CN107705300A - A kind of method that blank page detection is realized based on morphological transformation - Google Patents

A kind of method that blank page detection is realized based on morphological transformation Download PDF

Info

Publication number
CN107705300A
CN107705300A CN201710901117.0A CN201710901117A CN107705300A CN 107705300 A CN107705300 A CN 107705300A CN 201710901117 A CN201710901117 A CN 201710901117A CN 107705300 A CN107705300 A CN 107705300A
Authority
CN
China
Prior art keywords
mrow
image
msub
structural elements
gray
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
Application number
CN201710901117.0A
Other languages
Chinese (zh)
Inventor
赖南英
陶冰洁
严棚
王酉祥
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chengdu Great Bear Intelligent Technology Co Ltd
Original Assignee
Chengdu Great Bear Intelligent Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Chengdu Great Bear Intelligent Technology Co Ltd filed Critical Chengdu Great Bear Intelligent Technology Co Ltd
Priority to CN201710901117.0A priority Critical patent/CN107705300A/en
Publication of CN107705300A publication Critical patent/CN107705300A/en
Pending legal-status Critical Current

Links

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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/20—Image enhancement or restoration by the use of local operators
    • G06T5/30—Erosion or dilatation, e.g. thinning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/40—Analysis of texture
    • G06T7/41—Analysis of texture based on statistical description of texture
    • G06T7/44—Analysis of texture based on statistical description of texture using image operators, e.g. filters, edge density metrics or local histograms
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
    • 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
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image
    • G06T2207/10008—Still image; Photographic image from scanner, fax or copier
    • 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
    • G06T2207/20036—Morphological image processing

Abstract

The invention discloses a kind of method that blank page detection is realized based on morphological transformation.The Intelligent Recognition field belonged in machine vision commercial Application.Emphasis of the present invention is that the quick and accurate automatic detection to blank page is realized based on morphologic method.General principle converts for artwork to be carried out to the top cap in morphology first, get rid of the background parts in image, mean filter is carried out with the structural elements of font size again, the ratio of the point of a certain threshold value is exceeded according to gray scale in last image, determines whether current page is blank page.This method will focus on detection speed, take that some principles are simple but fast and effectively method.Test result indicates that this method not only efficient quick, and can ensure preferable accuracy of identification, has practical value.

Description

A kind of method that blank page detection is realized based on morphological transformation
Technical field
The present invention relates to the Intelligent Recognition field in machine vision commercial Application, more particularly to one kind to be based on morphological transformation The method for realizing blank page detection, for automatic detection blank page in non-contact scanning.
Background technology
With the popularization of computer and network technologies, electronic edition document by its it is easily stored utilize, propagate it is fast and simple Advantage received by increasing people.Therefore, the data originally using paper as carrier is converted into electronic edition document, The namely digitlization of paper document either in routine office work, or library management in application all increasingly Increase.
Non-contact scanning has obtained increasing application as a kind of new paper document digital form, it Certain room for promotion also be present at present.Non-contact scanning is scanned by the way of page turning, while scan text is left It is right page two, but some texts are one sides, therefore left side page may be blank page, while scan the left and right meetings of page two and cause wave Take.Can solve this problem by way of detecting page and whether being blank page.The present invention method be exactly tackle it is this Demand, by the current page of rapidly automatic detection, determine whether blank page, be automatically regulated to be single-page scan or pair of pages Scanning, the workload of people is reduced, further lift the practicality of non-contact scanning.Test result indicates that this method is not only high Effect is quick, and can ensure preferable accuracy of identification, has practical value.
The content of the invention
It is an object of the invention to：The present invention provides one kind for the weak point of non-contact scanning and is based on morphology The method that blank page detection is realized in conversion, for solving single pair of pages switching problem in non-contact scanning, is further lifted non- The practicality of contact type scanning.
The technical solution adopted by the present invention is as follows：
A kind of method that blank page detection is realized based on morphological transformation, it is characterised in that comprise the following steps：
Step 1：Initial pictures f (x, y) is read in, is converted into gray level image fgray(x, y), initial pictures size are M × N；
Step 2：Structural elements a is taken to gray level image fgray(x, y) carries out expansive working, the image f after being expandedswell (x, y)；
Step 3：Take the image f after structural elements b corrosion expansionsswell(x, y), the image f after being corrodedetch(x, y), then With the gray level image f in step 1gray(x, y) subtracts the image f after being expanded in step 2swell(x, y), after obtaining top cap conversion Image ftophat(x, y)；
Step 4：With with font size identical structural elements c and image ftophat(x, y) does convolution, obtains the figure after convolution As fconv(x, y)；
Step 5：By the image f after convolutionconv(x, y) binaryzation, obtain binary image fbinary(x, y)；
Step 6：The binary image f arrived of statistic procedure 5binaryBright spot accounts for the ratio of whole binary image in (x, y) Example, if the shared bright spot threshold value for being more than setting of bright spot, then it is assumed that be not blank page, conversely, being then considered blank page.
Further, the step 1 comprises the following steps that：
Step 11：Read initial pictures f (x, y)；
Step 12：Initial pictures f (x, y) is converted into gray level image fgray(x, y), wherein gray level image fgray(x's, y) The formula of the gray value of each pixel is as follows：
Wherein Rf(x, y) represents the pixel value of each pixel of initial pictures f ' (x, y) red channel；Gf(x, y) is represented just The pixel value of each pixel of beginning image f (x, y) green channel；Bf(x, y) represents that initial pictures f (x, y) blue channel is each The pixel value of pixel, (x, y) represent each pixel of initial pictures.
Further, the step 2 comprises the following steps that：
Step 21：The structural elements a of the size of font 1/3rd is taken, structural elements a sizes are m × n；
Step 22：With the inswept gray level image f of structural elements agray(x, y) each element, the image f after being expandedswell (x, y)；
Step 221：When sweeping to pixel (x, y), the value after point expansion is：
F (x+s, y+t) belongs to, when structural elements a center is at (x, y), the region where structural elements a；S, t are to make f (x + s, y+t) belong to current structure member a covering region constant.
Further, the step 3 comprises the following steps that：
Step 31：It is m × n to take with an equal amount of structural elements b of A structural elements a, structural elements b sizes；
Step 32：With the image f after the inswept expansions of structural elements bswell(x, y) each element, the image after being corroded fetch(x, y)；
Step 321：When sweeping to pixel (x, y), the value calculation formula after the spot corrosion is：
F (x+u, y+v) belongs to, when structural elements b center is at (x, y), the whole region where structural elements b；U, v are to make F (x+u, y+v) belongs to the constant in the region of current structure member b coverings.
Further, the step 4 comprises the following steps that：
Step 41：It is p × q to take with font size identical structural elements c, structural elements c size；
Step 42：With structural elements c and the image f after corrosionetch(x, y) carries out convolution, and concrete mode is as follows,
Step 421：From the image f after corrosionetch(x, y) upper left side begins stepping through image, mobile q/2 or p/2 every time Pixel, it is p/2 in short transverse, is q/2 on width；First move right, step-length q/2, terminate a line after still further below Mobile p/2, since high order end, convolution results areImage f after the image block of size, as convolutionconv(x, y)；
Step 422：For the image f after convolutionconvPixel in (x, y), its value calculation formula are
f(x0, y0) belong to structural elements c center at (x, y), the whole region where structural elements c.
Further, the step 5 comprises the following steps that：
Step 51：Calculate gray level image fconvThe average gray of (x, y)
Step 52：For gray value t, f is traveled throughconvThe all pixels point of (x, y), will be divided into two parts a little, respectively It is more than t pixel point set B for pixel point set A of the gray value less than or equal to t and gray value；
Step 53：The point in A and B is calculated respectively account for the number of all pixels point account for the number of all pixels point in A and B Ratio, it is designated as PAAnd PB, then the average gray value of pixel in A and B is calculated respectively, it is designated asWith
Step 54：Inter-class variance is calculated, inter-class variance calculation formula is：
Make t=1 successively, 2,3 ... 255, obtain all ICVt, compare as a result, working as
Using the gray value corresponding to inter-class variance maximum in all inter-class variances as binaryzation transform key t0；
Step 55：According to binaryzation transform key t0, by gray level image fgray(x, y) is converted into Binary Sketch of Grey Scale Image figure As fbinary(x, y).
Further, the step 6 comprises the following steps that：
Step 61:Count binary image fbinaryThe quantity of bright spot in (x, y), it is designated as h, bright spot proportion
Step 62：It is 2% to set bright spot threshold value, if
α ＞ 2%
Then image f (x, y) is not blank image；
If
α≤2%
Then image f (x, y) is blank image.In summary, by adopting the above-described technical solution, the beneficial effect of the present invention Fruit is：
1. in this programme, top cap conversion is carried out to image first, removes the background image in image, then according to by top cap Image after conversion carries out binaryzation, and so as to accelerate the speed of image judgement, precise and high efficiency judges whether the page is empty White page.
2. in this programme, due to being filtered out using top cap conversion to the background of image, therefore, it is possible to reduce Background The influence that non-legible point as in is judged blank page, so as to improve the accuracy of blank page judgement.
3.3. due to this programme can automatic decision current page whether non-NULL white page, accordingly, it is capable to sentencing according to this programme Break in non-contact scanning, adjust automatically uses single-page scan pattern or pair of pages scan pattern, and the workload for reducing people enters one The practicality of step lifting non-contact scanning.
Brief description of the drawings
Examples of the present invention will be described by way of reference to the accompanying drawings, wherein：
Fig. 1 is the schematic flow sheet of the present invention；
Fig. 2 is the initial pictures f (x, y) of text to be detected；
Fig. 3 is gray level image fgray(x,y)；
Fig. 4 is the image f after expansion processswell(x,y)；
Fig. 5 is the image f after processing corrosion treatmentetch(x,y)；
Fig. 6 is the image f after top cap convertstophat(x,y)；
Fig. 7 is the image f after convolutionconv(x,y)；
Fig. 8 is the image f after binaryzationbinary(x,y)。
Embodiment
All features disclosed in this specification, or disclosed all methods or during the step of, except mutually exclusive Feature and/or step beyond, can combine in any way.
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples The present invention is further elaborated.It should be appreciated that specific embodiment described herein is only to explain the present invention, and It is not used in the restriction present invention.
A kind of method that blank page detection is realized based on morphological transformation, is comprised the following steps：
Step 1：Initial pictures f (x, y) is read in, is converted into gray level image fgray(x, y), initial pictures size are M × N. Comprise the following steps that：
Step 11：Read initial pictures f (x, y).
Step 12：Initial pictures f (x, y) is converted into gray level image fgray(x, y), wherein gray level image fgray(x's, y) The formula of the gray value of each pixel is as follows：
Wherein Rf(x, y) represents the pixel value of each pixel of initial pictures f ' (x, y) red channel；Gf(x, y) is represented just The pixel value of each pixel of beginning image f (x, y) green channel；Bf(x, y) represents that initial pictures f (x, y) blue channel is each The pixel value of pixel, (x, y) represent each pixel of initial pictures.
Step 2：The structural elements a of the size of font 1/3rd or so is taken to expand gray level image fgray(x, y), after obtaining expansion Image fswell(x, y).
Comprise the following steps that：
Step 21：The structural elements a of the size of font 1/3rd is taken, structural elements a sizes are to be set to m × n.
Step 22：With the inswept gray level image f of structural elements agray(x, y) each element, the image f after being expandedswell (x, y).
Step 221：When sweeping to pixel (x, y), the value after point expansion is：
F (x+s, y+t) belongs to, when structural elements a center is at (x, y), the region where structural elements a；S, t are to make f (x + s, y+t) belong to current structure member a covering region constant.Step 3：Take the structural elements b of the size of font 1/3rd or so Image f after corrosion expansionswell(x, y), the image f after being corrodedetch(x, y), then with gray level image fgray(x, y) subtracts The image f gone after expansionswell(x, y) obtains the image f after top cap conversiontophat(x, y)；Comprise the following steps that：
Step 31：Take with an equal amount of structural elements b of structural elements a, be m × n.
Step 32：With the image f after the inswept expansions of structural elements bswell(x, y) each element, the image after being corroded fetch(x, y).
Step 321：When sweeping to pixel (x, y), the value after the spot corrosion is：
F (x+u, y+v) belongs to, when structural elements b center is at (x, y), the whole region where structural elements b；U, v are to make F (x+u, y+v) belongs to the constant in the region of current structure member b coverings.Step 4：With with font size identical structural elements c with Image ftophat(x, y) does convolution, obtains the image f after convolutionconv(x, y)；Comprise the following steps that：
Step 41：The structural elements c same or like with font size is taken, if structural elements c sizes are p × q.
Step 42：With structural elements c and the image f after corrosionetch(x, y) carries out convolution, and concrete mode is as follows
Step 421：From the image f after corrosionetch(x, y) upper left side begins stepping through image, mobile q/2 or p/2 every time Pixel (being p/2 in short transverse, be q/2 on width).First move right, step-length q/2, terminate a line after again to Lower mobile p/2, since high order end, convolution results areImage f after the image block of size, as convolutionconv(x, y)。
Step 422：For the image f after convolutionconvPixel in (x, y), its value are
f(x0, y0) belong to structural elements c center at (x, y), the whole region where structural elements c.
Step 5：By the image f after convolutionconv(x, y) binaryzation, obtain binary image fbinary(x, y)；Specific steps It is as follows：
Step 51：Calculate gray level image fconvThe average gray of (x, y), note average gray are
Step 52：For gray value t (0≤t≤255), f is traveled throughconvThe all pixels point of (x, y), will be divided into a little The pixel point set A of two parts, respectively gray value less than or equal to t and gray value are more than t pixel point set B.
Step 53：The point in A and B is calculated respectively account for the number of all pixels point account for the number of all pixels point in A and B Ratio, it is designated as PAAnd PB, then the average gray value of pixel in A and B is calculated respectively, it is designated asWith
Step 54：Calculate inter-class variance ICVt, inter-class variance ICVtCalculation formula be
Make t=1 successively, 2,3 ... 255, obtain all inter-class variance ICVt, compare as a result, working as
Using the gray value corresponding to inter-class variance maximum in all inter-class variances as binaryzation transform key t0；
Step 55：According to binaryzation transform key t0, by gray level image fgray(x, y) is converted into Binary Sketch of Grey Scale Image figure As fbinary(x, y)；
Step 6：Investigate fbinaryRatio in (x, y) shared by bright spot, if more than 2%, then it is assumed that it is not blank page, conversely, Then it is considered blank page；Comprise the following steps that：
Step 61:Investigate binary image fbinaryThe quantity of bright spot in (x, y), it is designated as h, bright spot proportion
Step 62：Bright spot threshold value is set, if α is more than the bright spot threshold value set, it is not blank sheet to judge image f (x, y) Picture.Otherwise, it is blank page to judge image f (x, y).
Such as set bright spot threshold value be 2% when, if
α ＞ 2%
It is not blank image then to think image f (x, y).
If
α≤2%
It is blank image then to think image f (x, y).

Claims (7)

  1. A kind of 1. method that blank page detection is realized based on morphological transformation, it is characterised in that comprise the following steps：
    Step 1：Initial pictures f (x, y) is read in, is converted into gray level image fgray(x, y), initial pictures size are M × N；
    Step 2：Structural elements a is taken to gray level image fgray(x, y) carries out expansive working, the image f after being expandedswell(x, y)；
    Step 3：Take the image f after structural elements b corrosion expansionsswell(x, y), the image f after being corrodedetch(x, y), then with step Gray level image f in rapid 1gray(x, y) subtracts the image f after being expanded in step 2swell(x, y), obtain the image after top cap conversion ftophat(x, y)；
    Step 4：With with font size identical structural elements c and image ftophat(x, y) does convolution, obtains the image f after convolutionconv (x, y)；
    Step 5：By the image f after convolutionconv(x, y) binaryzation, obtain binary image fbinary(x, y)；
    Step 6：The binary image f arrived of statistic procedure 5binaryBright spot accounts for the ratio of whole binary image in (x, y), if The shared bright spot threshold value for being more than setting of bright spot, then it is assumed that be not blank page, otherwise, then it is assumed that be blank page.
  2. A kind of 2. method that blank page detection is realized based on morphological transformation according to claim 1, it is characterised in that institute State comprising the following steps that for step 1：
    Step 11：Read initial pictures f (x, y)；
    Step 12：Initial pictures f (x, y) is converted into gray level image fgray(x, y), wherein gray level image fgray(x's, y) is each The formula of the gray value of pixel is as follows：
    <mrow> <msub> <mi>f</mi> <mrow> <mi>g</mi> <mi>r</mi> <mi>a</mi> <mi>y</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mrow> <msub> <mi>R</mi> <mi>f</mi> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>G</mi> <mi>f</mi> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>B</mi> <mi>f</mi> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> </mrow> <mn>3</mn> </mfrac> </mrow>
    Wherein Rf(x, y) represents the pixel value of each pixel of initial pictures f ' (x, y) red channel；Gf(x, y) represents initial graph As the pixel value of each pixel of f (x, y) green channel；Bf(x, y) represents each pixel of initial pictures f (x, y) blue channel The pixel value of point, (x, y) represent each pixel of initial pictures.
  3. A kind of 3. method that blank page detection is realized based on morphological transformation according to claim 1, it is characterised in that institute State comprising the following steps that for step 2：
    Step 21：The structural elements a of the size of font 1/3rd is taken, structural elements a sizes are m × n；
    Step 22：With the inswept gray level image f of structural elements agray(x, y) each element, the image f after being expandedswell(x, y)；
    Step 221：When sweeping to pixel (x, y), the value after point expansion is：
    <mrow> <msub> <mi>f</mi> <mrow> <mi>s</mi> <mi>w</mi> <mi>e</mi> <mi>l</mi> <mi>l</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> <mo>=</mo> <munder> <mrow> <mi>m</mi> <mi>a</mi> <mi>x</mi> </mrow> <mrow> <mo>(</mo> <mi>s</mi> <mo>,</mo> <mi>t</mi> <mo>)</mo> <mo>&amp;Element;</mo> <mi>a</mi> </mrow> </munder> <mo>{</mo> <mi>f</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>+</mo> <mi>s</mi> <mo>,</mo> <mi>y</mi> <mo>+</mo> <mi>t</mi> <mo>)</mo> </mrow> <mo>}</mo> </mrow>
    F (x+s, y+t) belongs to, when structural elements a center is at (x, y), the region where structural elements a；S, t are to make f (x+s, y+ T) constant in the region of current structure member a coverings is belonged to.
  4. A kind of 4. method that blank page detection is realized based on morphological transformation according to claim 1, it is characterised in that institute State comprising the following steps that for step 3：
    Step 31：It is m × n to take with an equal amount of structural elements b of structural elements a, structural elements b sizes；
    Step 32：With the inswept image f of structural elements bswell(x, y) each element, the image f after being corrodedetch(x, y)；
    Step 321：When sweeping to pixel (x, y), the value calculation formula after the spot corrosion is：
    <mrow> <msub> <mi>f</mi> <mrow> <mi>c</mi> <mi>o</mi> <mi>n</mi> <mi>v</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> <mo>=</mo> <munder> <mrow> <mi>m</mi> <mi>i</mi> <mi>n</mi> </mrow> <mrow> <mo>(</mo> <mi>u</mi> <mo>,</mo> <mi>v</mi> <mo>)</mo> <mo>&amp;Element;</mo> <mi>b</mi> </mrow> </munder> <mo>{</mo> <mi>f</mi> <mrow> <mo>(</mo> <mi>x</mi> <mo>+</mo> <mi>u</mi> <mo>,</mo> <mi>y</mi> <mo>+</mo> <mi>v</mi> <mo>)</mo> </mrow> <mo>}</mo> </mrow>
    F (x+u, y+v) belongs to, when structural elements b center is at (x, y), the whole region where structural elements b；U, v are to make f (x+ U, y+v) belong to current structure member b covering region constant.
  5. A kind of 5. method that blank page detection is realized based on morphological transformation according to claim 1, it is characterised in that institute State comprising the following steps that for step 4：
    Step 41：It is p × q to take with font size identical structural elements c, structural elements c size；
    Step 42：With structural elements c and the image f after corrosionetch(x, y) carries out convolution, and concrete mode is as follows,
    Step 421：From the image f after corrosionetch(x, y) upper left side begins stepping through image, every time mobile q/2 or p/2 pixel Point, it is p/2 in short transverse, is q/2 on width；First move right, step-length q/2, moved still further below after terminating a line P/2, since high order end, convolution results areImage f after the image block of size, as convolutionconv(x, y)；
    Step 422：For the image f after convolutionconvPixel in (x, y), its value calculation formula are：
    <mrow> <msub> <mi>f</mi> <mrow> <mi>c</mi> <mi>o</mi> <mi>n</mi> <mi>v</mi> </mrow> </msub> <mrow> <mo>(</mo> <mi>x</mi> <mo>,</mo> <mi>y</mi> <mo>)</mo> </mrow> <mo>=</mo> <mfrac> <mn>1</mn> <mrow> <mi>p</mi> <mi>q</mi> </mrow> </mfrac> <munder> <mo>&amp;Sigma;</mo> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>y</mi> <mn>0</mn> </msub> <mo>)</mo> <mo>&amp;Element;</mo> <mi>c</mi> </mrow> </munder> <mi>f</mi> <mrow> <mo>(</mo> <msub> <mi>x</mi> <mn>0</mn> </msub> <mo>,</mo> <msub> <mi>y</mi> <mn>0</mn> </msub> <mo>)</mo> </mrow> </mrow>
    f(x0, y0) belong to structural elements c center at (x, y), the whole region where structural elements c.
  6. A kind of 6. method that blank page detection is realized based on morphological transformation according to claim 1, it is characterised in that institute State comprising the following steps that for step 5：
    Step 51：Calculate gray level image fconvThe average gray of (x, y)
    Step 52：For gray value t, f is traveled throughconvThe all pixels point of (x, y), will be divided into two parts, respectively gray scale a little Pixel point set A of the value less than or equal to t and gray value are more than t pixel point set B；
    Step 53：The point in A and B is calculated respectively account for the number of all pixels point account for the ratio of the number of all pixels point in A and B Example, is designated as PAAnd PB, then the average gray value of pixel in A and B is calculated respectively, it is designated asWith
    Step 54：Inter-class variance is calculated, inter-class variance calculation formula is：
    <mrow> <msub> <mi>ICV</mi> <mi>t</mi> </msub> <mo>=</mo> <msub> <mi>P</mi> <mi>A</mi> </msub> <mo>&amp;times;</mo> <msup> <mrow> <mo>(</mo> <mover> <msub> <mi>X</mi> <mi>A</mi> </msub> <mo>&amp;OverBar;</mo> </mover> <mo>-</mo> <mover> <mi>X</mi> <mo>&amp;OverBar;</mo> </mover> <mo>)</mo> </mrow> <mn>2</mn> </msup> <mo>+</mo> <msub> <mi>P</mi> <mi>B</mi> </msub> <mo>&amp;times;</mo> <msup> <mrow> <mo>(</mo> <msub> <mi>X</mi> <mi>B</mi> </msub> <mo>-</mo> <mover> <mi>X</mi> <mo>&amp;OverBar;</mo> </mover> <mo>)</mo> </mrow> <mn>2</mn> </msup> </mrow>
    Make t=1 successively, 2,3 ... 255, obtain all ICVt, compare as a result, working as
    <mrow> <msub> <mi>ICV</mi> <msub> <mi>t</mi> <mn>0</mn> </msub> </msub> <mo>=</mo> <mi>m</mi> <mi>a</mi> <mi>x</mi> <mrow> <mo>(</mo> <msub> <mi>ICV</mi> <mi>t</mi> </msub> <mo>)</mo> </mrow> </mrow>
    Using the gray value corresponding to inter-class variance maximum in all inter-class variances as binaryzation transform key t0；
    Step 55：According to binaryzation transform key t0, by gray level image fgray(x, y) is converted into Binary Sketch of Grey Scale Image image fbinary(x, y).
  7. A kind of 7. method that blank page detection is realized based on morphological transformation according to claim 1, it is characterised in that institute State comprising the following steps that for step 6：
    Step 61:Count binary image fbinaryThe quantity of bright spot in (x, y), it is designated as h, bright spot proportion
    <mrow> <mi>&amp;alpha;</mi> <mo>=</mo> <mfrac> <mi>h</mi> <mrow> <mfrac> <mrow> <mn>2</mn> <mi>M</mi> </mrow> <mi>p</mi> </mfrac> <mo>&amp;times;</mo> <mfrac> <mrow> <mn>2</mn> <mi>N</mi> </mrow> <mi>q</mi> </mfrac> </mrow> </mfrac> <mo>=</mo> <mfrac> <mrow> <mi>h</mi> <mi>p</mi> <mi>q</mi> </mrow> <mrow> <mn>4</mn> <mi>M</mi> <mi>N</mi> </mrow> </mfrac> </mrow>
    Step 62：It is 2% to set bright spot threshold value, if
    α ＞ 2%
    Then image f (x, y) is not blank image；
    If
    α≤2%
    Then image f (x, y) is blank image.
CN201710901117.0A 2017-09-28 2017-09-28 A kind of method that blank page detection is realized based on morphological transformation Pending CN107705300A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710901117.0A CN107705300A (en) 2017-09-28 2017-09-28 A kind of method that blank page detection is realized based on morphological transformation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710901117.0A CN107705300A (en) 2017-09-28 2017-09-28 A kind of method that blank page detection is realized based on morphological transformation

Publications (1)

Publication Number Publication Date
CN107705300A true CN107705300A (en) 2018-02-16

Family

ID=61175260

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710901117.0A Pending CN107705300A (en) 2017-09-28 2017-09-28 A kind of method that blank page detection is realized based on morphological transformation

Country Status (1)

Country Link
CN (1) CN107705300A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108765335A (en) * 2018-05-25 2018-11-06 电子科技大学 A kind of forest fire detection method based on remote sensing images
CN109543554A (en) * 2018-10-30 2019-03-29 深圳怡化电脑股份有限公司 Bill detection method, device, terminal and computer readable storage medium

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030235451A1 (en) * 2002-06-24 2003-12-25 Brown Mark Louis Apparatus and method for suppressing the printing of nearly-blank pages
CN1556503A (en) * 2004-01-09 2004-12-22 清华大学 Fast morphology corrusion expansion method in image treatment
CN102612833A (en) * 2009-09-10 2012-07-25 萨热姆文献简易股份有限公司 Method for detecting blank pages, and corresponding multifunctional machine
CN103077388A (en) * 2012-10-31 2013-05-01 浙江大学 Rapid text scanning method oriented to portable computing equipment
CN106558033A (en) * 2016-10-19 2017-04-05 安徽中科华澄智能科技股份有限公司 Binary Images Processing method based on morphological dilations erosion algorithm
CN107133929A (en) * 2017-04-27 2017-09-05 湖北工业大学 Low quality file and picture binary coding method based on background estimating and energy minimization

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030235451A1 (en) * 2002-06-24 2003-12-25 Brown Mark Louis Apparatus and method for suppressing the printing of nearly-blank pages
CN1556503A (en) * 2004-01-09 2004-12-22 清华大学 Fast morphology corrusion expansion method in image treatment
CN102612833A (en) * 2009-09-10 2012-07-25 萨热姆文献简易股份有限公司 Method for detecting blank pages, and corresponding multifunctional machine
CN103077388A (en) * 2012-10-31 2013-05-01 浙江大学 Rapid text scanning method oriented to portable computing equipment
CN106558033A (en) * 2016-10-19 2017-04-05 安徽中科华澄智能科技股份有限公司 Binary Images Processing method based on morphological dilations erosion algorithm
CN107133929A (en) * 2017-04-27 2017-09-05 湖北工业大学 Low quality file and picture binary coding method based on background estimating and energy minimization

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
殷苏民等: "基于顶帽变换和最大类间方差法的图像分割方法研究", 《科学技术与工程》 *
陆玲等: "《数字图像处理》", 31 July 2007 *
陈爱军等: "《数字图像处理及其MATLAB实现》", 31 July 2008 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108765335A (en) * 2018-05-25 2018-11-06 电子科技大学 A kind of forest fire detection method based on remote sensing images
CN108765335B (en) * 2018-05-25 2022-08-02 电子科技大学 Forest fire detection method based on remote sensing image
CN109543554A (en) * 2018-10-30 2019-03-29 深圳怡化电脑股份有限公司 Bill detection method, device, terminal and computer readable storage medium
CN109543554B (en) * 2018-10-30 2020-11-13 深圳怡化电脑股份有限公司 Bill detection method, device, terminal and computer readable storage medium

Similar Documents

Publication Publication Date Title
CN104732227B (en) A kind of Location Method of Vehicle License Plate based on definition and luminance evaluation
CN101789080B (en) Detection method for vehicle license plate real-time positioning character segmentation
CN106096610B (en) A kind of file and picture binary coding method based on support vector machines
CN101042735B (en) Image binarization method and device
CN103258198B (en) Character extracting method in a kind of form document image
CN107766854B (en) Method for realizing rapid page number identification based on template matching
CN105374015A (en) Binary method for low-quality document image based on local contract and estimation of stroke width
EP2605186B1 (en) Method and apparatus for recognizing a character based on a photographed image
CN105117677A (en) QR code feature detection method and system
CN104239909A (en) Method and device for recognizing images
CN102779273A (en) Human-face identification method based on local contrast pattern
CN102096802A (en) Face detection method and device
CN107705300A (en) A kind of method that blank page detection is realized based on morphological transformation
CN102610104A (en) Onboard front vehicle detection method
CN109741273A (en) A kind of mobile phone photograph low-quality images automatically process and methods of marking
CN115082776A (en) Electric energy meter automatic detection system and method based on image recognition
CN109858484B (en) Multi-class transformation license plate correction method based on deflection evaluation
CN106846815A (en) A kind of efficient intelligent traffic administration system big data analysis system
Zheng et al. A fast adaptive binarization method based on sub block OSTU and improved sauvola
CN104881663A (en) Method and device for discriminating selected result of check box
CN114155313A (en) Rapid graph automatic drawing method based on image recognition
CN115588208A (en) Full-line table structure identification method based on digital image processing technology
CN101702238A (en) Motion segmentation method based on relief image
CN110807348A (en) Method for removing interference lines in document image based on greedy algorithm
CN112070081B (en) Intelligent license plate recognition method based on high-definition video

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20180216