CN108961251A - A kind of raw cotton fault and defects inspecting and recognition methods and system - Google Patents
A kind of raw cotton fault and defects inspecting and recognition methods and system Download PDFInfo
- Publication number
- CN108961251A CN108961251A CN201810818716.0A CN201810818716A CN108961251A CN 108961251 A CN108961251 A CN 108961251A CN 201810818716 A CN201810818716 A CN 201810818716A CN 108961251 A CN108961251 A CN 108961251A
- Authority
- CN
- China
- Prior art keywords
- image
- gray
- fault
- obtains
- unit
- 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.)
- Pending
Links
- 229920000742 Cotton Polymers 0.000 title claims abstract description 100
- 238000000034 method Methods 0.000 title claims abstract description 60
- 230000007547 defect Effects 0.000 title claims abstract description 23
- 239000012535 impurity Substances 0.000 claims abstract description 50
- 238000003709 image segmentation Methods 0.000 claims abstract description 20
- 238000003708 edge detection Methods 0.000 claims abstract description 18
- 230000002708 enhancing effect Effects 0.000 claims abstract description 12
- 239000000284 extract Substances 0.000 claims abstract description 7
- 238000012545 processing Methods 0.000 claims description 42
- 238000001914 filtration Methods 0.000 claims description 25
- 238000004422 calculation algorithm Methods 0.000 claims description 21
- 230000005540 biological transmission Effects 0.000 claims description 17
- 238000005286 illumination Methods 0.000 claims description 13
- 230000001629 suppression Effects 0.000 claims description 13
- 230000011218 segmentation Effects 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 10
- 230000009977 dual effect Effects 0.000 claims description 8
- 230000000877 morphologic effect Effects 0.000 claims description 8
- 230000003044 adaptive effect Effects 0.000 claims description 7
- 230000007797 corrosion Effects 0.000 claims description 7
- 238000005260 corrosion Methods 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 7
- 238000003672 processing method Methods 0.000 claims description 7
- 238000007689 inspection Methods 0.000 claims description 2
- 238000001514 detection method Methods 0.000 abstract description 39
- 230000006872 improvement Effects 0.000 abstract description 4
- 230000000694 effects Effects 0.000 description 19
- 239000000835 fiber Substances 0.000 description 15
- 230000009286 beneficial effect Effects 0.000 description 10
- 238000005516 engineering process Methods 0.000 description 10
- 238000002474 experimental method Methods 0.000 description 5
- 238000012935 Averaging Methods 0.000 description 4
- 230000033001 locomotion Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000011049 filling Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 239000013589 supplement Substances 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 239000003245 coal Substances 0.000 description 2
- 239000002019 doping agent Substances 0.000 description 2
- 238000009987 spinning Methods 0.000 description 2
- 208000003464 asthenopia Diseases 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 230000008451 emotion Effects 0.000 description 1
- 210000002615 epidermis Anatomy 0.000 description 1
- 238000005530 etching Methods 0.000 description 1
- 238000005194 fractionation Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000001678 irradiating effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003333 near-infrared imaging Methods 0.000 description 1
- 238000011017 operating method Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000005622 photoelectricity Effects 0.000 description 1
- 239000000049 pigment Substances 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000009738 saturating Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 210000003491 skin Anatomy 0.000 description 1
- 238000010183 spectrum analysis Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 239000004753 textile Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G06T5/70—
-
- G06T5/94—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20032—Median filtering
Abstract
The present invention relates to a kind of raw cotton fault and defects inspecting and recognition methods and systems, and wherein method includes: irradiation sample cotton upper and lower surface, the color image of the sample cotton upper and lower surface after acquisition irradiation；The color image is pre-processed, gray level image is obtained；Image procossing is carried out to the gray level image, obtains target image；The characteristic value of the target image is extracted, and the characteristic value of the target image is compared with presetting database, obtains the recognition result of fault and impurity and display.Identification and classification by computer to the acquisition of cotton image, the enhancing of target image and improvement, image segmentation, edge detection and target image, extract validity feature value, realize the detection and identification to fault contained by raw cotton and impurity, improve detection and accuracy of identification, it can identify a plurality of types of raw cotton faults and impurity, the drawbacks of eliminating manual sorting reduces detection and identification cost.
Description
Technical field
The present invention relates to fiber check and measure technical field more particularly to a kind of raw cotton fault and defects inspecting and recognition methods and
System.
Background technique
Raw cotton Harmful Defect and impurity mainly include mote, rope silk, soft seed epidermis, bearded mote, stiff piece, cotton knot with
And chemical fibre, silk, hair, plastic ties, fiber crops, colored fibre etc..Most of rejecting in China, for raw cotton impurity and fault
Cotton spinning enterprise at present still by the way of manual sorting, need to carry out cotton packet fractionation one by one by worker and pick, not only work
Work amount is heavy, takes time and effort, inefficiency, and detects effect and be affected by human factors quite greatly.In general, manual sorting's raw cotton
Impurity fault requires that a biggish place is specially arranged, and quantity and the qualification of worker will tie in, allow in this way
Financial resources investment increases；And worker, because of the high-intensitive same movement of repetition, subjective emotion can generate the fluctuation of certain amplitude simultaneously
There is different degrees of visual fatigue；Cotton after sorting simultaneously, which also needs to be packaged again, could put into spinning link use, this
The relatively complicated complexity of one process.For large-scale Cotton Textile Enterprises, the cotton that need to be sorted daily is innumerable, so
Huge workload is even more so that manual sorting is difficult.
There are many research method at present about raw coal nature context of detection, and the method and technique means of use mainly include light
Spectral analysis technology, ultraviolet light detection, ultrasonic technology, near-infrared imaging technology, sensor technology, X-ray transmission, machine view
Feel technology and Visible-light CCD imaging etc..For artificial vision's sorting, the range that machine can perceive is broader,
Whether detect speed or effect it is reasonable rapidly compared with manual sorting, in this respect the technology of American-European some countries just more at
It is ripe perfect, have developed relevant raw coal nature detector, such as the different fine detection machine of Jossi System company of Switzerland
Vision Shield Inspect4 and MagicEye M1, Te Lvcile Textile Co., Ltd. of Germany foreign matter ejector
SECUROMAT SCFO, Loptex company of Italy photoelectricity ultrasonic wave raw cotton de-burring machine Single, Tandem and H.P. tri-
Series, the relatively good different fiber removing machine Fibre i for mainly having Bremen Corporation of India that furthermore Asia is done, Chinese longitude and latitude are spun
Knit the foreign fiber sorter JWF0011 and foreign fiber and micro-dust separator JWF1051 of Machinery Co., Ltd..
Though ultrasonic technology can identify various different fibres, it is slower that it detects speed, and it is lesser different to detect area
Property fiber；It may recognize that foreign fiber of different sizes using optical sensor technology, but be limited only to foreign pigment, for saturating
Bright or white different fibre can not but identify；External mechanical vision inspection technology is to very thin raw cotton impurity and white fault
Detection effect be not obvious, be not suitable for China cotton contain miscellaneous situation.
In terms of academic research, Jin Shoufeng et al. utilizes high performance CCD video camera in the base of MATLAB image processing techniques
Completion realizes the detection and identification of cotton foreign fiber on plinth.Computer is former according to the collected foreign fiber in raw cotton of CCD camera
Beginning picture extracts relevant parameter, carries out edge detection to the different fibre in raw cotton using the method that Canny algorithm and dual threshold combine,
And the approximate location of foreign fiber in raw cotton is obtained using center of gravity positioning mode；Tae etc. chooses colourful CCD video camera and acquires moits figure
Piece selects the smoits in suitable threshold value identification raw cotton, mostly broken leaf or kind skin according to 8 neighborhood processings, and can calculate
The quantity and distribution of impurity.Although the above scholar can substantially position to raw cotton impurity and quantity statistics, in raw cotton defect
The aspect that accurately identifies of point and impurity also needs to continue deeper into research.
Summary of the invention
The technical problem to be solved by the present invention is to solve the above shortcomings of the prior art and to provide a kind of raw cotton faults and miscellaneous
Quality detection and recognition methods and system.
The technical scheme to solve the above technical problems is that a kind of raw cotton fault and defects inspecting and identification side
Method, comprising the following steps:
Step 1: irradiation sample cotton upper and lower surface, the color image of the sample cotton upper and lower surface after acquisition irradiation；
Step 2: the color image being pre-processed, gray level image is obtained；
Step 3: image procossing being carried out to the gray level image, obtains target image；
Step 4: extract the characteristic value of the target image, and by the characteristic value of the target image and presetting database into
Row compares, and obtains the recognition result of fault and impurity and display.
The beneficial effects of the present invention are: color image is pre-processed by acquisition of the computer to cotton image, it is right
Pretreated gray level image carries out image procossing, obtains clearly target image, extracts the validity feature value in target image
Fault contained by raw cotton and impurity are identified, fast to the detection and recognition speed of raw cotton fault and impurity, Detection accuracy is high,
The drawbacks of capable of identifying a plurality of types of raw cotton faults and impurity, eliminating manual sorting, reduces detection and identification cost.
Based on the above technical solution, the present invention can also be improved as follows:
Further, pretreated specific steps in the step 2 are as follows:
Step 21: gray processing processing being carried out to the color image in step 1 using weighted average method, described in extraction
Tri- channel components of R, G, B of color image, and tri- channel components of described R, G, B are weighted and averaged, obtain initial ash
It spends image Gray (i, j)；
Step 22: enhancing the contrast of the initial gray image Gray (i, j) using histogram enhancement processing method, obtain
To Continuous Tone Image Z (i, j)；
Step 23: noise removal process being carried out to the Continuous Tone Image Z (i, j) using median filtering, is obtained described
Gray level image.
The beneficial effect of above-mentioned further scheme is: being weighted and averaged using weighted average method to R, G, B three-component
More reasonable gray level image can be obtained, global histogram is formed using histogram enhancement processing method, enhances gray level image
Local contrast makes image become clear, highlights the gray level of target image, facilitate subsequent image procossing.
Further, the step 22 specifically:
Step 221: a histogram equalization processing being carried out to the initial gray image Gray (i, j), draws gray scale
Histogram；
Step 222: the grey level histogram being modified using histogram specification method, obtains the middle gray
Image Z (i, j).
The beneficial effect of above-mentioned further scheme is: forming the overall situation using histogram equalization and histogram specification method
Histogram, enhance the local contrast of gray level image, so that image is become clear, highlight the gray level of target image, facilitate
Subsequent image procossing.
Further, the specific steps of the step 3 are as follows:
Step 31: image segmentation being carried out using gray level image described in adaptive threshold image segmentation, obtains segmentation gray scale
Image T (i, j)；
Step 32: edge detection being carried out to the segmentation gray level image T (i, j) using edge detection operator, obtains edge
Gray level image Y (i, j)；
Step 33: the edge gray table being obtained as Y (i, j) progress Morphological scale-space using corrosion and expansive working
The target image.
The beneficial effect of above-mentioned further scheme is: examining to the enhancing of image degree of comparing and improvement, image segmentation, edge
The filling and repairing of survey and target image, it is ensured that obtain the clearly image of raw cotton fault and impurity.
Further, using the adaptive threshold image segmentation based on Mean Shift algorithm in the step 31.
The beneficial effect of above-mentioned further scheme is: the motion profile of Mean Shift algorithm is one and is intended to convergence point
Smooth route, by Mean Shift algorithm carry out mould point search, mould point cluster and annex zonule, in target image not
With target area accurately divided, and the processing arithmetic speed of image segmentation is fast, it is available it is smoother clearly
Image, and it is possible to prevente effectively from the result of segmentation is excessively fine crushing, guarantee obtains the target of more visible raw cotton fault and impurity
Image.
Further, step 32 specific steps are as follows:
Step 321: noise removal process is carried out to the segmentation gray level image T (i, j) using 2-d gaussian filters template,
Obtain smoothed image；
Step 322: the gray scale for calculating the smoothed image using first-order difference local derviation is obtained in the partial derivative in the direction x, y
The gradient magnitude and gradient direction of the gray scale of the smoothed image；
Step 323: using non-maxima suppression method, the marginal point of the smoothed image is accurately positioned and refines side
Edge obtains non-maxima suppression image N (i, j)；
Step 324: detecting the edge of the non-maxima suppression image N (i, j) using dual threashold value-based algorithm, obtain the side
Edge gray level image Y (i, j).
The beneficial effect of above-mentioned further scheme is: being calculated using gaussian filtering, first-order difference local derviation, non-maxima suppression
The step of method and dual threashold value-based algorithm, carries out edge detection, and arithmetic speed is fast, is not limited by directionality and noise, precision
Height can also refine edge, the operation without additional refinement edge, so that it may obtain more visible in addition to enhancing edge strength
Accurate edge gray table is as Y (i, j).
Further, the characteristic value in the step 4 includes the eccentricity of the target image and the face of the target image
Product perimeter ratio, the presetting database are stored with the area perimeter of the eccentricity comprising fault and impurity and the fault and impurity
The relation table of ratio.
The beneficial effect of above-mentioned further scheme is: learnt by many experiments, the eccentricity of different faults and impurity with
There are certain rules for the correlation of area perimeter ratio, by fault and the eccentricity of impurity and area perimeter than one-to-one
Relation table, calculates the eccentricity and area perimeter ratio of target image, and is compared with relation table, available fault and
The recognition result of impurity.Since cotton fault and dopant species are more, complex shape, the feature of different types of fault and impurity is not
Together, traditional identification method to fault and impurity needs to set different identification sides for different types of fault and impurity
Formula, or identified using different threshold values, and identification method of the invention is without setting different identification method or threshold value
Different types of fault and impurity are identified well, and detection is high with recognition accuracy, and speed is fast, and practicability is high.
In order to solve technical problem of the invention, a kind of raw cotton fault and defects inspecting and identifying system, packet are additionally provided
Include image acquisition units, image processing unit, image extraction unit, recognition unit and display unit；
Described image acquisition unit for being irradiated to sample cotton upper and lower surface, and acquires on the sample cotton after irradiation
The color image of lower surface；
Described image processing unit obtains gray level image, also for pre-processing the collected color image
For the gray level image to be carried out image procossing, target image is obtained；
The extraction unit, for extracting the characteristic value of the target image；
The recognition unit obtains fault and miscellaneous for the characteristic value of the target image to be compared with database
The recognition result of matter；
The display unit, for showing the recognition result of the fault and impurity.
The beneficial effects of the present invention are: the color image on sample cotton surface is acquired by image acquisition units, it will be collected
Image is sent to image processing unit, obtains target image after treatment, and the spy of target image is extracted by extraction unit
Value indicative identified by recognition unit, finally shows obtained recognition result by display unit, complete detection to fault with
Identification, detection and recognition speed are fast, and Detection accuracy is high, can identify a plurality of types of raw cotton faults and impurity, eliminate artificial point
The drawbacks of picking reduces detection and identification cost.
Further, described image acquisition unit includes illumination unit, transmission units and Image Acquisition subelement；
The illumination unit is used to irradiate the upper surface of the sample cotton；
The transmission units are used to irradiate the lower surface of the sample cotton；
Described image acquisition subelement is used to acquire the upper surface of the sample cotton after irradiation and the lower surface of the sample cotton
The color image；
The beneficial effect of above-mentioned further scheme is: upper surface of the light to sample cotton of different wave length is generated by illumination unit
It is irradiated, and the lower surface of sample cotton is irradiated as supplement light source by transmission units, guarantee all sites of sample cotton
It can be irradiated to, avoid occurring dash area in image, then following table on sample cotton is irradiated to by the acquisition of Image Acquisition subelement
The color image in face, obtained color image screen are more complete, clear.
Further, described image processing unit includes gray processing unit, contrast enhancement unit, filter unit, image point
Cut unit, edge detection unit and Morphological scale-space unit；
The gray processing unit obtains initial for carrying out gray processing to the color image using weighted average method
Gray level image Gray (i, j)；
The contrast enhancement processing unit, for using histogram enhancement processing method to the initial gray image
The processing enhancing of Gray (i, j) degree of comparing, obtains Continuous Tone Image Z (i, j)；
The filter unit, for being carried out at noise remove using median filtering method to the Continuous Tone Image Z (i, j)
Reason, obtains the gray level image；
Described image cutting unit is obtained for carrying out image segmentation to the gray level image using Mean Shift algorithm
To segmentation gray level image T (i, j)；
The edge detection unit, for carrying out edge to the segmentation gray level image T (i, j) using edge detection operator
Detection, obtains edge gray table as Y (i, j)；
The Morphological scale-space unit, for using corrosion and expansive working to the edge gray table as Y (i, j) is carried out
Filling and repairing obtains the target image.
The beneficial effect of above-mentioned further scheme is: being enhanced by the gray processing unit in image processing unit, contrast
Unit, filter unit, image segmentation unit, edge detection unit and Morphological scale-space unit carry out gray processing, histogram to image
One systems such as figure enhancing processing, median filtering, the image segmentation based on Mean shift algorithm, edge detection and corrosion expansive working
Column processing, it is ensured that obtain clear smooth target image, be convenient for subsequent identification work, improve detection and accuracy of identification.
Detailed description of the invention
Fig. 1 is the flow diagram of a kind of raw cotton fault of the invention and defects inspecting and recognition methods；
Fig. 2-1 is that the filtered image effect figure of 3*3 neighborhood averaging is used in one embodiment of the present of invention；
Fig. 2-2 is in one embodiment of the present of invention using the image effect figure after 3*3 neighborhood Wiener filtering；
Fig. 2-3 is in one embodiment of the present of invention using the image effect figure after 3*3 neighborhood median filtering；
Fig. 3-1 is that the filtered image effect figure of 7*7 neighborhood averaging is used in one embodiment of the present of invention；
Fig. 3-2 is in one embodiment of the present of invention using the image effect figure after 7*7 neighborhood Wiener filtering；
Fig. 3-3 is in one embodiment of the present of invention using the image effect figure after 7*7 neighborhood median filtering；
Fig. 4 is the structural schematic diagram of one embodiment of a kind of raw cotton fault of the present invention and defects inspecting and identifying system.
In attached drawing, parts list represented by the reference numerals are as follows:
1, Image Acquisition subelement, 2, illumination unit, 3, transmission units, 4, sample stage, 5, light-transmitting plate, 6, sample cotton, 7, mirror
Head, 8, display, 9, processor, 10, data line.
Specific embodiment
The principle and features of the present invention will be described below with reference to the accompanying drawings, and the given examples are served only to explain the present invention, and
It is non-to be used to limit the scope of the invention.
With reference to the accompanying drawing, the present invention will be described.
Embodiment one, as shown in Figure 1, showing for a kind of raw cotton fault of the invention and defects inspecting and the process of recognition methods
It is intended to, comprising the following steps:
S1: irradiation sample cotton upper and lower surface, the color image of the sample cotton upper and lower surface after acquisition irradiation；
S2: the color image is pre-processed, gray level image is obtained；
S3: image procossing is carried out to the gray level image, obtains target image；
S4: the characteristic value of the target image is extracted, and the characteristic value of the target image and presetting database are carried out
It compares, obtains the recognition result of fault and impurity and display.
By acquisition of the computer to cotton image in the present embodiment, the color image is pre-processed, to pre- place
Gray level image after reason carries out image procossing, obtains clearly target image, extracts the validity feature value pair of the target image
Fault and impurity contained by raw cotton are identified, fast to the detection and recognition speed of raw cotton fault and impurity, and Detection accuracy is high, energy
The drawbacks of identifying a plurality of types of raw cotton faults and impurity, eliminating manual sorting reduces detection and identification cost.
Preferably, pretreated specific steps in the S2 are as follows:
S21: gray processing processing is carried out to the color image in step 1 using weighted average method, extracts the coloured silk
The component in tri- channels R, G, B of chromatic graph picture, and tri- channel components of described R, G, B are weighted and averaged, obtain initial ash
It spends image Gray (i, j), wherein used weighted average formula are as follows:
Gray (i, j)=0.3R (i, j)+0.59G (i, j)+0.11B (i, j)
Wherein, R (i, j) is the R channel components of the color image, and G (i, j) is the G channel components of the color image,
B (i, j) is the channel B component of the color image；
Based on experience value, R, G, B three-component are weighted and averaged using weighted average method can obtain it is relatively reasonable just
Beginning gray level image Gray (i, j).
S22: enhancing the overall contrast of the initial gray image Gray (i, j) using histogram enhancement processing method,
Obtain Continuous Tone Image Z (i, j)；
S23: noise removal process is carried out to the Continuous Tone Image Z (i, j) using median filtering, obtains the gray scale
Image.
Median filtering is a kind of nonlinear signal processing technology that can effectively inhibit noise based on sequencing statistical theory, it
The intermediate value of all pixels point gray value in the gray value of pixel each in the image point neighborhood is replaced, allows surrounding
Pixel value is close to true value, to eliminate isolated noise spot, and median filtering has preferable protection to make at the edge of image
With solving most linear filterings blurred picture this disadvantage while denoising, obtain preferable image restoration effect.
The present embodiment uses the two dimension pattern plate of 7*7 neighborhood, obtains two dimension median filter output image, the as described grayscale image
Picture:
F (i, j)=med { Z (i-k, j-l), (k, l ∈ W) }
Wherein, using the optimal value that the two dimension pattern plate of 7*7 neighborhood is after test of many times, referring to fig. 2-1,2-2,2-3,
Shown in 3-1,3-2,3-3.The two dimension pattern plate of 3*3 neighborhood and the two dimension pattern plate of 7*7 neighborhood have been respectively adopted in an experiment, has made respectively
It is filtered experiment with average filter, Wiener filtering and median filtering, Fig. 2-1 is using the filtered image of 3*3 neighborhood averaging
Effect picture, Fig. 2-2 be using the image effect figure after 3*3 neighborhood Wiener filtering, Fig. 2-3 be using 3*3 neighborhood median filtering after
Image effect figure, Fig. 3-1 is using the filtered image effect figure of 7*7 neighborhood averaging, and Fig. 3-2 is using 7*7 neighborhood wiener
Filtered image effect figure, Fig. 3-3 are using the image effect figure after 7*7 neighborhood median filtering.Experimental result is shown in 3*3
In neighborhood, there are still noises for image after average filter, and the effect of median filtering is fine, not only eliminates noise, but also smooth
Image；In the neighborhood of 7*7, there is the phenomenon that having obscured image in average filter, but the effect of median filtering is splendid；Again to intermediate value
Two neighborhoods of filtering compare, and can intuitively see very much, and when using 7*7 neighborhood, de-noising and smooth effect are compared with 3*3 neighbour
Domain will be got well, so carrying out denoising to cotton image using 7*7 neighborhood median filtering.
Preferably, the S22 specifically:
S221: a histogram equalization processing is carried out to the initial gray image Gray (i, j), after being equalized
The gray value S of each pixelk, to SkRounding treatment is carried out, the grey level after determining equalization, and calculate corresponding each SkPicture
Prime number mesh draws grey level histogram；Wherein, the gray value S of each pixelkFormula are as follows:
Wherein, rjFor the gray level of the initial gray image Gray (i, j), njFor each gray-scale number of pixels, Pr
(rj) it is each gray-scale probability density after equalization；
Because after histogram equalization processing, originally the gray scale of fewer pixel can be assigned to other gray scale and go, as
Plain Relatively centralized, tonal range becomes larger after processing, and contrast becomes larger, and clarity becomes larger, so can effectively enhance image.Pass through this
Kind method, brightness can be preferably distributed on the histogram, can be used for enhancing local contrast without influencing whole pair
Degree of ratio.
S222: the grey level histogram is modified using histogram specification method, obtains the middle gray figure
As Z (i, j)；
Histogram specification operation is carried out to image, so as to purposefully to certain grey level distribution models in image
Image in enclosing is enhanced, and the histogram of original image and the histogram of desired image are known, PrIt (r) is the present embodiment
The probability density function for the intensity profile that middle original image is once equalized, PzIt (z) is the intensity profile of desired image
Probability density function is modified gained grey level histogram in S221, with Pz(z) shape represented by, it is specific real
It is now as follows:
S2221: the formula of each gray scale in S221 is write as functional integration functional form:
S2222: to desired image histogram equalization processing method described in S221, transforming function transformation function is obtained:
Its inverse process are as follows:
Z=G-1(u)
S2223: due to it is above-mentioned be the equalization to same image, the probability density function of treated original image and phase
Hope the probability density function of image be it is equal, so s=u, and z=G-1(u)=G-1(T (r)), the figure of resulting each gray scale
As the as described Continuous Tone Image Z (i, j).
Preferably, the specific steps of S3 are as follows:
S31: carrying out image segmentation using gray level image described in adaptive threshold image segmentation, obtains segmentation gray level image
T(i,j)；
S32: edge detection is carried out to the segmentation gray level image T (i, j) using edge detection operator, obtains edge gray scale
Image Y (i, j)；
S33: the edge gray table is obtained described as Y (i, j) progress Morphological scale-space using corrosion and expansive working
Target image.
To the enhancing of image degree of comparing and improvement, image segmentation, edge detection and target image filling and repairing, can be with
Guarantee to obtain the clearly image of raw cotton fault and impurity.
The adaptive threshold image segmentation based on Mean Shift algorithm is used in the S31.
To obtain more visible gray level image, need to be split image processing.The characteristic point of one sub-picture at least may be used
To extract five dimensions (x, y, r, g, b), including the space two-dimensional coordinate space (x, y) and three-dimensional color space (r, g, b), and use
Practical Mean shift algorithm is exactly the mould point for finding this quintuple space, since different points can finally converge to different peaks
Value completes the purpose of image segmentation so these points are formed one kind.Mean shift algorithm has the advantages that uniqueness,
Its vector has constringency performance, its motion profile is the smooth route for being intended to extreme point, in Mean shift algorithm
When searching for mould point (extreme point), two are generally all not more than 90 ° with the angle between continuous vector, it can be promoted to move
Stability enhancing, smoothly movement can promote the processing arithmetic speed of image segmentation to accelerate to Mean shift algorithm relatively, obtain
Steady and audible target area information required for iamge description and analysis.Meanwhile Mean shift algorithm not only can be to not
With target area accurately divided, it is often more important that this algorithm it is possible to prevente effectively from divide result it is excessively fine crushing,
It can obtain relatively clear raw cotton fault and impurity image.Specific implementation step is the prior art, and which is not described herein again.
Preferably, the S32 specific steps are as follows:
S321: noise removal process is carried out to the segmentation gray level image T (i, j) using 2-d gaussian filters template, is obtained
To smoothed image；
Since the image border property after adaptive threshold fuzziness is bad, and image border generally comprises some related images
Characteristics extraction and target identification important information, it is therefore desirable to enhance the intensity of image border.Since noise also focuses on
High-frequency signal, it is easy to be identified as pseudo-edge, therefore remove noise using gaussian filtering first, reduce the identification of pseudo-edge.
Gaussian filtering is the prior art, is specifically repeated no more.
S322: calculating partial derivative of the gray scale in the direction x, y of the smoothed image using first-order difference local derviation, obtains described
The gradient magnitude and gradient direction of the gray scale of smoothed image:
Wherein, GxAnd GyThe partial derivative of the gray scale of the respectively described smoothed image in the x and y direction；
S323: using non-maxima suppression method, edge be accurately positioned and refined to the marginal point of the smoothed image,
Obtain non-maxima suppression image N (i, j)；
Since gradient magnitude is located substantially at four direction, respectively represent horizontal, vertical and two diagonal lines (0 °, 45 °,
90 °, 135 °), by the gradient of the gray scale of the pixel on the gradient magnitude of the gray scale of current pixel point and its positive and negative gradient direction
Amplitude is compared, if the gradient magnitude of current pixel point is maximum, is retained its value, as image border point, is otherwise inhibited, will
Its shade of gray value sets 0, as image background.
S324: detecting the edge of the non-maxima suppression image N (i, j) using dual threashold value-based algorithm, if high threshold is τ1,
Low threshold is τ2, the gradient magnitude of the pixel in the non-maxima suppression image N (i, j) is compared with dual threshold and is sentenced
It is fixed, the edge gray table is obtained as Y (i, j).
It is as follows specifically to compare to determine method:
The gradient magnitude of pixel in the non-maxima suppression image N (i, j) is compared judgement with dual threshold,
Then gradient magnitude is greater than τ1Pixel be image border (strong edge point), gradient magnitude be less than τ2Pixel be image background,
When in the non-maxima suppression image N (i, j) pixel between τ1And τ2Between, then check that the gradient magnitude of its adjacent pixels is
It is no to exist greater than τ1Pixel, if it exists then be image weak marginal point, if it does not exist then be image background, to be had
There is the image at the edge of registration.It is the prior art that accumulative histogram, which calculates dual threshold, is specifically repeated no more.
Preferably, after the step S324 further include:
S325: hysteresis bounds tracking is implemented as follows:
S3251: 8 connection area images of all weak marginal points are obtained；
S3252: 8 connection area images of the weak marginal point are carried out using breadth first algorithm or depth-priority-searching method
Search, if there are a pixels and strong edge point at the weak edge of a connection in 8 connection area images of the weak marginal point
Connection, then retain the weak edge of the connection, otherwise inhibit.
Since weak marginal point may be genuine edge, it is also possible to accurate to obtain caused by noise or color change
Edge needs to curb weak marginal point caused by noise or color change, be tracked by hysteresis bounds, can obtain edge more
It is accurate edge gray table as Y (i, j).Breadth first algorithm and depth-priority-searching method are the prior art, are specifically repeated no more.
Preferably, the characteristic value in the S4 includes the eccentricity of the target image and the area week of the target image
Long ratio, the presetting database are stored with the eccentricity comprising fault and impurity and the area of the corresponding fault and impurity week
The relation table of long ratio.
The relation table is obtained by many experiments, referring specifically to table 1:
The relation table of the eccentricity and area/perimeter of 1 fault of table and impurity
Eccentricity is mathematically to state ellipse to the departure degree with it with the positive round of reference point, geometric
Concept is just expressed as height fat or thin degree of the closed curved section on its major axes orientation by image, and iconology is arrived in reflection
In indicate the oblate state of body form, and do the resulting numerical value of division arithmetic using the area of target image and perimeter and exist
Certain range rule.It is learnt by many experiments, the eccentricity and the correlation of the ratio between area perimeter of different faults and impurity
There are certain rules, therefore the present embodiment establishes the ratio between eccentricity and area perimeter of fault and impurity one according to experimental result
One-to-one correspondence table, calculates the eccentricity and the ratio between area and perimeter of target image, and is compared with relation table, can be with
Obtain the recognition result of fault and impurity.Since cotton fault and dopant species are more, complex shape, different types of fault and miscellaneous
The feature of matter is different, and traditional identification method to fault and impurity needs different for different types of fault and impurity setting
Identification method, or identified using different threshold values, and the identification method of the present embodiment is without setting different identification methods
Or threshold value, different types of fault and impurity can be identified well, and detection is high with recognition accuracy, and speed is fast, practicability
It is high.
Etching operation is used to edge gray level image in the present embodiment S3, is to eliminate the edge in edge gray table picture
Burr, the boundary of smoothed image, and use expansive working, be in order to by edge gray table picture edge or inside hole
It fills out, smooth boundary, therefore using corrosion and expansive working, more smooth target image can be obtained, operating method is
The prior art specifically repeats no more.
Embodiment two, as shown in figure 4, for a kind of raw cotton fault of the present invention and defects inspecting and identifying system an implementation
The structural schematic diagram of example, including detection case and terminal；The detection case upper end is equipped with Image Acquisition subelement 1, and described image is adopted
Collection subelement 1 is used to acquire the color image of the sample cotton upper and lower surface after irradiation, is equipped with illumination unit 2 in the middle part of the detection case,
The transmission units illumination unit 2 is used to irradiate the upper surface of the sample cotton, and the detection case bottom is equipped with transmission units 3, institute
Transmission units 3 are stated for irradiating the lower surface of the sample cotton, the lower part of the detection case is equipped with sample stage 5, on the sample stage 4
Equipped with light-transmitting plate 5, sample cotton 6 is placed on the light-transmitting plate 5；
The terminal includes display 8 and processor 9, and the processor 9 is used to pre-process the color image,
Gray level image is obtained, the processor 9 is also used to carry out image procossing to the gray level image, obtains target image, the place
Reason device 9 is also used to extract the characteristic value of the target image, and the characteristic value of the target image and presetting database are carried out
It compares, obtains the recognition result of fault and impurity；The display 8 is used to show the recognition result of the fault and impurity；
The display 8 is electrically connected with the processor 9, described image acquire subelement 1 by data line 10 with it is described
Processor 9 is electrically connected, and the illumination unit 2 and the transmission units 3 are electrically connected with the processor 9.
Sample cotton is irradiated by the light that the illumination unit in detection case generates different wave length, and passes through detection case bottom
Transmission units as supplement light source, guarantee that all sites of sample cotton can be irradiated to, avoid occurring shadow part in image
Divide, then is irradiated to the color image on sample cotton surface by the Image Acquisition subelement acquisition of detection upper box part, it will by data line
Collected colo r image transmission after terminal carries out processing analysis to image, obtains fault corresponding to target image to terminal
Type, to realize the detection and identification to raw cotton fault and impurity, detection and recognition speed are fast, and Detection accuracy is high, can know
The drawbacks of not a plurality of types of raw cotton faults and impurity, elimination manual sorting, reduces detection and identification cost.
Preferably, described image acquisition subelement 1 is CCD camera, and the CCD camera is equipped with camera lens.
Small in size using CCD camera high sensitivity, the service life is long, anti-vibration excellent, the distortion of institute's acquired image
It is small, picture is clear.
Preferably, the illumination unit 2 is two groups of LED light, and two groups of LED light are separately positioned in the middle part of the detection case
Two sides.
The two sides of sample cotton oblique upper are arranged in two groups of LED light, be subject to its light source energy maximum magnitude covering sample cotton, by setting
Two groups of LED light in detection case on both sides of the middle are set, guarantee that LED light can be irradiated to all sites on sample cotton surface to the maximum extent,
So that acquired image is more comprehensively.
Preferably, the transmission units 3 are evenly distributed and are arranged described using the central axes of the light-transmitting plate 5 as benchmark line
Detection case bottom.
The transmission units being evenly distributed by detection case bottom irradiate the supplement of sample cotton, to cover sample to the maximum extent
Whole positions on cotton surface avoid dash area occur in the collected color image of institute.
Raw cotton fault and defects inspecting in the present embodiment and identifying system include power supply, described image acquire subelement 1,
The illumination unit 2, the transmission units 3, the display 8 and the processor 9 with the power electric connection.Pass through electricity
Source is Image Acquisition subelement, illumination unit, transmission units, display and processor continued power.
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all in spirit of the invention and
Within principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (10)
1. a kind of raw cotton fault and defects inspecting and recognition methods, which comprises the following steps:
Step 1: irradiation sample cotton upper and lower surface, and acquire the color image of the sample cotton upper and lower surface after irradiation；
Step 2: the color image being pre-processed, gray level image is obtained；
Step 3: image procossing being carried out to the gray level image, obtains target image；
Step 4: extracting the characteristic value of the target image, and the characteristic value of the target image and presetting database are compared
It is right, obtain the recognition result of fault and impurity and display.
2. a kind of raw cotton fault according to claim 1 and defects inspecting and recognition methods, which is characterized in that the step
Pretreated specific steps in 2 are as follows:
Step 21: using weighted average method to the color image carry out gray processing processing, extract the color image R,
G, tri- channel components of B, and tri- channel components of described R, G, B are weighted and averaged, obtain initial gray image Gray (i,
j)；
Step 22: using histogram enhancement processing method to initial gray image Gray (i, j) degree of the comparing enhancing at
Reason, obtains Continuous Tone Image Z (i, j)；
Step 23: noise removal process being carried out to the Continuous Tone Image Z (i, j) using median filtering, obtains the gray scale
Image.
3. a kind of raw cotton fault according to claim 2 and defects inspecting and recognition methods, which is characterized in that the step
22 specifically:
Step 221: a histogram equalization processing being carried out to the initial gray image Gray (i, j), and it is straight to draw gray scale
Fang Tu；
Step 222: the grey level histogram being modified using histogram specification method, obtains the Continuous Tone Image
Z(i,j)。
4. a kind of raw cotton fault according to claim 1 and defects inspecting and recognition methods, which is characterized in that the step
3 specific steps are as follows:
Step 31: image segmentation being carried out to the gray level image using adaptive threshold image segmentation, obtains segmentation grayscale image
As T (i, j)；
Step 32: edge detection being carried out to the segmentation gray level image T (i, j) using edge detection operator, obtains edge gray scale
Image Y (i, j)；
Step 33: the edge gray table being obtained described as Y (i, j) progress Morphological scale-space using corrosion and expansive working
Target image.
5. a kind of raw cotton fault according to claim 4 and defects inspecting and recognition methods, which is characterized in that the step
The adaptive threshold image segmentation based on Mean Shift algorithm is used in 31.
6. a kind of raw cotton fault according to claim 4 and defects inspecting and recognition methods, which is characterized in that the step
32 specific steps are as follows:
Step 321: noise remove being carried out to the gray level image T (i, j) after dividing processing using 2-d gaussian filters template
Processing, obtains smoothed image；
Step 322: calculating partial derivative of the gray scale in the direction x, y of the smoothed image using first-order difference local derviation, obtain described
The gradient magnitude and gradient direction of the gray scale of smoothed image；
Step 323: being accurately positioned using marginal point of the non-maxima suppression method to the smoothed image and refine edge, obtained
To non-maxima suppression image N (i, j)；
Step 324: detecting the edge of the non-maxima suppression image N (i, j) using dual threashold value-based algorithm, obtain the edge ash
It spends image Y (i, j).
7. a kind of raw cotton fault according to claim 1-6 and defects inspecting and recognition methods, which is characterized in that
Characteristic value in the step 4 includes the eccentricity of the target image and the area perimeter ratio of the target image, described pre-
If database purchase has the relationship of the eccentricity comprising fault and impurity with the corresponding fault and the area perimeter ratio of impurity
Table.
8. a kind of raw cotton fault and defects inspecting and identifying system, which is characterized in that including image acquisition units, image procossing list
Member, image extraction unit, recognition unit and display unit；Described image acquisition unit, for shining sample cotton upper and lower surface
It penetrates, and acquires the color image of the sample cotton upper and lower surface after irradiation；Described image processing unit is used for collected institute
It states color image to be pre-processed, obtains gray level image, be also used to the gray level image carrying out image procossing, obtain target figure
Picture；
The extraction unit, for extracting the characteristic value of the target image；
The recognition unit obtains fault and miscellaneous for the characteristic value of the target image to be compared with presetting database
The recognition result of matter；
The display unit, for showing the recognition result of the fault and impurity.
9. a kind of raw cotton fault according to claim 8 and defects inspecting and identifying system, which is characterized in that described image
Acquisition unit includes illumination unit, transmission units and Image Acquisition subelement；
The illumination unit is used to irradiate the upper surface of the sample cotton；
The transmission units are used to irradiate the lower surface of the sample cotton；
Described image acquisition subelement is used to acquire the institute of the upper surface of the sample cotton after irradiation and the lower surface of the sample cotton
State color image.
10. a kind of raw cotton fault according to claim 9 and defects inspecting and identifying system, which is characterized in that the figure
As processing unit includes gray processing unit, contrast enhancement processing unit, filter unit, image segmentation unit, edge detection list
Member and Morphological scale-space unit；
The gray processing unit obtains initial gray for carrying out gray processing to the color image using weighted average method
Image Gray (i, j)；
The contrast enhancement processing unit, for using histogram enhancement processing method to the initial gray image Gray
(i, j) degree of comparing enhancing processing, obtains Continuous Tone Image Z (i, j)；
The filter unit, for carrying out noise removal process to the Continuous Tone Image Z (i, j) using median filtering method,
Obtain the gray level image；
Described image cutting unit is divided for carrying out image segmentation to the gray level image using Mean Shift algorithm
Cut gray level image T (i, j)；
The edge detection unit, for carrying out edge inspection to the segmentation gray level image T (i, j) using edge detection operator
It surveys, obtains edge gray table as Y (i, j)；
The Morphological scale-space unit, for using corrosion and expansive working to the edge gray table as Y (i, j) is filled
Repairing, obtains the target image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810818716.0A CN108961251A (en) | 2018-07-24 | 2018-07-24 | A kind of raw cotton fault and defects inspecting and recognition methods and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810818716.0A CN108961251A (en) | 2018-07-24 | 2018-07-24 | A kind of raw cotton fault and defects inspecting and recognition methods and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108961251A true CN108961251A (en) | 2018-12-07 |
Family
ID=64464214
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810818716.0A Pending CN108961251A (en) | 2018-07-24 | 2018-07-24 | A kind of raw cotton fault and defects inspecting and recognition methods and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108961251A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109738436A (en) * | 2018-12-28 | 2019-05-10 | 安徽财经大学 | A kind of unginned cotton impurity quantifies rapid detection system and method |
| CN110009603A (en) * | 2019-03-14 | 2019-07-12 | 广州供电局有限公司 | High-voltage cable insulating detection method and high-tension cable maintaining method |
| CN111476795A (en) * | 2020-02-27 | 2020-07-31 | 浙江工业大学 | Binary icon notation method based on breadth-first search |
| CN111537519A (en) * | 2020-04-11 | 2020-08-14 | 嘉兴市昊鸣纺织有限公司 | Automatic detection and supervision system for cloth |
| CN111665199A (en) * | 2019-03-06 | 2020-09-15 | 东莞中科蓝海智能视觉科技有限公司 | Wire and cable color detection and identification method based on machine vision |
| CN112586539A (en) * | 2020-11-08 | 2021-04-02 | 泰州市华仕达机械制造有限公司 | Self-adaptive belt return control platform |
| CN113182214A (en) * | 2021-04-21 | 2021-07-30 | 江南大学 | Sorting device for foreign fibers of raw cotton and using method |
| CN113261428A (en) * | 2021-06-18 | 2021-08-17 | 湖北省农业科学院经济作物研究所 | Whole-process quality control method for machine-harvested cotton |
| CN114638827A (en) * | 2022-05-18 | 2022-06-17 | 卡松科技股份有限公司 | Visual detection method and device for impurities of lubricating oil machinery |
| CN117011550A (en) * | 2023-10-08 | 2023-11-07 | 超创数能科技有限公司 | Impurity identification method and device in electron microscope photo |
| CN117522281A (en) * | 2024-01-05 | 2024-02-06 | 山东通广电子股份有限公司 | Tool and instrument warehouse-in and warehouse-out management method and system based on visual identification |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0491954A4 (en) * | 1990-05-28 | 1993-09-08 | Kanebo Ltd. | Method of inspecting cleanliness of top and device used therefor |
| CN101587079A (en) * | 2009-05-08 | 2009-11-25 | 清华大学 | Multispectral chromatic polarization detection method of foreign impurity in the textile raw material |
| CN103698341A (en) * | 2013-12-31 | 2014-04-02 | 中华全国供销合作总社郑州棉麻工程技术设计研究所 | System for detecting cotton rolling quality and method for detecting cotton rolling quality based on image |
| CN104751443A (en) * | 2014-12-12 | 2015-07-01 | 郑州轻工业学院 | Cotton fault detecting and identifying method based on multi-spectrum technology |
-
2018
- 2018-07-24 CN CN201810818716.0A patent/CN108961251A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0491954A4 (en) * | 1990-05-28 | 1993-09-08 | Kanebo Ltd. | Method of inspecting cleanliness of top and device used therefor |
| CN101587079A (en) * | 2009-05-08 | 2009-11-25 | 清华大学 | Multispectral chromatic polarization detection method of foreign impurity in the textile raw material |
| CN103698341A (en) * | 2013-12-31 | 2014-04-02 | 中华全国供销合作总社郑州棉麻工程技术设计研究所 | System for detecting cotton rolling quality and method for detecting cotton rolling quality based on image |
| CN104751443A (en) * | 2014-12-12 | 2015-07-01 | 郑州轻工业学院 | Cotton fault detecting and identifying method based on multi-spectrum technology |
Non-Patent Citations (2)
| Title |
|---|
| 赵小川: "《MATLAB图像处理能力提高与应用案例》", 31 January 2014, 北京航空航天大学出版 * |
| 赵西娜: "基于图像处理技术的原棉疵点及杂质特征识别的研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅰ辑》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109738436A (en) * | 2018-12-28 | 2019-05-10 | 安徽财经大学 | A kind of unginned cotton impurity quantifies rapid detection system and method |
| CN111665199A (en) * | 2019-03-06 | 2020-09-15 | 东莞中科蓝海智能视觉科技有限公司 | Wire and cable color detection and identification method based on machine vision |
| CN110009603A (en) * | 2019-03-14 | 2019-07-12 | 广州供电局有限公司 | High-voltage cable insulating detection method and high-tension cable maintaining method |
| CN111476795A (en) * | 2020-02-27 | 2020-07-31 | 浙江工业大学 | Binary icon notation method based on breadth-first search |
| CN111537519B (en) * | 2020-04-11 | 2023-02-28 | 嘉兴市昊鸣纺织有限公司 | Automatic detection and supervision system for cloth |
| CN111537519A (en) * | 2020-04-11 | 2020-08-14 | 嘉兴市昊鸣纺织有限公司 | Automatic detection and supervision system for cloth |
| CN112586539A (en) * | 2020-11-08 | 2021-04-02 | 泰州市华仕达机械制造有限公司 | Self-adaptive belt return control platform |
| CN113182214A (en) * | 2021-04-21 | 2021-07-30 | 江南大学 | Sorting device for foreign fibers of raw cotton and using method |
| CN113261428B (en) * | 2021-06-18 | 2022-04-12 | 湖北省农业科学院经济作物研究所 | Whole-process quality control method for machine-harvested cotton |
| CN113261428A (en) * | 2021-06-18 | 2021-08-17 | 湖北省农业科学院经济作物研究所 | Whole-process quality control method for machine-harvested cotton |
| CN114638827A (en) * | 2022-05-18 | 2022-06-17 | 卡松科技股份有限公司 | Visual detection method and device for impurities of lubricating oil machinery |
| CN117011550A (en) * | 2023-10-08 | 2023-11-07 | 超创数能科技有限公司 | Impurity identification method and device in electron microscope photo |
| CN117011550B (en) * | 2023-10-08 | 2024-01-30 | 超创数能科技有限公司 | Impurity identification method and device in electron microscope photo |
| CN117522281A (en) * | 2024-01-05 | 2024-02-06 | 山东通广电子股份有限公司 | Tool and instrument warehouse-in and warehouse-out management method and system based on visual identification |
| CN117522281B (en) * | 2024-01-05 | 2024-04-16 | 山东通广电子股份有限公司 | Tool and instrument warehouse-in and warehouse-out management method and system based on visual identification |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108961251A (en) | A kind of raw cotton fault and defects inspecting and recognition methods and system | |
| CN109934802B (en) | Cloth defect detection method based on Fourier transform and image morphology | |
| CN108765402B (en) | Non-woven fabric defect detection and classification method | |
| CN104597057B (en) | A kind of column Diode facets defect detecting device based on machine vision | |
| CN107610104A (en) | Crack detecting method at a kind of elevator compensation chain R based on machine vision | |
| CN108320280A (en) | The crater image method for detecting abnormality of view-based access control model clarity and contours extract | |
| CN109685766A (en) | A kind of Fabric Defect detection method based on region fusion feature | |
| CN109668909A (en) | A kind of glass defect detection method | |
| CN113252568A (en) | Lens surface defect detection method, system, product and terminal based on machine vision | |
| Dixit et al. | Image texture analysis-survey | |
| CN109345524A (en) | A kind of bearing open defect detection system of view-based access control model | |
| CN106780464A (en) | A kind of fabric defect detection method based on improvement Threshold segmentation | |
| CN103473776A (en) | Method and system for comparing image definition and automatic focusing control method | |
| CN105426843A (en) | Single-lens palm vein and palmprint image acquisition apparatus and image enhancement and segmentation method | |
| Feng et al. | A separating method of adjacent apples based on machine vision and chain code information | |
| CN102860814A (en) | OCT (Optical Coherence Tomography) synthetic fundus image optic disc center positioning method and equipment | |
| CN110335233A (en) | Express-way guard-rail plates defect detecting system and method based on image processing techniques | |
| CN104123731A (en) | Method for dividing low-contrast-ratio ginned cotton mulching film image | |
| CN109949294A (en) | A kind of fracture apperance figure crack defect extracting method based on OpenCV | |
| CN115100104A (en) | Defect detection method, device and equipment for glass ink area and readable storage medium | |
| CN106093051A (en) | Paper roll tangent plane burr detection method based on machine vision and device | |
| WO2018046407A1 (en) | Particle boundary identification | |
| CN110648312A (en) | Method for identifying wool and cashmere fibers based on scale morphological characteristic analysis | |
| CN109870730A (en) | A kind of method and system for X-ray machine image analytic degree test body regular inspection | |
| CN109387524A (en) | Thread defect detection method and device based on linearly polarized photon |
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 |