liquor Impurity Detection System based on machine vision.
For bottled liquor, particularly circular transparent bottled liquor, the present invention provides a kind of based on machine vision
Bottled liquor method for detecting impurities, first acquisition continuous multiple frames image, then are handled to obtain containing movement to multiple image
The image of target completes the detection to bottled liquor impurity further according to moving target outline shape.The present invention further uses one
The improved second order difference algorithm of kind is realized under a large amount of jamming patterns using the continuity of sequence image over time and space
The problem of being partitioned into thin objects.
Bottled liquor method for detecting impurities provided by the invention based on machine vision, includes the following steps:
(1) bottled liquor is overturn 180 °;
(2) video is acquired to the bottled liquor after overturning;
(3) N frame image, each frame image and previous frame image or next frame image are extracted from the sequential frame image of video
Between be spaced M frame image, wherein N is positive odd number, and M is integer more than or equal to 0;
(4) the N frame image of extraction is pre-processed, removes picture noise;
(5) it is analyzed using the image that second order difference algorithm obtains pretreatment, extracts the figure containing moving target
Picture;
(6) profile of moving target is extracted, and calculates the ratio of profile long axis and short axle;
(7) obtained profile ratio of long axis to short axis is compared with the threshold value of setting, if profile ratio of long axis to short axis belong to 1.0~
1.1, moving target is bubble;If profile ratio of long axis to short axis is greater than 1.4, moving target is impurity.
The above-mentioned bottled liquor method for detecting impurities based on machine vision, generates big to reduce due to liquid motion
Bubble is measured, the present invention moves wine liquid by the way of 180 ° of overturning.It, should relative to the rotation emergency stop mode generallyd use
Method can be realized the generation that bubble is reduced while liquid and impurity that may be present in bottle move.Especially work as use
When being inverted 180 ° of overturning (being first inverted body, under bottleneck is super, body is then overturn 180 ° again), impurity can be made to be located at bottle
Near neck, overturning can make impurity as far as possible near central axes, and the feelings of missing inspection occur close to bottle wall due to impurity to reduce
Condition.
The above-mentioned bottled liquor method for detecting impurities based on machine vision, and provided through testing repeatedly, detect light used
Source and lighting method are related to dopant species, acquire video from bottled liquor front, when the impurity of detection is white, to bottled
White wine using dark background, bottom to light lighting method;When the impurity of detection is black, using back to the illumination side of light
Formula;It can be conducive to obtain high-quality video image in this way.For round bottle, it is also easy to produce the reflection and refraction effect of light, from
And a large amount of interference is generated around bottle wall;In order to prevent due to round bottle this physical characteristic generate interference,
One layer of optical filtering paper is set before into the light source of bottle, to reduce interference of the light to imaging.
The above-mentioned bottled liquor method for detecting impurities based on machine vision improves bottled in order to ensure image processing effect
White wine defects inspecting accuracy includes the sequential frame image of at least three frames in the video.
The above-mentioned bottled liquor method for detecting impurities based on machine vision is ensuring bottled liquor defects inspecting accuracy
Meanwhile detection efficiency is improved, the N is 3 or 5, and the value of the M is:2≤M≤5.
The above-mentioned bottled liquor method for detecting impurities based on machine vision removes picture noise to eliminate background interference,
The N frame image of extraction is pre-processed, remove picture noise, including it is following step by step:
(41) Gamma correction is carried out to the N frame image that lifts, and the image after Gamma is corrected is converted into gray level image;
(42) it is handled using the gray level image that weighted mean filter algorithm obtains step (41), filters out image and make an uproar
Sound.
Above-mentioned steps (41) carry out Gamma correction to the N frame image of extraction, with adjust bright dark areas contrast (such as
Reduce the contrast of highlight regions), thus the interference that reflection and refraction effect of the more efficient reduction due to light generate.It can adopt
The conventional Gamma bearing calibration disclosed with this field.Image after Gamma is corrected again is converted into gray level image.
The purpose of above-mentioned steps (42) is removal picture noise, the conventional filtering algorithm that can have been disclosed using this field
Removal noise is filtered to image, the present invention is using weighted mean filter algorithm (Li Xiufeng, Soviet Union based on intermediate value
Lan Hai, Rong Huifang, Chen Hua improve Mean Filtering Algorithm and application study [J] microcomputer information, 2008 (01):235-236+
202.) obtained gray level image is filtered, detection target is retained while removing picture noise, so that inspection
The result of survey is more accurate.The above-mentioned bottled liquor method for detecting impurities based on machine vision, the present invention are calculated using second order difference
Method handles step (4) pretreated image, including it is following step by step:
(51) by N frame image, adjacent two field pictures carry out calculus of differences and obtain (N-1) width difference image;
(52) by (N-1) width difference image, the adjacent two width difference image of every two is divided into one group, and two width in every group are poor
Partial image carries out calculus of differences respectively and energy accumulation operation obtains associated second order difference image and energy accumulation image;
(53) obtained energy accumulation image second order difference image subtraction associated therewith is obtained containing movement mesh
Target image.
By using second order difference algorithm, the moving target in bottle outlet can be separated, solves the static state such as background and body
The interference of subject image bring.
The above-mentioned bottled liquor method for detecting impurities based on machine vision in step (5), contains moving target to extraction
Image using Otsu ' s maximum variance between clusters carry out binarization threshold processing, to efficiently separate background image, as far as possible
Eliminate background interference;And the bright spot in image less than 4 pixels is removed with morphology opening operation 2 × 2.
The above-mentioned bottled liquor method for detecting impurities based on machine vision, moving target profile long axis are referred respectively to short axle
The length and width of moving target profile boundary rectangle.
The above-mentioned bottled liquor method for detecting impurities based on machine vision, in order to distinguish impurity and bubble, according to miscellaneous
The features of shape of matter and bubble is classified, and when moving target profile ratio of long axis to short axis belongs to 1.0~1.1, moving target is true
It is set to bubble;When moving target profile long axis and short axle ratio are when being greater than 1.1 and being less than or equal between 1.4, it may be possible to bubble
It is also likely to be impurity, therefore in order to improve the accuracy to defects inspecting, profile ratio of long axis to short axis is greater than to 1.4 moving target
It is determined as impurity.
When cannot achieve the differentiation of impurity or bubble based on moving target profile, can repeat the above steps (1)~
(7), it is detected again.
Invention further provides a kind of bottled liquor Impurity Detection System based on machine vision, including:
Image collecting device, for acquiring video to 180 ° of overturning of bottled liquor;
Image processing apparatus is handled for the video to image acquisition device, is completed to bottled liquor impurity
Detection;Described image processing unit includes:
Frame image extraction unit extracts N frame image, each frame image and previous frame image from the sequential frame image of video
Or M frame image is spaced between next frame image, wherein N is positive odd number, and M is the positive integer more than or equal to 0;
Image pre-processing unit pre-processes the N frame image of extraction, removes picture noise;
Movement destination image extraction unit is handled using the image that second order difference algorithm obtains pretreatment, is extracted
Out containing the image of moving target;
Contours extract and ratio of long axis to short axis computing unit, extract the profile of moving target, and calculate profile long axis and short axle
Ratio;
Judging unit compares obtained profile ratio of long axis to short axis with the threshold value of setting, if profile ratio of long axis to short axis category
In 1.0~1.1, moving target is bubble;If profile ratio of long axis to short axis is greater than 1.4, moving target is impurity.
The above-mentioned bottled liquor Impurity Detection System based on machine vision, the detection system are provided with black defects inspecting station
And/or white defects inspecting station, black defects inspecting station and/or white defects inspecting station configure an image collecting device;It is right
180 ° of bottled liquor is overturn in the Image Acquisition window alignment of black detection station, image collecting device, while utilizing light source
From bottled liquor back to light;For white detection station, the Image Acquisition window alignment of image collecting device overturns 180 °
Bottled liquor, while utilizing light source from bottled liquor bottom to light, and bottled liquor back uses dark background.Due to black impurity
It is more readily detected and detection accuracy is higher, first black impurity is detected, then dialogue colored foreign is detected, can improve
Detection accuracy is further increased while detection efficiency.Therefore, when setting black impurity detection station and white detection station,
Bottled liquor is first entered black detection station to detect, if containing black impurity, is directly filtered out bottled liquor, no
Need to enter next white defects inspecting station;If, into white defects inspecting station, judged whether without black impurity
Contain white impurity.
Compared with prior art, the invention has the advantages that:
(1) present invention moves wine liquid using overturning mode first, and acquires the consecutive image of wine liquid movement, then by changing
Into second order difference algorithm, isolate moving target, finally according to moving target profile attributes, judge whether bottled liquor contains
There is impurity, this detection mode not only can reduce interference effect brought by bubble, and solve to avoid a large amount of bubble is generated
Background of having determined and body bring interference effect;Therefore the high-precision detection to bottled liquor impurity may be implemented;
(2) bottled liquor is overturn it so that impurity be made to be located near bottleneck in inversion state before overturning by the present invention
Rear impurity is further reduced due to impurity is close to bottle wall that there is a situation where missing inspections as far as possible near central axes;
(3) present invention can realize the detection to two class impurity of black and white in bottled liquor respectively, ensure defects inspecting essence
Detection efficiency is improved while spending;
(4) present invention is carrying out data processing to the bottled liquor wine liquid consecutive image of acquisition, extracts moving target process
In, by the way that optical filtering paper and the Gamma in later period correction is arranged at light source rear, it is anti-that the light due to caused by light is former can be effectively reduced
Penetrate and reflect interference;
(5) present invention is further carried out using image containing moving target of Otsu ' the s maximum variance between clusters to extraction
Thresholding processing, can efficiently separate out background, further decrease the interference of background;
(6) present invention can rely on existing equipment to realize, easy to operate, efficient, be suitable in liquid impurity, particularly bottle
Dress is promoted the use of in white wine defects inspecting field.
Detailed description of the invention
Fig. 1 is bottled liquor white defects inspecting flow diagram of the embodiment of the present invention 1 based on machine vision.
Fig. 2 is that the present invention is based on the bottled liquor defects inspecting schematic diagrames of machine vision;Wherein (a) bottled liquor inversion is turned over
Turn schematic diagram, (b) bottled liquor white defects inspecting schematic diagram, (c) bottled liquor black impurity detection schematic diagram;In figure, 1-
Video camera, 2- bottled liquor, 3- red filter paper, 4-LED light source, 5- black barn door.
Fig. 3 is the flow diagram for extracting moving target in the embodiment of the present invention 1 using second order difference algorithm.
Fig. 4 is that the present invention is based on the bottled liquor defects inspecting effect diagrams of machine vision;Wherein (a), (e), (i) and
(i1) acquisition image, second order difference processing image, moving target profile and the movement of fiber impurity in bottled liquor are respectively corresponded
Objective contour enlarged drawing, (b), (f), (j) and (j1) respectively correspond the acquisition image of black slag impurity, second order difference in bottled liquor
Image, moving target profile and moving target profile enlarged drawing are handled, (c), (g), (k) and (k1) respectively correspond in bottled liquor
The acquisition image of glass chip, second order difference processing image, moving target profile and moving target profile enlarged drawing, (d), (h),
(l) and (l1) respectively corresponds the acquisition image of White Flocculus impurity in bottled liquor, second order difference processing image, moving target
Profile and moving target profile enlarged drawing.
Fig. 5 is bottled liquor black impurity testing process schematic diagram of the embodiment of the present invention 2 based on machine vision.
Fig. 6 is the flow diagram for extracting moving target in the embodiment of the present invention 2 using second order difference algorithm.
Specific embodiment
The embodiment of the present invention is provided below with reference to attached drawing, and technical solution of the present invention is carried out into one by embodiment
Clear, the complete explanation of step.Obviously, the embodiment is only a part of the embodiments of the present invention, rather than whole implementation
Example.Based on the content of present invention, those of ordinary skill in the art are obtained all without making creative work
Other embodiments belong to the range that the present invention is protected
Embodiment 1
The present embodiment is for detecting the white impurity such as glass chip, fiber hair, to the bottled liquor based on machine vision
Defects inspecting process is described in detail.
Bottled liquor white defects inspecting process provided in this embodiment, as shown in Figure 1, including the following steps:
(1) bottled liquor is overturn 180 °.
Due to bottled liquor body there are scale, decorative pattern, the impurity of absorption and during production and transport because collision
Etc. a plurality of types of interference such as microgroove for generating of reasons so that the small foreign matter in detection wine liquid becomes difficult.Such as Fig. 2 (a) institute
Show, be first inverted the bottled liquor 2 containing white impurity, bottled liquor 2 is overturn 180 ° when into area to be tested, it is this
It is inverted the mode of entrance of overturning, the existing interference of bottle is opposing stationary in multiple images, and liquid in bottle and may deposit
Impurity move along 2 central axes of bottled liquor due to flipping upside down for bottle body from being located near bottleneck, obtain sequence image.This
Sample can not only be reduced due to impurity close to bottle wall and there is a situation where missing inspections, relative to existing rotation emergency stop mode, is produced
Raw bubble is less, to reduce interference of the bubble to defects inspecting to the greatest extent, improves defects inspecting precision.
(2) video is acquired to the bottled liquor after overturning.
The present embodiment shoots video using the bottled liquor 2 after 1 pair of high-definition camera overturning.In order to improve dialogue colored foreign
Detection accuracy, as shown in Fig. 2 (b), for the present embodiment when detecting white impurity, 2 back of bottled liquor uses black barn door 5
Dark background is formed, LED light source 4 is irradiated from 2 bottom of bottled liquor, while adding one layer of red filter paper 3 in the front of LED light source 4
(used high-definition digital camera is more preferable to red effect), to reduce interference of the light to imaging.
(3) 3 frame images, each frame image and previous frame image or next frame image are extracted from the sequential frame image of video
Between be spaced 2 frame images, in this way can while ensuring to extract moving target from image, reduce due to interference caused by
Uncertain factor bring error.Those skilled in the art can be according to impurity moving speed, two consecutive frame figure of appropriate adjustment
The frame number being spaced as between.
(4) 3 frame images of extraction are pre-processed, remove picture noise, including it is following step by step:
(41) routine Gamma is carried out to the 3 frame images that lift to correct, reduce the contrast of highlight regions, and by the school Gamma
Image after just is converted into gray level image;
(42) it is handled, is filtered using the gray level image that the weighted mean filter algorithm based on intermediate value obtains step (41)
Retain object to be measured while except picture noise.
The present embodiment is included the following steps using the weighted average filtering algorithm based on intermediate value:
(A) with X=[x (i, j)]M×NIndicate input picture, wherein x (i, j) indicates (i, j) point in image grayscale value matrix
Locate the gray value of pixel, W3(i, j) represent center pixel (i, j), size as 3 × 3 a window (due to impurity it is smaller because
This takes 3 × 3, and this field can choose suitable window according to impurity size is estimated).Then
(B) W is used first3(i, j) is at X=[x (i, j)]M×NW is read in middle scanning3The gray value x of each pixel in (i, j)
(i, j), and gray value is sorted by size, take maximum value MAX and the minimum value MIN in gray value.
(C) if x (i, j) is equal to maximum value MAX or minimum value MIN, enabling multiplication weight is 0, i.e., does not consider that they are right
The influence of image, then according to formula
A=x1(i,j)×m1+x2(i,j)×m2+x3(i,j)×m3+x4(i,j)×m4+x5(i,j)×m3+x6(i,j)×
m2+x7(i,j)×m1
A is obtained, wherein x1(i,j)、x2(i,j)...x7(i, j) be remove maximum value MAX or minimum value MIN from it is small to
The gray value of longer spread, m1、m2、m3、m4For weight corresponding to each gray value, weight value rule is m4These corresponding ashes
Intermediate value in angle value takes coefficient maximum, m3、m2、m1It is gradually reduced, and m1、m2、m3、m4Belong to (0,1), m1、m2、m3、m4It can be with
It is set according to above-mentioned rule.
(D) it enables
And Z value is assigned to pixel center point x (i, j)=Z in window institute scanning area.
Obtained gray level image is handled according to above-mentioned steps (A)-(D), it can filtering image noise.
(5) it is analyzed using the image that second order difference algorithm obtains pretreatment, extracts the figure containing moving target
Picture.
This step is analyzed using the image that a kind of improved second order difference algorithm obtains pretreatment, is isolated small
Moving target, as shown in figure 3, specifically including following steps:
(51) by three frame images, adjacent two field pictures carry out calculus of differences and obtain two width difference images.
With f (x, y, t-2), f (x, y, t), f (x, y, t+2) indicate step (4) treated three frame images, first frame figure
As being spaced 2 frame images, the second frame image f (x, y, t) and third frame figure between f (x, y, t-2) and the second frame image f (x, y, t)
As being spaced 2 frame images between f (x, y, t+2).First frame image f (x, y, t-2) and the second frame image f (x, y, t) do absolute difference,
Second frame image f (x, y, t) and third frame image f (x, y, t+2) do absolute difference, respectively obtain two width difference image d(t-2,t)
(x, y) and d(t,t+2)(x,y);Here operation completion can be carried out by the function absdiff () called directly in OpenCV.
d(t-2,t)(x, y)=| f (x, y, t)-f (x, y, t-2) |
d(t,t+2)(x, y)=| f (x, y, t)-f (x, y, t+2) |
(52) two width difference images are subjected to calculus of differences again and obtain second order difference image, two width difference images are carried out
Energy accumulation operation obtains energy accumulation image.
By two width difference image d in the present embodiment(t-2,t)(x, y) and d(t,t+2)(x, y) do absolute difference obtain image D (x,
y);Operation completion can be carried out by the function absdiff () called directly in OpenCV.
D (x, y)=| d(t-2,t)(x,y)-d(t,t+2)(x,y)|。
By two width difference image d in the present embodiment(t-2,t)(x, y) and d(t,t+2)(x, y) carries out energy accumulation and obtains image E
(x,y);Operated by the function addWeighted (src1, scr2,1, result) called directly in OpenCV
At src1 and src2 is respectively image d here(t-2,t)(x, y) and d(t,t+2)(x,y)。
E (x, y)=d(t-2,t)(x,y)+d(t,t+2)(x,y)。
(53) obtained energy accumulation image second order difference image subtraction associated therewith is obtained containing movement mesh
Target image.
Image E (x, y) and image D (x, y) are subtracted each other into the gray level image F obtained containing moving target in the present embodiment
(x,y).It can be operated by calling directly the function addWeighted in OpenCV (src1, scr2, -1, result)
It completes, wherein src1 and src2 is respectively image E (x, y) and image D (x, y) here.
F (x, y)=E (x, y)-D (x, y).
Therefore, the moving target that the enhancing of an intermediate frame is only remained in the image obtained, reduces body itself and exists
Interference.
In order to eliminate as much as background interference, the present embodiment further uses Otsu ' s maximum variance between clusters to extraction
Image containing moving target carries out binarization threshold processing, to efficiently separate background image;And use morphology opening operation 2
Less than the bright spot of 4 pixels in × 2 removal images.
(6) profile of moving target is extracted, and calculates the ratio of profile long axis and short axle.
In the present embodiment, moving target profile long axis refers respectively to the length and width of moving target profile boundary rectangle with short axle
(7) obtained profile ratio of long axis to short axis is compared with the threshold value of setting, if profile ratio of long axis to short axis belong to 1~
1.1, moving target is bubble;If profile ratio of long axis to short axis is greater than 1.4, moving target is impurity.
In order to distinguish impurity and bubble, classified according to the features of shape of impurity and bubble, when moving target wheel
When wide ratio of long axis to short axis belongs to 1.0~1.1, moving target is determined as bubble;When moving target profile long axis and short axle ratio exist
Greater than 1.1 and when being less than or equal between 1.4, it may be possible to which bubble is also likely to be impurity, therefore in order to improve the standard to defects inspecting
Exactness, the moving target by profile ratio of long axis to short axis greater than 1.4 are determined as impurity.
When cannot achieve the differentiation of impurity or bubble based on moving target profile, can repeat the above steps (1)~
(7), it is detected again.
It is imitated in order to illustrate the detection of the bottled liquor white method for detecting impurities provided in this embodiment based on machine vision
Fruit is that white impurity is added in bottled liquor with fiber hair, glass chip and White Flocculus, according to above-mentioned steps (1)-
(7) dialogue colored foreign is detected, obtained corresponding acquisition image (Fig. 4 (a), (c) and (d)), second order difference processing image (figure
4 (e), (g) and (h)), moving target profile (Fig. 4 (i), (k) and (l)) and moving target profile enlarged drawing (Fig. 4 (i1), (k1)
(l1)).It is provided from Fig. 4 (i1), the long axis and short axle ratio of moving target profile are 6.3, therefore, it is determined that moving target is
Impurity;It is provided from Fig. 4 (k1), the long axis and short axle ratio of moving target profile are 1.6, therefore, it is determined that moving target is miscellaneous
Matter;It is provided from Fig. 4 (l1), the long axis and short axle ratio of moving target profile are 2.0, therefore, it is determined that moving target is impurity.
The above analysis result is consistent substantially with the contaminant size of setting.Therefore, the detection method provided through this embodiment may be implemented
Accurate detection to bottled liquor white impurity.
Embodiment 2
The present embodiment for detecting black slag black impurity, to the bottled liquor defects inspecting process based on machine vision into
Row detailed description.
Bottled liquor black impurity detection process provided in this embodiment, as shown in figure 5, including the following steps:
(1) bottled liquor is overturn 180 °.
Due to bottled liquor body there are scale, decorative pattern, the impurity of absorption and during production and transport because collision
Etc. a plurality of types of interference such as microgroove for generating of reasons so that the small foreign matter in detection wine liquid becomes difficult.Such as Fig. 2 (a) institute
Show, be first inverted the bottled liquor 2 containing white impurity, bottled liquor 2 is overturn 180 ° when into area to be tested, it is this
It is inverted the mode of entrance of overturning, the existing interference of bottle is opposing stationary in multiple images, and liquid in bottle and may deposit
Impurity move along 2 central axes of bottled liquor due to flipping upside down for bottle body from being located near bottleneck, obtain sequence image.This
Sample can not only be reduced due to impurity close to bottle wall and there is a situation where missing inspections, relative to existing rotation emergency stop mode, is produced
Raw bubble is less, to reduce interference of the bubble to defects inspecting to the greatest extent, improves defects inspecting precision.
(2) video is acquired to the bottled liquor after overturning.
The present embodiment shoots video using the bottled liquor 2 after 1 pair of high-definition camera overturning.In order to improve to black impurity
Detection accuracy, as shown in Fig. 2 (c), the present embodiment when detecting black impurity, LED light source 4 from 2 back illuminated of bottled liquor,
In such a way that back is to light.
(3) 5 frame images, each frame image and previous frame image or next frame image are extracted from the sequential frame image of video
Between be spaced 5 frame images, in this way can while ensuring to extract moving target from image, reduce due to interference caused by
Uncertain factor bring error.Those skilled in the art can be according to impurity moving speed, two consecutive frame image of appropriate adjustment
Between the frame number that is spaced.
(4) 5 frame images of extraction are pre-processed, remove picture noise, including it is following step by step:
(41) routine Gamma is carried out to the 5 frame images that lift to correct, reduce the contrast of highlight regions, and by the school Gamma
Image after just is converted into gray level image;
(42) it is handled, is filtered using the gray level image that the weighted mean filter algorithm based on intermediate value obtains step (41)
Retain object to be measured while except picture noise.
The present embodiment is included the following steps using the weighted average filtering algorithm based on intermediate value:
(A) with X=[x (i, j)]M×NIndicate input picture, wherein x (i, j) indicates (i, j) point in image grayscale value matrix
Locate the gray value of pixel, W3(i, j) represent center pixel (i, j), size as 3 × 3 a window (due to impurity it is smaller because
This takes 3 × 3, and this field can choose suitable window according to impurity size is estimated).Then
(B) W is used first3(i, j) is at X=[x (i, j)]M×NW is read in middle scanning3The gray value x of each pixel in (i, j)
(i, j), and gray value is sorted by size, take maximum value MAX and the minimum value MIN in gray value.
(C) if x (i, j) is equal to maximum value MAX or minimum value MIN, enabling multiplication weight is 0, i.e., does not consider that they are right
The influence of image, then according to formula
A=x1(i,j)×m1+x2(i,j)×m2+x3(i,j)×m3+x4(i,j)×m4+x5(i,j)×m3+x6(i,j)×
m2+x7(i,j)×m1
A is obtained, wherein x1(i,j)、x2(i,j)...x7(i, j) be remove maximum value MAX or minimum value MIN from it is small to
The gray value of longer spread, m1、m2、m3、m4For weight corresponding to each gray value, weight value rule is m4These corresponding ashes
Intermediate value in angle value takes coefficient maximum, m3、m2、m1It is gradually reduced, and m1、m2、m3、m4Belong to (0,1), m1、m2、m3、m4It can be with
It is set according to above-mentioned rule.
(D) it enables
And Z value is assigned to pixel center point x (i, j)=Z in window institute scanning area.
Obtained gray level image is handled according to above-mentioned steps (A)-(D), it can filtering image noise.
(5) it is analyzed using the image that second order difference algorithm obtains pretreatment, extracts the figure containing moving target
Picture.
This step is analyzed using the image that a kind of improved second order difference algorithm obtains pretreatment, is isolated small
Moving target, as shown in fig. 6, specifically including following steps:
(51) by five frame images, adjacent two field pictures carry out calculus of differences and obtain two width difference images.
With f (x, y, t-10), f (x, y, t-5), f (x, y, t), f (x, y, t+5), f (x, y, t+10) are indicated at step (4)
Five frame images after reason are spaced 5 frame images between adjacent two field pictures.Adjacent two field pictures do absolute difference, obtain four width difference
Image (difference image 1 to difference image 4);Here it can be grasped by the function absdiff () called directly in OpenCV
It completes.
(52) four width difference images are divided into two groups, two width difference images in every group carry out calculus of differences and energy respectively
Accumulation operation obtains associated second order difference image and energy accumulation image.
Difference image 1 and difference image 2 are done into absolute difference in the present embodiment and obtain image second order difference 1, it can be by straight
It connects and the function absdiff () in OpenCV is called to carry out operation completion.Difference image 1 and difference image 2 are subjected to energy accumulation
Obtain energy accumulation 1, can by call directly the function addWeighted in OpenCV (src1, scr2,1, result) into
Row operation is completed, and src1 and src2 is respectively difference image 1 and difference image 2 here
Difference image 3 and difference image 4 are done into absolute difference in the present embodiment and obtain image second order difference 2, it can be by straight
It connects and the function absdiff () in OpenCV is called to carry out operation completion.Difference image 3 and difference image 4 are subjected to energy accumulation
Obtain energy accumulation 2, can by call directly the function addWeighted in OpenCV (src1, scr2,1, result) into
Row operation is completed, and src1 and src2 is respectively difference image 3 and difference image 4 here.
(53) obtained energy accumulation image second order difference image subtraction associated therewith is obtained containing movement mesh
Target image.
The image 1 obtained containing moving target is subtracted each other into energy accumulation 1 and second order difference 1 in the present embodiment, by energy
Accumulation 2 subtracts each other the image 2 obtained containing moving target with second order difference 2, these can be by calling directly in OpenCV
Function addWeighted (src1, scr2, -1, result) carries out operation completion, and wherein src1 is energy accumulation 1 or energy here
2, scr2 of amount accumulation is second order difference 1 or second order difference 2.
Therefore, centre one is only remained in the image 1 containing moving target and the image containing moving target 2 obtained
The moving target of the enhancing of frame reduces the existing interference of body itself.And the image 2 containing moving target with containing movement mesh
Logo image 1 is fortune function state of the same movement target in different time, can not only be reduced because uncertain caused by interference
Factor bring error, while two of same moving target are detected with the accuracy that can increase judgement.
In order to eliminate as much as background interference, the present embodiment further uses Otsu ' s maximum variance between clusters to extraction
Image 1 containing moving target and the image containing moving target 2 carry out binarization threshold processing, to efficiently separate Background
Picture;And the bright spot in image less than 4 pixels is removed with morphology opening operation 2 × 2.
(6) profile of moving target is extracted, and calculates the ratio of profile long axis and short axle.
In the present embodiment, any one in the image 1 containing moving target and the image containing moving target 2 is research pair
As calculating the long axis and short axle of moving target profile.
(7) obtained profile ratio of long axis to short axis is compared with the threshold value of setting, if profile ratio of long axis to short axis belong to 1~
1.1, moving target is bubble;If profile ratio of long axis to short axis is greater than 1.4, moving target is impurity.
In order to distinguish impurity and bubble, classified according to the features of shape of impurity and bubble, when moving target wheel
When wide ratio of long axis to short axis belongs to 1.0~1.1, moving target is determined as bubble;When moving target profile long axis and short axle ratio exist
Greater than 1.1 and when being less than or equal between 1.4, it may be possible to which bubble is also likely to be impurity, therefore in order to improve the standard to defects inspecting
Exactness, the moving target by profile ratio of long axis to short axis greater than 1.4 are determined as impurity.
When cannot achieve the differentiation of impurity or bubble based on moving target profile, can repeat the above steps (1)~
(7), it is detected again.
The present embodiment is using black slag as black impurity, obtained corresponding acquisition image, second order difference processing image, movement mesh
It marks profile and moving target profile enlarged drawing is respectively Fig. 4 (b), (f), (j), (j1), provided from Fig. 4 (j1), moving target
The long axis and short axle ratio of profile are 1.7, therefore, it is determined that moving target is impurity.The contaminant size of above analysis result and setting
Substantially it is consistent.Therefore, the accurate detection to bottled liquor black impurity may be implemented in the detection method provided through this embodiment.
Embodiment 3
A kind of bottled liquor Impurity Detection System based on machine vision is present embodiments provided, may be implemented to bottled white
The detection of wine black impurity and white impurity, the detection system include:
Black defects inspecting station, including bottled liquor bottle support member and LED light source, LED light source is from bottled liquor
Back is to light;
White defects inspecting station, including bottled liquor bottle support member, LED light source, red filter paper and black shading
Plate, the part that bottle support member is directed at bottle bottom is transparent configuration, in order to light transmission;LED light source is from bottled liquor bottom
Portion is provided with red filter paper before LED light source to light;Black barn door is set to bottled liquor back, to make bottled white
Wine back is dark background;
Image collecting device, for overturning 180 ° bottled liquor acquire video, can for high-definition digital video camera,
CCD industrial camera etc.;
Image processing apparatus is handled for the video to image acquisition device, is completed to bottled liquor impurity
Detection;Described image processing unit includes:
Frame image extraction unit extracts N frame image, each frame image and previous frame image from the sequential frame image of video
Or M frame image is spaced between next frame image, wherein N is positive odd number, and M is the positive integer more than or equal to 0;
Image pre-processing unit pre-processes the N frame image of extraction, removes picture noise;
Movement destination image extraction unit is handled using the image that second order difference algorithm obtains pretreatment, is extracted
Out containing the image of moving target;
Contours extract and ratio of long axis to short axis computing unit, extract the profile of moving target, and calculate profile long axis and short axle
Ratio;
Judging unit compares obtained profile ratio of long axis to short axis with the threshold value of setting, if profile ratio of long axis to short axis category
In 1.0~1.1, moving target is bubble;If profile ratio of long axis to short axis is greater than 1.4, moving target is impurity.
An image collecting device can be configured with black defects inspecting station and white defects inspecting station, it can also be only with
One image collecting device, needs to configure the mobile platform for carrying image collecting device at this time, enables image collecting device
It is enough to switch between different detection stations.Since black impurity is more readily detected and detection accuracy is higher, first to black impurity into
Row detection, then dialogue colored foreign are detected, and can further increase detection accuracy while improving detection efficiency.Therefore,
When setting black impurity detection station and white detection station, bottled liquor is first entered into black detection station and is detected,
If directly filtered out bottled liquor containing black impurity, do not need to enter next white defects inspecting station;If not yet
There is black impurity, into white defects inspecting station, judges whether containing white impurity.
It is above-mentioned by frame image extraction unit, image pre-processing unit, movement destination image extraction unit, contours extract and length
Short axle ratio calculation unit, judging unit composition image processing apparatus can be loaded into computer or with image processing function
In processor, to complete the functions such as image procossing and impurity judgement.
Claims (10)
1. a kind of bottled liquor method for detecting impurities based on machine vision, it is characterised in that include the following steps:
(1) bottled liquor is overturn 180 °;
(2) video is acquired to the bottled liquor after overturning;
(3) N frame image is extracted from the sequential frame image of video, between each frame image and previous frame image or next frame image
It is spaced M frame image, wherein N is positive odd number, and M is the positive integer more than or equal to 0;
(4) the N frame image of extraction is pre-processed, removes picture noise;
(5) it is analyzed using the image that second order difference algorithm obtains pretreatment, extracts the image containing moving target;
(6) profile of moving target is extracted, and calculates the ratio of profile long axis and short axle;
(7) obtained profile ratio of long axis to short axis is compared with the threshold value of setting, if profile ratio of long axis to short axis belongs to 1.0~1.1,
Moving target is bubble;If profile ratio of long axis to short axis is greater than 1.4, moving target is impurity.
2. the bottled liquor method for detecting impurities according to claim 1 based on machine vision, it is characterised in that when to black
When colored foreign detects, the bottled liquor after overturning is adopted using back to light, while to the bottled liquor after overturning in step (2)
Collect video.
3. the bottled liquor method for detecting impurities according to claim 1 based on machine vision, it is characterised in that work as dialogue
When colored foreign detects, use dark background, bottom to light the bottled liquor after overturning in step (2), while to the bottle after overturning
It fills white wine and acquires video.
4. special according to claim 1 to the bottled liquor method for detecting impurities described in 3 any claims based on machine vision
Sign is the sequential frame image in the video comprising at least three frames.
5. the bottled liquor method for detecting impurities based on machine vision according to claim 4, it is characterised in that the N is 3
Or 5, the value of the M is:2≤M≤5.
6. the bottled liquor method for detecting impurities based on machine vision according to claim 1, it is characterised in that the step
(4) include it is following step by step:
(41) Gamma correction is carried out to the N frame image that lifts, and the image after Gamma is corrected is converted into gray level image;
(42) it is handled using the gray level image that weighted mean filter algorithm obtains step (41), filtering image noise.
7. the bottled liquor method for detecting impurities based on machine vision according to claim 1, it is characterised in that step (5) packet
Include it is following step by step:
(51) by N frame image, adjacent two field pictures carry out calculus of differences and obtain (N-1) width difference image;
(52) by (N-1) width difference image, the adjacent two width difference image of every two is divided into one group, two width difference diagrams in every group
As progress calculus of differences and energy accumulation operation obtain associated second order difference image and energy accumulation image respectively;
(53) obtained energy accumulation image second order difference image subtraction associated therewith is obtained containing moving target
Image.
8. according to claim 1 or the 7 bottled liquor method for detecting impurities based on machine vision, it is characterised in that step
(5) in, binarization threshold processing is carried out using Otsu ' s maximum variance between clusters to the image containing moving target of extraction, and
The bright spot in image less than 4 pixels is removed with morphology opening operation 2 × 2.
9. a kind of bottled liquor Impurity Detection System based on machine vision, it is characterised in that including:
Image collecting device, for acquiring video to 180 ° of overturning of bottled liquor;
Image processing apparatus is handled for the video to image acquisition device, completes the inspection to bottled liquor impurity
It surveys;Described image processing unit includes:
Frame image extraction unit, extracts N frame image from the sequential frame image of video, each frame image and previous frame image or under
M frame image is spaced between one frame image, wherein N is positive odd number, and M is the positive integer more than or equal to 0;
Image pre-processing unit pre-processes the N frame image of extraction, removes picture noise;
Movement destination image extraction unit is handled using the image that second order difference algorithm obtains pretreatment, extracts and contain
There is the image of moving target;
Contours extract and ratio of long axis to short axis computing unit, extract the profile of moving target, and calculate the ratio of profile long axis and short axle
Value;
Judging unit compares obtained profile ratio of long axis to short axis with the threshold value of setting, if profile ratio of long axis to short axis belongs to 1.0
~1.1, moving target is bubble;If profile ratio of long axis to short axis is greater than 1.4, moving target is impurity.
10. the bottled liquor Impurity Detection System based on machine vision according to claim 9, it is characterised in that the detection system
System is provided with black defects inspecting station and/or white defects inspecting station, and black defects inspecting station and/or white defects inspecting station are matched
Set an image collecting device;
For black detection station, the Image Acquisition window alignment of image collecting device overturns 180 ° of bottled liquor, while benefit
With light source from bottled liquor back to light;
For white detection station, the Image Acquisition window alignment of image collecting device overturns 180 ° of bottled liquor, while benefit
With light source from bottled liquor bottom to light, and bottled liquor back uses dark background.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810451360.1A CN108896574B (en) | 2018-05-11 | 2018-05-11 | Bottled liquor impurity detection method and system based on machine vision |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810451360.1A CN108896574B (en) | 2018-05-11 | 2018-05-11 | Bottled liquor impurity detection method and system based on machine vision |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108896574A true CN108896574A (en) | 2018-11-27 |
CN108896574B CN108896574B (en) | 2020-03-17 |
Family
ID=64342840
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810451360.1A Active CN108896574B (en) | 2018-05-11 | 2018-05-11 | Bottled liquor impurity detection method and system based on machine vision |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108896574B (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109741282A (en) * | 2019-01-16 | 2019-05-10 | 清华大学 | A kind of multiframe bubble stream image processing method based on Predictor Corrector |
CN110006917A (en) * | 2019-05-10 | 2019-07-12 | 陈昌涛 | A kind of online vision detection system of bottled liquor impurity and its implementation |
CN110811325A (en) * | 2019-11-13 | 2020-02-21 | 重庆科文净水设备有限公司 | Intelligent drinking water system |
WO2020188730A1 (en) * | 2019-03-19 | 2020-09-24 | 日本電気株式会社 | Inspection device, inspection method, and non-transitory computer-readable medium |
CN111830044A (en) * | 2019-04-19 | 2020-10-27 | 无锡维胜威信息科技有限公司 | White spirit quality detection system |
CN111950466A (en) * | 2020-08-13 | 2020-11-17 | 北京福通互联科技集团有限公司 | Bottled liquor detection method and detection device based on trajectory cascade algorithm |
CN112381776A (en) * | 2020-11-09 | 2021-02-19 | 深圳前海微众银行股份有限公司 | Method and device for determining impurities of contents in transparent container |
CN112418105A (en) * | 2020-11-25 | 2021-02-26 | 湖北工业大学 | High maneuvering satellite time sequence remote sensing image moving ship target detection method based on difference method |
CN113284116A (en) * | 2021-05-28 | 2021-08-20 | 荆门汇易佳信息科技有限公司 | Real-time image analysis type red wine semi-finished product motion detection system |
CN114124626A (en) * | 2021-10-15 | 2022-03-01 | 西南交通大学 | Signal noise reduction method and device, terminal equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1588015A (en) * | 2004-09-28 | 2005-03-02 | 上海金纳德制药设备科技发展有限公司 | On-line lamp detecting method and its device using machine visual system |
CN101403705A (en) * | 2008-11-17 | 2009-04-08 | 湖南大学 | Foreign body machine vision recognition method and device for assembly line special-shaped bottling liquid |
CN101413904A (en) * | 2008-11-25 | 2009-04-22 | 湖南大学 | Impurity real time machine vision detection method and apparatus in special-shaped bottle-packaging solution |
CN101806752A (en) * | 2010-03-12 | 2010-08-18 | 长沙图创机电科技有限公司 | Method and equipment for visual detection of visible foreign matters in bottled liquid medicine |
-
2018
- 2018-05-11 CN CN201810451360.1A patent/CN108896574B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1588015A (en) * | 2004-09-28 | 2005-03-02 | 上海金纳德制药设备科技发展有限公司 | On-line lamp detecting method and its device using machine visual system |
CN101403705A (en) * | 2008-11-17 | 2009-04-08 | 湖南大学 | Foreign body machine vision recognition method and device for assembly line special-shaped bottling liquid |
CN101413904A (en) * | 2008-11-25 | 2009-04-22 | 湖南大学 | Impurity real time machine vision detection method and apparatus in special-shaped bottle-packaging solution |
CN101806752A (en) * | 2010-03-12 | 2010-08-18 | 长沙图创机电科技有限公司 | Method and equipment for visual detection of visible foreign matters in bottled liquid medicine |
Non-Patent Citations (2)
Title |
---|
张辉 等: "基于机器视觉的保健酒可见异物检测系统研究与开发", 《仪器仪表学报》 * |
陈文广: "基于机器视觉的酒液异物在线检测技术研究", 《万方学位论文数据库》 * |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109741282A (en) * | 2019-01-16 | 2019-05-10 | 清华大学 | A kind of multiframe bubble stream image processing method based on Predictor Corrector |
JP7238965B2 (en) | 2019-03-19 | 2023-03-14 | 日本電気株式会社 | Inspection device, inspection method, and program |
WO2020188730A1 (en) * | 2019-03-19 | 2020-09-24 | 日本電気株式会社 | Inspection device, inspection method, and non-transitory computer-readable medium |
JPWO2020188730A1 (en) * | 2019-03-19 | 2020-09-24 | ||
CN111830044A (en) * | 2019-04-19 | 2020-10-27 | 无锡维胜威信息科技有限公司 | White spirit quality detection system |
CN110006917A (en) * | 2019-05-10 | 2019-07-12 | 陈昌涛 | A kind of online vision detection system of bottled liquor impurity and its implementation |
CN110811325A (en) * | 2019-11-13 | 2020-02-21 | 重庆科文净水设备有限公司 | Intelligent drinking water system |
CN111950466A (en) * | 2020-08-13 | 2020-11-17 | 北京福通互联科技集团有限公司 | Bottled liquor detection method and detection device based on trajectory cascade algorithm |
CN112381776A (en) * | 2020-11-09 | 2021-02-19 | 深圳前海微众银行股份有限公司 | Method and device for determining impurities of contents in transparent container |
CN112418105A (en) * | 2020-11-25 | 2021-02-26 | 湖北工业大学 | High maneuvering satellite time sequence remote sensing image moving ship target detection method based on difference method |
CN113284116A (en) * | 2021-05-28 | 2021-08-20 | 荆门汇易佳信息科技有限公司 | Real-time image analysis type red wine semi-finished product motion detection system |
CN114124626A (en) * | 2021-10-15 | 2022-03-01 | 西南交通大学 | Signal noise reduction method and device, terminal equipment and storage medium |
CN114124626B (en) * | 2021-10-15 | 2023-02-17 | 西南交通大学 | Signal noise reduction method and device, terminal equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN108896574B (en) | 2020-03-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108896574A (en) | A kind of bottled liquor method for detecting impurities and system based on machine vision | |
CN110146516B (en) | Fruit grading device based on orthogonal binocular machine vision | |
CN111189837B (en) | Cigarette appearance online detection method and device | |
CN111696077A (en) | Wafer defect detection method based on wafer Det network | |
CN110310275B (en) | Chain conveyor belt defect detection method based on image processing | |
CN112452801A (en) | High-efficiency online detection device for defects of glass bottles and working method thereof | |
CN102162797A (en) | Algorithm for detecting glass bottle neck damage and bottle bottom dirt | |
CN104835166A (en) | Liquid medicine bottle foreign matter detection method based on machine visual detection platform | |
CN105784722A (en) | Method and system of detecting visible foreign matters in medicine liquid bottles on flow line | |
CN113252568A (en) | Lens surface defect detection method, system, product and terminal based on machine vision | |
JP2004525340A (en) | Inspection system and inspection method for container having opening | |
CN105973904A (en) | Edible oil impurity detection method based on image background probability graph | |
CN109360195A (en) | The visible detection method of foreign particles in a kind of bottle-packaging solution | |
CN109709108B (en) | Bottled transparent liquid impurity detection equipment and method | |
CN101825582A (en) | Method and device for detecting wall of cylindrical transparent bottle | |
CN109507205A (en) | A kind of vision detection system and its detection method | |
Fu et al. | Medicine glass bottle defect detection based on machine vision | |
CN112129783A (en) | Transparent medicine bottle bottom defect detection device and detection method based on deep learning | |
CN201819889U (en) | Full-automatic intelligent lamp detection system | |
CN108520260A (en) | The recognition methods of visible foreign matters in bottled oral solution | |
CN109406539A (en) | A kind of transparent medicine bottle bottom buildup defect detecting system and method | |
CN102024143B (en) | Method for tracking and identifying medicinal liquid impurity images on high-speed medicament production line | |
CN110044919A (en) | A kind of detection device and its detection method for mirror surfaces scratch | |
CN109626302A (en) | A kind of online test method of beer foreign matter in bottles | |
CN209606340U (en) | Transparent medicine bottle bottom buildup defect detecting system |
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 |