WO2016015548A1 - 一种用于检测预定区域中特定标识图像的方法及装置 - Google Patents

一种用于检测预定区域中特定标识图像的方法及装置 Download PDF

Info

Publication number
WO2016015548A1
WO2016015548A1 PCT/CN2015/083277 CN2015083277W WO2016015548A1 WO 2016015548 A1 WO2016015548 A1 WO 2016015548A1 CN 2015083277 W CN2015083277 W CN 2015083277W WO 2016015548 A1 WO2016015548 A1 WO 2016015548A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
predetermined area
connected domain
specific identification
detecting
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.)
Ceased
Application number
PCT/CN2015/083277
Other languages
English (en)
French (fr)
Inventor
程刚
楚汝峰
张伦
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.)
Alibaba Group Holding Ltd
Original Assignee
Alibaba Group Holding 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 Alibaba Group Holding Ltd filed Critical Alibaba Group Holding Ltd
Publication of WO2016015548A1 publication Critical patent/WO2016015548A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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 OR CALCULATING; 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 OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/24—Aligning, centring, orientation detection or correction of the image
    • G06V10/247—Aligning, centring, orientation detection or correction of the image by affine transforms, e.g. correction due to perspective effects; Quadrilaterals, e.g. trapezoids

Definitions

  • the present application relates to the field of computer applications, and in particular, to a method and apparatus for acquiring a specific identifier in a standard card.
  • the identification of a specific identifier belongs to the object detection category, and the specific identifier is generally used to indicate a certain feature of the object. For the observer, the identifier can be used as a reference for identifying the object. Certain markers typically have a distinct shape and a significant spatial location that can be more easily identified. For bank cards, the identifiers generally include the graphic identification of the card issuing organization, the text identification, etc., and identifying these marks has the following important meanings:
  • the identification of the identifier can be used as an important reference for the authenticity of the bank card.
  • Markers generally have clear size specifications and relative position marks. Legal bank cards will print the marks according to the standard. Counterfeit bank cards have low manufacturing costs and low cost. Generally, the printing control of the marks is not strict. Through the identification of the markers, the size ratio and relative position of the markers are obtained, and further comparison with the bank card of the real standard can provide an effective reference for authenticating the authenticity of the bank card.
  • the identification of the identifier can provide a reference identifier for the detection and correction of the bank card.
  • the identifier Compared with the entire bank card, the identifier has a uniform style specification, so generally there is a clear edge identifier, and the edge of the identifier is recognized, and the affine transformation identifier of the current captured image can be obtained according to the edge, so The image is corrected to facilitate accurate detection and location of subsequent bank cards.
  • the description herein as a border is not particularly accurate, and an accurate description is an edge identifier.
  • the identification of the identifier allows the service provider to perform a precise advertisement push service based on the specific identifier.
  • the identifier often indicates the specific identification of the bank card, such as the card issuing organization and the scope of use. Therefore, the identification of the identifier allows the merchant to accurately push the card for the advertisement, thereby improving the quality of the service, and in addition, obtaining the identifier of the identifier. After that, combined with technologies such as augmented reality, multimedia display can be realized.
  • Object detection means finding a specific logo or interest in an image or a video that is given.
  • the object detection system usually outputs the circumscribed rectangular frame or the precise contour of the object.
  • Object detection can be roughly divided into sliding window and generalized Hough transform.
  • the sliding window method is the current mainstream object detecting method, and the sliding window method is to perform sliding scanning on multiple scales and multiple angles of the input image by using the trained template, and find the external window of the target object by determining the maximum corresponding position.
  • the sliding window method used in the prior art object detecting method has a large computational complexity and a relatively high computational complexity when analyzing and scanning data, thereby making the detection time longer, and the method is more complicated for the object structure. Detection. For the identification of a more fixed structure, it is obvious that the sliding window method does not apply.
  • the present application provides a method for detecting a specific identification image in a predetermined area to solve the problem of high computational complexity and large computational complexity in the prior art detection process.
  • the present application provides a method for detecting a specific identification image in a predetermined area, including:
  • the acquiring the image of the predetermined area or the predetermined area as the image to be inspected includes:
  • Locating the preview image within the area acquiring the predetermined area or including a predetermined area An image of the domain, and the image within the bounding area of the frame is taken as the target image to be inspected.
  • the calculating the range of the area of the preview image means: obtaining a virtual border parameter of the preview image according to the actual size parameter of the predetermined area and the resolution parameter of the preview image, specifically:
  • the width of the virtual border is: a predetermined area or an actual width of the image including the predetermined area multiplied by the height of the preview image divided by the predetermined area or the actual height of the image including the predetermined area multiplied by the virtual border and the preview image. a ratio; the height of the virtual border is: a preview image height multiplied by a ratio of a virtual border to the preview image;
  • the width of the virtual border is: the width of the preview image multiplied by the ratio of the virtual border to the preview image;
  • the height of the virtual border is: the actual height of the predetermined area or the image containing the predetermined area multiplied by the width of the preview image divided by a predetermined area or an actual width of the image including the predetermined area multiplied by a ratio of the virtual border to the preview image;
  • the acquiring the predetermined area or the image including the predetermined area as the target image to be inspected includes: capturing the specific identification image according to the specific location information in the predetermined area or the predetermined area image of the specific identification image The area of the area as the target image to be inspected.
  • the specific location information is a location in which a specific identification image set in advance is located in a predetermined area or on an image including a predetermined area.
  • the specific position is an upper right corner or a lower right corner or an upper left corner or a lower left corner of an image located in the predetermined area or containing the predetermined area.
  • intercepting the area range including the specific identification image as the target image to be detected includes:
  • An image area including the specific identification in the predetermined area or the image including the predetermined area is intercepted as the target image to be inspected based on the position information.
  • an upper limit value of the threshold interval is a product of the acquired predetermined area image or a width, a height, and an upper limit coefficient including a predetermined area image; and a lower limit value is the acquired predetermined area image or includes a predetermined area image The product of the width, height, and lower bound coefficients.
  • comparing the image feature of the candidate connected domain region with the corresponding image feature of the specific identifier image comprises: using an individual image feature comparison or a combined image feature comparison; wherein the image feature includes The volume ratio of the connected domain, the long axis direction of the connected ellipse, the outer contour shape of the connected domain, and/or the circumscribed rectangle aspect ratio of the connected domain.
  • image features of the candidate connected domain region are acquired one by one, and image features of the candidate connected domain region are compared with corresponding image features of the specific identification image, and candidate connections with the same or closest matching result are compared.
  • the domain area is regarded as the area where the specific identification image is located, including:
  • the candidate connected domain regions are filtered by using the volume ratio of the connected domain and the long axis direction of the connected ellipse as the combined image feature, specifically:
  • the candidate connected domain region is again filtered by using the connected domain outer contour shape as a separate condition, specifically:
  • the acquiring sampling pixel points of the upper and lower edges and the left and right edges of the edge contour image includes:
  • Scanning is performed on the condition that there are two intersections with the filled image, and the intersection points are taken as sampling pixel points;
  • the selecting the upper and lower edge sampling pixel points and the left and right edge sampling pixel points within the set threshold range include:
  • the threshold range is maxDist ⁇ maxDist x 5%.
  • the specific identification image is a logo graphic of a bank card association organization or a bank card issuance organization.
  • the present application further provides an apparatus for detecting a specific identification image in a predetermined area, including:
  • An acquiring unit configured to acquire the predetermined area or an image including the predetermined area as an image to be inspected
  • a binarization processing unit configured to perform binarization processing on the image to be inspected, obtain a binarized image, and invert the image of the binarized image, both of which are referred to as binary images of the target to be inspected ;
  • a connected domain calibration unit configured to calibrate a connected domain of the target binary image to be detected, and use a connected domain in the calibrated connected domain to satisfy a set threshold interval as a candidate connected domain;
  • an image feature comparison unit configured to acquire the image features of the candidate connected domain one by one, and compare the image features of the candidate connected domain with the corresponding image features of the standard specific identifier image, and compare the results to be the same or closest
  • the candidate connected domain is considered to be the location of the specific logo image.
  • the obtaining unit comprises:
  • a preview image acquisition unit configured to acquire a preview image of the predetermined area or an image including the predetermined area
  • a calculating unit configured to calculate a region range of the preview image
  • a frame unit configured to frame the preview image within the range of the region, acquire the predetermined region or an image including the predetermined region, and use the image within the framed region as the target image to be inspected.
  • the calculating unit refers to: obtaining a virtual border parameter of the preview image according to an actual size parameter of the predetermined area and a resolution parameter of the preview image, specifically:
  • the width of the virtual border is: a predetermined area or an actual width of the image including the predetermined area multiplied by the height of the preview image divided by the predetermined area or the actual height of the image including the predetermined area multiplied by the virtual border and the preview image. a ratio; the height of the virtual border is: a preview image height multiplied by a ratio of a virtual border to the preview image;
  • the width of the virtual border is: the width of the preview image multiplied by the ratio of the virtual border to the preview image;
  • the height of the virtual border is: the actual height of the predetermined area or the image containing the predetermined area multiplied by the width of the preview image divided by a predetermined area or an actual width of the image including the predetermined area multiplied by a ratio of the virtual border to the preview image;
  • the obtaining unit comprises:
  • An intercepting unit for specializing in the predetermined area or the predetermined area image according to the specific identification image The position information is intercepted, and the range of the area including the specific identification image is intercepted as the target image to be inspected.
  • the specific location information is a location in which a specific identification image set in advance is located in a predetermined area or on an image including a predetermined area.
  • the specific position is an upper right corner or a lower right corner or an upper left corner or a lower left corner of an image located in the predetermined area or containing the predetermined area.
  • the intercepting unit comprises:
  • a traversing unit configured to traverse the standard image template according to the predetermined area or an image including the predetermined area, and obtain a specific identification image by using a correspondence between the standard image template and the predetermined area or an image including a predetermined area a location or location information in an image containing a predetermined area;
  • the object to be detected image determining unit is configured to intercept, according to the position information, an image area including the specific identifier in the predetermined area or the image including the predetermined area as the target image to be inspected.
  • an upper limit value of the threshold interval is a product of the predetermined area or an image width, a height, and an upper limit coefficient including a predetermined area; and a lower limit is an image width, a height, and a lower limit of the predetermined area or a predetermined area The product of the coefficients.
  • comparing the image feature of the candidate connected domain region with the corresponding image feature of the specific identifier image comprises: using an individual image feature comparison or a combined image feature comparison; wherein the image feature includes The volume ratio of the connected domain, the long axis direction of the connected ellipse, the outer contour shape of the connected domain, and/or the circumscribed rectangle aspect ratio of the connected domain.
  • the method includes: combining image feature screening units, specifically:
  • An external border obtaining unit is configured to obtain an external border of the candidate connected domain area
  • a volume ratio calculating unit configured to calculate a volume ratio of a pixel of a foreground color in the candidate connected domain to an area of the outer border in the circumscribed frame
  • a volume ratio threshold value determining unit configured to determine whether the volume ratio is greater than a threshold range of a preset specific identification image; if not, discarding the candidate connected domain region; if yes, filling the candidate connected domain region;
  • a filling calculation unit configured to calculate a direction of the circumscribed ellipse long axis of the candidate connected domain region after the filling, and obtain an angle between the direction and the set reference axis;
  • An angle threshold determining unit is configured to determine whether the angle is within a predetermined threshold range, and if so, The candidate connected domain region is regarded as the region where the specific identification image is located; if not, the candidate connected domain region is discarded.
  • the individual image feature screening unit comprises:
  • a padding unit configured to fill the candidate connected domain region to obtain a padding image
  • An edge contour acquiring unit configured to obtain an edge contour image of the filled image according to detecting an edge of the filled image
  • a sampling unit configured to respectively acquire sampling pixel points of the upper and lower edges and the left and right edges of the edge contour image, and select the sampling pixel points and the left and right edge sampling pixel points in the set threshold range;
  • a sampling pixel point fitting unit configured to respectively acquire sampling pixel points of the upper and lower edges and the left and right edges of the edge contour image, and select the sampling pixel points and the left and right edge sampling pixel points in the set threshold range;
  • a line segment combination determining unit configured to determine whether the line segment combination information and the intersection location information of the intersection point location information and the standard specific identifier image are the same; if yes, the candidate connected domain region is regarded as a specific identifier image region; , the output does not detect the specific identification image area.
  • the sampling unit comprises:
  • Scanning unit performs scanning on the condition that there are two intersections with the filled image, and the intersection points are used as sampling pixel points;
  • the distance threshold determining unit is configured to determine whether the distance between the upper and lower edges of the sampled pixel points and the distance between the left and right edges are within a set threshold range, and if so, select; if not, discard.
  • the threshold range is maxDist ⁇ maxDist x 5%.
  • An affine transformation unit configured to calculate an affine transformation matrix according to the predetermined area or the location information including the specific identifier in the predetermined area image and the respective intersection location information;
  • a related information acquiring unit configured to acquire related information according to the affine transformation matrix.
  • the specific identification image is a logo graphic of a bank card association organization or a bank card issuance organization.
  • the present invention obtains an image of a predetermined area as a target image to be inspected, performs binarization processing on the image to be inspected, and calibrates the connected domain in the binarized image as a candidate connected domain image, and images of the candidate connected domains in the candidate connected domain image A connected domain that matches the characteristics of the specific identifier image is selected to determine that the specific identification information is detected. Since the present application compares only the predetermined area or the image feature of the candidate connected-domain image in the predetermined area image with the standard-specific identification image feature, when the data is scanned and analyzed, it is performed in a smaller area. Therefore, this method can reduce the amount of calculation in the detection process, and reduce the computational complexity, and can quickly complete the detection according to the feature comparison.
  • FIG. 1 is a flow chart of a method for detecting a specific identification image in a predetermined area provided by the present application
  • FIG. 2 is a schematic diagram of calculation of a virtual border in a method for detecting a specific identification image in a predetermined area provided by the present application;
  • FIG. 3 is a schematic diagram of a virtual edge in a method for detecting a specific identification image in a predetermined area provided by the present application
  • FIG. 4 is a schematic diagram of a first embodiment of intercepting a range of regions including a specific identification image in a method for detecting a specific identification image in a predetermined area provided by the present application;
  • FIG. 5 is a flow chart of a second embodiment of intercepting a range of regions including a specific identification image in a method for detecting a specific identification image in a predetermined area provided by the present application;
  • FIG. 6 is a schematic diagram of a binarized image after performing binarization processing on a target image to be inspected in a method for detecting a specific logo image in a predetermined area according to the present application;
  • FIG. 7 is a flowchart of a first screening of candidate connected domain regions in a method for detecting a specific identification image in a predetermined area provided by the present application;
  • FIG. 8 is a flowchart of a second screening of a candidate communication area in a method for detecting a specific identification image in a predetermined area provided by the present application;
  • FIG. 9 is a schematic diagram of detecting a top and bottom of a UnionPay logo in a method for detecting a specific logo image in a predetermined area provided by the present application;
  • FIG. 10 is a structural block diagram of an apparatus for detecting a specific identification image in a predetermined area provided by the present application.
  • the area where the bank card face is located is the predetermined area
  • the UnionPay logo on the bank card is a specific identifier, which illustrates the specific implementation process of the present application.
  • the method of the present application is not limited to the UnionPay logo on the bank card. Identification, the method of the present application can be used for all identification scenarios having specific known identifiers in a predetermined area, the specific known identifiers may be product identifications, corporate logos, etc., the predetermined area of the product identification may be the product itself; the logo of the enterprise The predetermined area may be any carrier itself that contains the logo.
  • FIG. 1 is a flowchart of a first embodiment of a method for detecting a specific identification image in a predetermined area.
  • the method includes the following steps:
  • Step S100 Acquire an image of a predetermined area as an image to be inspected
  • Step S200 Perform binarization processing on the image to be inspected to obtain a binarized image, and an inverse image of the binarized image, both of which are referred to as a target binary image to be inspected;
  • Step S300 calibrating the connected domain of the to-be-detected target binary image, and using the area of the connected domain in the calibrated connected domain to satisfy the set threshold interval as the candidate connected domain region;
  • Step S400 Acquire image features of the candidate connected domain region one by one, and compare image features of the candidate connected domain region with corresponding image features of the standard specific identifier image, and compare candidate connected domains with the same or closest comparison result
  • the area is considered to be the location of the specific logo image.
  • the above detection method can reduce the amount of calculation in the detection process, and reduce the computational complexity, and can quickly complete the detection according to the feature comparison.
  • the step S100 acquiring an image of a predetermined area as a target image to be inspected.
  • the predetermined area in the step may be a bank card face area, that is, an image of a bank card face or an image containing a card face area
  • the acquisition mode may be obtained by using a mobile terminal or automatically obtained, for example, by The imaging function of the mobile terminal acquires a predetermined area or an image containing the predetermined area.
  • the present application may first obtain the bank card map obtained when acquiring the bank card image or the image containing the bank card area. Like or including a preview image of the bank card face area image, and according to the actual size of the standard bank card and the resolution of the preview image, obtain the length and width of the virtual frame that frames the preview image; and then according to the length and width of the virtual frame Calculate the upper, lower, left and right borders of the virtual border, as follows:
  • FIG. 2 is a schematic diagram of virtual border calculation in a method for detecting a specific identification image in a predetermined area provided by the present application.
  • the standard size parameters of the bank card card can be known.
  • PreviewWidth ⁇ PreviewHeight the ratio parameter of the virtual border to the preview image is DisplayRatio; when calculating the virtual border, first determine whether the PreviewHeight ⁇ CardWidthGdt (card width) is greater than (excluding equal to) PreviewWidth (preview image width) ⁇ CardHeightGdt (card height), if yes, then:
  • ScaledWidth CardWidthGdt ⁇ PreviewHeight/CardHeightGdt ⁇ DisplayRatio;
  • ScaledHeight PreviewHeight ⁇ DisplayRatio
  • ScaledHeight CardHeightGdt ⁇ PreviewWidth/CardWidthGdt ⁇ DisplayRatio;
  • ScaledWidth PreviewWidth ⁇ DisplayRatio
  • the upper and lower left and right borders of the virtual border may be determined according to the resolution parameter of the screen of the mobile terminal, as follows:
  • FIG. 3 is a schematic diagram of a virtual edge in a method for detecting a specific identification image in a predetermined area provided by the present application.
  • the virtual border is overlaid on the bank card image or the bank card in the image area of the bank card, and the area within the range covered by the virtual frame is used as the target image to be detected, thereby reducing the detection range. It can also be based on the specific location of the UnionPay logo in the standard bank within the coverage of the virtual border.
  • the information intercepts the range of the area including the UnionPay logo image, and uses the area range as the target image to be inspected, thereby further reducing the detection range, thereby improving the detection speed and reducing the amount of calculation.
  • FIG. 4 is a schematic diagram of a first embodiment of intercepting a range of regions including a specific identification image in a method for detecting a specific identification image in a predetermined area provided by the present application.
  • the location information of the UnionPay logo on the bank card is the pre-set position.
  • the UnionPay logo area is located on the right side of the card and is divided into lower positions. Both format and top format.
  • the UnionPay logo When the UnionPay logo adopts the lower format, it must be located in the lower right corner of the card.
  • the specific requirements are: the width of the UnionPay logo is 22mm, the height is 15mm, the distance from the right edge of the UnionPay logo to the right edge of the card is 2mm, and the lower edge of the UnionPay logo is to the card. The distance to the lower edge is 2mm.
  • the UnionPay logo When the UnionPay logo adopts the upper format, it must be located in the upper right corner of the card.
  • the specific requirements are: the width of the UnionPay logo is 22mm, the height is 15mm, the distance from the right edge of the UnionPay logo to the right edge of the card is 2mm, and the upper edge of the UnionPay logo is attached to the card. The distance from the upper edge is 2mm.
  • the resolution (width ⁇ height) is: Width ⁇ Height
  • the images to be captured are [Width/2, Width] ⁇ [0, Height] and [Width/2, Width] ⁇ [Height/2, Height] .
  • FIG. 5 is a flowchart of a second embodiment for intercepting a range of regions including a specific identification image in a method for detecting a specific identification image in a predetermined area;
  • Step S101 Establish a standard image template of the bank card image
  • Step S102 traverse the standard image template according to the bank card image or an image including a bank card area, and obtain a UnionPay logo image by using a correspondence between the standard image template and the bank card image or an image including a bank card area. Location information in the bank card image or in the image of the bank card area;
  • Step S103 The area range including the UnionPay logo image is intercepted according to the location information, and the area range of the intercepted UnionPay logo image is used as the target image to be detected.
  • the present application can directly intercept the image area including the UnionPay logo according to the bank card image or the image containing the bank card area according to the specific location information of the UnionPay logo or according to the bank card standard image template.
  • the image area is used as the target image to be inspected.
  • On the basis of the virtual border only the portion including the UnionPay logo image area in the virtual border area may be intercepted, and the image information other than the virtual border may be discarded, thereby improving the detection efficiency.
  • Step S200 Perform binarization processing on the image to be inspected to obtain a binarized image, and an inverse image of the binarized image, which are both referred to as a target binary image to be inspected.
  • FIG. 6 is a schematic diagram of a binarized image after performing binarization processing on a target image to be inspected in a method for detecting a specific logo image in a predetermined area according to the present application;
  • the target image to be inspected that is, the image of the region range of the intercepted image including the UnionPay logo image is binarized.
  • Image binarization refers to setting the gray value of the pixel on the image to 0 or 255, that is, presenting the entire image with a distinct black and white visual effect; that is, a grayscale image of 256 brightness levels.
  • a binarized image that still reflects the overall and local features of the image is obtained by appropriate threshold selection.
  • the gray image is binarized to obtain a binarized image, which is beneficial to the image collection property when the image is further processed. It is only related to the position of a point whose pixel value is 0 or 255, and the multi-level value of the pixel is no longer involved, making the processing simple, and the processing and compression amount of data are reduced.
  • the Otsu method can be used to divide the grayscale image into two parts: the background and the target (foreground).
  • the Otsu method also known as the maximum inter-class variance method, is an adaptive threshold determination method that divides an image into two parts, a background and a target, and an inter-class between the background and the target according to the gray specificity of the image. The larger the variance, the greater the difference between the background and the target image.
  • the number of foreground pixels accounts for the ratio of the image and the average gray value; the number of background pixels accounts for the ratio of the image and the average gray value, and the variance of the foreground and background images is calculated.
  • the gray value corresponding to the maximum variance is the global threshold.
  • the image is binarized according to the global threshold.
  • the white pixel (target pixel) in the binarized image is to be marked, and the image containing the UnionPay logo is binarized as the target image to be detected, the UnionPay logo may be It will be binarized into foreground (1) or background (0). Therefore, after the binarized image is obtained, the binarized image can be inverted, and both the binarized image and the inverted image are used. Target binary image to be checked.
  • the most important method of binary image analysis is the connected region marker, which is used for all binary image analysis. Basically, it makes each individual connected area form an identified block by marking the white pixels (targets) in the binary image. Further, we can obtain the contours, circumscribed rectangles, centroids, and invariants of these blocks. Equal geometric parameters.
  • the connected-domain labeling operation of the binary image is to be adjacent to each other in a bitmap image combined with white pixels (usually binary images with "1") and black pixels (usually represented by "0"). Or a set of pixels having a pixel value of "1" is extracted.
  • the binary image contains black (pixel value is 0) white (pixel value is 1 or 255) two colors, respectively as the target color and background color, and the marking algorithm only marks the target pixel.
  • the step S300 calibrating the connected domain of the target binary image to be detected, and using the connected domain in the calibrated connected domain to satisfy the set threshold interval as the candidate connected domain.
  • Connected domain calibration is the basis of all binary image analysis.
  • the connected domain labeling operation of a binary image is a combination of white pixels (usually binary images with "1") and black pixels (usually represented by "0").
  • the bitmap image the pixel sets having the pixel value "1" adjacent to each other are extracted, that is, the binary image contains black (pixel value is 0) white (pixel value is 1 or 255) two colors.
  • the calibration algorithm only calibrates the target pixel (white pixel) in the connected domain.
  • 4 adjacencies and 8 adjacencies there are two common adjacencies: 4 adjacencies and 8 adjacencies.
  • 8 adjacent points have a total of 8 points, including the diagonal position of the point, namely: up, down, left, right, top left and top right.
  • the set formed by the points connecting the upper, lower, left, right, or upper, lower, left, right, upper left, and upper right is marked as a connected domain.
  • the connected domain calibration method is adopted for the binarized image, and the connected domain in the binarized image is calibrated, that is, the calibration pixel value is 1. All connected domains. Threshold judgment is performed on each of the independent connected domains that are calibrated. If the number of pixels in the connected domain is within a given threshold interval, the connected domain is retained, otherwise the inverted image in the binary image of the target is connected. The domain is calibrated, and the number of connected domain pixels in the inverted image is thresholded. If the number of connected domains is within the threshold interval, the connected domain is reserved, otherwise the connected domain is discarded.
  • the number of pixels in the connected domain may be within a reasonable range, and the reasonable range may set a threshold interval belonging to the connected domain of the UnionPay logo image.
  • Width and Height are respectively the width and height of the acquired image of the predetermined area or the image containing the predetermined area, that is, Width and Height may be the width and height of the image to be inspected, and the lower limit coefficient and the upper limit may be The empirical values are 0.011 and 0.043, respectively.
  • a threshold interval may also be set according to a reasonable range in which the number of pixels of the specific identifier connected domain is within.
  • Step S400 Acquire image features of the candidate connected domain region one by one, and compare image features of the candidate connected domain region with corresponding image features of the standard specific identifier image, and connect candidate structures with the same or close comparison structure
  • the domain area is considered to be the location of the specific logo image.
  • the specific implementation process of the step is to compare the image features of the candidate UnionPay logo connected domain region with the standard UnionPay logo image, and compare the corresponding image features of the two, which may be performed by using separate image feature comparison or combined image feature ratio.
  • image features include: a volume ratio including a connected domain, a connected domain ellipsoid long axis direction, a connected domain outer contour shape, and/or a connected domain circumscribed rectangle aspect ratio, etc., it is also understood that the feature belongs to an image.
  • Shape features, and shape features may also include distance features, image relief features, length features, and the like.
  • the connected domain region of the candidate UnionPay logo image may be filtered by using the volume ratio of the connected domain and the long axis direction of the outer ellipse of the connected domain as the first combined image feature.
  • FIG. 7 is a flowchart of a first screening of candidate connected domain regions in a method for detecting a specific identification image in a predetermined area.
  • the specific steps include:
  • Step S701 Acquire an external border of the candidate connected domain area
  • scanning may be performed inward from the outermost edge of the candidate UnionPay logo connected domain image. After detecting the foreground color (white pixel), scanning is stopped to obtain an external border, and the external border is rectangular.
  • Step S702 Calculate a volume ratio of a pixel of a foreground color in the candidate connected domain to an area of the outer border in the outer border;
  • the number of pixels in the foreground color (white) in the outer border of the rectangle is counted, and the number of white pixels is divided by the area of the outer border of the rectangle to obtain the connected domain volume ratio.
  • Step S703 determining whether the volume ratio is greater than a threshold range of a specific identification image set in advance, If not, discarding the candidate connected domain region; if yes, filling the candidate connected domain region;
  • the threshold value Tsolid of the standard UnionPay logo connected domain volume ratio is set in advance, and it is determined whether the candidate UnionPay logo connected domain volume ratio matches the image feature of the standard UnionPay logo connected domain volume ratio.
  • the threshold Tsolid can be obtained by using the information of the standard UnionPay logo image, for example, obtaining a standard UnionPay logo image from various angles, and calculating a volume ratio thereof to the circumscribing matrix, thereby selecting a threshold value Tsolid, identifying the connected domain volume ratio and the Tsolid by the candidate UnionPay For comparison, it is determined whether the candidate connected domain is a UnionPay logo image.
  • the candidate UnionPay identity connection domain is a UnionPay identity connectivity domain, because there may be a non-UnionPay identity in the candidate UnionPay identity connectivity domain that is identical to the standard UnionPay identity connectivity domain volume ratio image feature. Connected domain, therefore, the present application further combines another image feature, that is, the image feature of the long axis direction of the connected domain ellipse, to further determine the candidate UnionPay logo. After determining that the candidate UnionPay logo connected domain volume ratio is greater than the threshold Tsolid, the image feature determination of the long axis direction of the connected domain ellipse is entered, as described in step S704 below.
  • Step S704 Filling the candidate connected domain area
  • the foreground color (white) of the UnionPay logo candidate connected domain area is filled.
  • Step S705 Calculate a direction of the filled UnionPay identifier candidate connected domain region after the filling
  • the direction of the connected domain can be calculated by using an ellipse fitting method.
  • the direction of the connected domain is the direction of the connected domain with the long-axis direction of the circumscribed ellipse obtained on the connected domain.
  • the ellipse fitting is performed by performing edge detection on the candidate connected domains, recording points on the edges, and performing ellipse fitting on the points.
  • the coordinate axis is established with the lower left corner of the connected domain as the origin, and the X-axis of the coordinate axis is defined as the reference axis, and the ellipse long axis forms an angle with the X-axis.
  • the angle formed between the long axis of the ellipse and the X axis can also be calculated.
  • Step S706 determining whether the included angle is within a predetermined threshold range. If yes, the candidate connected domain area of the UnionPay logo is regarded as the area where the UnionPay logo image is located; if not, the candidate connected domain area of the UnionPay logo is discarded. ;
  • the threshold range of the included angle may be elliptically fitted to the standard UnionPay logo image according to the above, and the angle between the long axis direction of the ellipse and the reference axis is calculated, because it is possible to obtain the target image of the UnionPay logo to be inspected.
  • the present application is based on the angle between the long axis of the ellipse and the reference axis in the standard UnionPay logo.
  • the threshold range of the angle is to avoid the influence of the judgment due to the problem of the angle of acquisition.
  • the reference axis here is also the X axis.
  • the above-mentioned reference axis setting can be set according to actual judgment, that is, it can also be a Y-axis, and can satisfy the requirement of judging the characteristics of the long-axis direction of the connected ellipse in the connected domain.
  • the second volume ratio judgment of the UnionPay identification candidate connected domain image may be performed again on the basis of the above.
  • the threshold value is not the above Tsolid, but another threshold value Tsolid' set for this judgment, and the detection determination process is as described above in steps S701 to S706. It should be noted that the pixel value of the inner layer connected domain is different from the pixel value of the outer connected domain, so the threshold is also different.
  • the present application determines, on the basis of the foregoing steps, that the angle between the major axis of the ellipse and the reference axis is within a predetermined threshold range.
  • the candidate connected domain region is subjected to a second screening by using the connected domain outer contour shape as a separate condition.
  • FIG. 8 is a method for detecting a specific identifier image in a predetermined area according to the present application, and performing a second screening of the candidate connected domain region by using a connected outer contour shape as a separate condition. flow chart.
  • Step S801 Fill the candidate connected domain area to obtain a filled image.
  • the candidate connected domain region may be filled by using the filling manner in step S704.
  • Step S802 detecting an edge of the filled image to obtain an edge contour image of the filled image
  • the edge detection method is a commonly used technique in image processing, which can capture important features of objects in an image.
  • Common edge detection methods include: edge detection, gradient edge detection, Roberts edge detection operator, Sobel edge detection algorithm. Sub, Prewitt edge detection operators and Laplace edge detection operators, and so on.
  • the Sobel edge detection operator is used to detect the edge of the filled image.
  • the Sobel edge detection operator is mainly based on the gray-scale weight difference between the upper and lower adjacent points of the pixel, and reaches the extreme value detection edge at the edge.
  • Sobel operators one for detecting horizontal edges and the other for detecting vertical edges.
  • the Sobel operator weights the influence of the position of the pixel, which can reduce the degree of edge blur. This effect is better.
  • edge detection operators can also be used to detect the edges of the filled image according to different needs.
  • Step S803 respectively acquiring sampling pixel points of the upper and lower edges and the left and right edges of the edge contour image, and selecting the upper and lower edge sampling pixel points and the left and right edge sampling pixel points within a set threshold range;
  • the sampling pixels of the upper and lower edges and the left and right edges in the edge contour image are sampled by scanning.
  • the edge contour of the filled image is a rectangular image, and the upper and lower edges and the left and right edges of the rectangular image are respectively scanned, and the scanning manners of the upper and lower edges and the left and right edges can be performed in the same manner. The above lower edge is illustrated.
  • the threshold range of the maximum distance is set to select the sampling pixel, and the threshold range can be set at maxDist ⁇ maxDist ⁇ 5%, and the pixel within the threshold is selected as the sampling pixel. Points, pixels outside the threshold range are discarded. Thereby, the sampling pixel points of the upper and lower edges of the rectangular image are obtained, and similarly, the sampling pixel points of the left and right edges can be obtained.
  • Step S804 fitting the selected sampling pixel points to obtain line segment combination information of the edge contour image.
  • Step S604 can use the polynomial curve fitting method to fit the selected sampled pixel points by using the polynomial curve fitting method, thereby obtaining four line segments, and combining the four line segments by using the sampling pixel points of the four sides of the candidate connected-domain rectangular image selected in step S803.
  • the four vertices are obtained.
  • the specific implementation of step S804 is an example for the bank card UnionPay logo. It can be understood that if other identifiers can adopt different fitting methods or calculation methods, the sampling pixel points are fitted.
  • Step S805 determining whether the line segment combination information is the same as the line segment combination information of the standard specific identification image, and if so, treating the candidate connected domain area as the specific identification image area; if not, outputting the undetected specific identification image area.
  • the UnionPay identification candidate connected domain region is a UnionPay logo image.
  • the candidate UnionPay logo is detected, and the system can output the four corner points of the UnionPay logo (refer to FIG. 9).
  • the candidate area may be determined to be a non-banking icon. If the detected UnionPay identification candidate area is empty, the user is prompted to not detect the UnionPay logo. If the number of the UnionPay logo candidate areas detected is more than one, all candidate areas are provided to the user for user autonomous selection.
  • the image features are shape features including parallelograms, four vertices, and the like; when the specific logo is a text or product logo or When the enterprise logo is used, it can be detected by the image features of the application.
  • the logo of each issuing bank is described by taking the logo of China Merchants Bank as an example. Please refer to Figure 3.
  • the image feature of the logo image of China Merchants Bank can be Each of the vertices, triangles, or embossed attributes of the circle, the letter M, and the letter M can be detected by the above detection methods by using the feature attributes.
  • the method for detecting a specific identifier provided by the present application is not limited to the above-mentioned UnionPay logo.
  • the affine transformation matrix of the UnionPay logo image may be calculated according to the detected four vertex information of the UnionPay logo and the vertex information of the standard UnionPay logo; according to the matrix, other information on the bank card may be calculated.
  • the affine transformation matrix can be obtained as follows:
  • the coordinates of the four vertices of the standard UnionPay logo are [x1, y1], [x2, y2], [x3, y3], and [x4, y4], and the coordinates of the four vertices in the candidate connected-domain image are detected.
  • the affine transformation matrix M can be calculated according to the correspondence of 4 points, for each of the detected images.
  • the pixel can calculate the image after the projective transformation, and the image in which the image is located is the positive view image.
  • the embodiment of the present invention discloses an embodiment of a method for detecting a specific identification image in a predetermined area.
  • the present application further discloses an embodiment for detecting a specific identification image device in a predetermined area.
  • FIG. 10 is a structural diagram of an embodiment of an apparatus for detecting a specific identification image in a predetermined area according to the present application. Since the device embodiment is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method embodiment.
  • the device embodiments described below are merely illustrative.
  • the present application provides an apparatus for detecting a specific identification image in a predetermined area, including:
  • the acquiring unit A100 is configured to acquire the predetermined area or an image including the predetermined area as an image to be inspected;
  • the binarization processing unit A200 is configured to perform binarization processing on the image to be inspected to obtain a binarized image, and an inverse image of the binarized image, which are both called target binary values image;
  • a connectivity domain calibration unit A300 configured to calibrate a connected domain of the to-be-detected target binary image, and use a connected domain in the calibrated connected domain to satisfy a set threshold interval as a candidate connected domain;
  • the image feature comparison unit A400 is configured to acquire image features of the candidate connected domain one by one, and compare the image features of the candidate connected domain with corresponding image features of the standard specific identification image, and compare or match the image with the same
  • the candidate connected domain is considered to be the location of the specific identification image.
  • the obtaining unit A100 includes:
  • a preview image acquisition unit configured to acquire a preview image of the predetermined area or an image including the predetermined area
  • a calculating unit configured to calculate a region range of the preview image; the calculating unit is configured to obtain a virtual border parameter of the preview image according to an actual size parameter of the predetermined region and a resolution parameter of the preview image, specifically :
  • the width of the virtual border is: a predetermined area or an actual width of the image including the predetermined area multiplied by the height of the preview image divided by the predetermined area or the actual height of the image including the predetermined area multiplied by the virtual border and the preview image. a ratio; the height of the virtual border is: a preview image height multiplied by a ratio of a virtual border to the preview image;
  • the width of the virtual border is: the width of the preview image multiplied by the ratio of the virtual border to the preview image;
  • the height of the virtual border is: the actual height of the predetermined area or the image containing the predetermined area multiplied by the width of the preview image divided by a predetermined area or an actual width of the image including the predetermined area multiplied by a ratio of the virtual border to the preview image;
  • a frame unit configured to frame the preview image within the range of the region, acquire the predetermined region or an image including the predetermined region, and use the image within the framed region as the target image to be inspected.
  • the obtaining unit A100 further includes:
  • an intercepting unit configured to intercept, according to the specific location information of the specific identifier image in the predetermined area or the predetermined area image, a range of the area including the specific identification image as the target image to be detected.
  • the specific location information in the intercepting unit is a location where the specific identification image set in advance is located in a predetermined area or on the image including the predetermined area.
  • the specific location may be known according to the records in the third volume of the UnionPay Card Business Operation Regulations, "Card No. BIN Number and Identification Rules", and the specific location of the UnionPay logo is located.
  • the interception unit can also be implemented as follows:
  • a traversing unit configured to traverse the standard image template according to the predetermined area or an image including the predetermined area, and obtain a specific identification image by using a correspondence between the standard image template and the predetermined area or an image including a predetermined area a location or location information in an image containing a predetermined area;
  • the object to be detected image determining unit is configured to intercept, according to the position information, an image area including the specific identifier in the predetermined area or the image including the predetermined area as the target image to be inspected.
  • the upper limit value of the threshold interval in the connected-domain calibration unit A300 is the product of the predetermined region or the image width, the height, and the upper limit coefficient including the predetermined region; the lower limit is the image width and height of the predetermined region or the predetermined region. The product of the lower bound coefficient.
  • the image feature comparison unit A400 includes a combined image feature screening unit or an individual image feature screening unit, by using separate image feature comparison or combined image feature comparison; wherein the image feature includes a volume ratio of the connected domain and a connected domain externally The ellipse long axis direction, the connected domain outer contour shape, and/or the connected domain circumscribed rectangle aspect ratio.
  • the combined image feature screening unit is specifically:
  • An external border obtaining unit is configured to obtain an external border of the candidate connected domain area
  • a volume ratio calculating unit configured to calculate a volume ratio of a pixel of a foreground color in the candidate connected domain to an area of the outer border in the circumscribed frame
  • a volume ratio threshold determining unit configured to determine whether the volume ratio is greater than a preset specific identification map a threshold range of the image; if not, discarding the candidate connected domain region; if so, filling the candidate connected domain region;
  • a filling calculation unit configured to calculate a direction of the circumscribed ellipse long axis of the candidate connected domain region after the filling, and obtain an angle between the direction and the set reference axis;
  • the angle threshold determining unit is configured to determine whether the angle is within a predetermined threshold range, and if yes, the candidate connected domain area is regarded as the area where the specific identification image is located; if not, the candidate connected domain area is discarded.
  • a separate image feature screening unit comprising:
  • a padding unit configured to fill the candidate connected domain region to obtain a padding image
  • An edge contour acquiring unit configured to obtain an edge contour image of the filled image according to detecting an edge of the filled image
  • a sampling unit configured to respectively acquire sampling pixel points of the upper and lower edges and the left and right edges of the edge contour image, and select the sampling pixel points and the left and right edge sampling pixel points in the set threshold range;
  • a sampling pixel point fitting unit configured to respectively acquire sampling pixel points of the upper and lower edges and the left and right edges of the edge contour image, and select the sampling pixel points and the left and right edge sampling pixel points in the set threshold range;
  • a line segment combination determining unit configured to determine whether the line segment combination information and the intersection location information of the intersection point location information and the standard specific identifier image are the same; if yes, the candidate connected domain region is regarded as a specific identifier image region; , the output does not detect the specific identification image area.
  • the sampling unit includes:
  • Scanning unit performs scanning on the condition that there are two intersections with the filled image, and the intersection points are used as sampling pixel points;
  • the distance threshold determining unit is configured to determine whether the distance between the upper and lower edges of the sampled pixel points and the distance between the left and right edges are within a set threshold range, and if so, select; if not, discard.
  • the threshold range may be maxDist ⁇ maxDist ⁇ 5%, where maxDist is the maximum distance of two sampled pixel points.
  • the application further includes an affine transformation unit for calculating an affine transformation matrix according to a predetermined area or location information including a specific identifier in the predetermined area image and each intersection location information; and a related information acquisition unit for performing the affine according to the affine The transformation matrix obtains relevant information.
  • a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
  • processors CPUs
  • input/output interfaces network interfaces
  • memory volatile and non-volatile memory
  • the memory may include non-persistent memory, random access memory (RAM), and/or non-volatile memory in a computer readable medium, such as read only memory (ROM) or flash memory.
  • RAM random access memory
  • ROM read only memory
  • Memory is an example of a computer readable medium.
  • Computer readable media including both permanent and non-persistent, removable and non-removable media may be implemented by any method or technology.
  • the information can be computer readable instructions, data structures, modules of programs, or other data.
  • Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory. (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disk read only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, Magnetic tape cartridges, magnetic tape storage or other magnetic storage devices or any other non-transportable media can be used to store information that can be accessed by a computing device.
  • computer readable media does not include non-transitory computer readable media, such as modulated data signals and carrier waves.
  • embodiments of the present application can be provided as a method, system, or computer program product.
  • the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment in combination of software and hardware.
  • the application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)
  • Geometry (AREA)

Abstract

一种用于检测预定区域中特定标识图像的方法，包括：获取预定区域的图像，作为待检目标图像（S100）；对待检目标图像进行二值化处理，获得二值化图像，及二值化图像的取反图像，二者均称为待检目标二值图像（S200）；标定待检目标二值图像的连通域，并将标定的连通域中像素个数满足设定阈值区间的连通域所在区域作为候选连通域区域（S300）；逐一获取候选连通域区域的图像特征，并将候选连通域区域的图像特征与标准特定标识图像的相应图像特征进行比对，将比对结果相同或最接近的候选连通域区域视为特定标识图像所在位置（S400）。该方法能够减小检测过程中的运算量，并降低计算复杂度，进而提高检测速度。同时，还提供一种用于检测预定区域中特定标识图像的装置。

Description

一种用于检测预定区域中特定标识图像的方法及装置 技术领域
本申请涉及计算机应用领域，具体涉及一种获取标准卡片中特定标识的方法和装置。
背景技术
特定标识的识别属于物体检测范畴，特定标识一般用于表明物体的某种特征，对于观察者来说，标识可以用来作为识别物体的参考依据。特定标识物通常具有明显的外形和显著的空间位置，能够更容易的被识别。对于银行卡来说，标识物一般包括发卡组织的图形标识，文字标识等，识别这些标识具有如下重要意义：
(1)标识物的识别可以作为鉴别银行卡真伪的重要参考依据。
标识物一般都具有明确的尺寸规格和相对位置标识，合法的银行卡都会按照标准对标识物进行印刷，而仿冒的银行卡由于做工低廉，成本较低，一般对标识物的印刷控制不严，通过对标志物的识别，得到标识物的尺寸比例和相对位置，进一步通过与真实标准的银行卡比对，可以为鉴别银行卡的真伪提供有效的参考。
(2)标识物的识别可以为银行卡的检测和校正提供参考标识。
相对于整个银行卡而言，标识物会有统一的样式规定，因此一般也就有明确的边缘标识，识别得到标识物的边缘，可以根据此边缘得到当前拍摄图像的仿射变换标识，因此可以对图像进行校正以利于后续的银行卡精确检测和定位。这里描述为边框标识不是特别准确，准确的描述是边缘标识。
(3)标识物的识别可以让服务提供商根据特定的标识物进行精准广告推送服务。
标识物往往表明了银行卡的特定标识，例如发卡组织和使用范围，因此标识物的识别可以让商家针对该银行卡进行精确地广告推送，提高服务的质量，除此之外，获取标识物标识后，结合增强现实等技术，可以实现多媒体的广告展示。
物体检测是指，在给出的一幅图像或一段视频中，找出特定标识或感兴趣 物体在图像或视频中的位置，根据具体的使用场景，物体检测系统通常最后输出物体的外接矩形框或者精确的轮廓。
物体检测大致可分为滑动窗口和广义霍夫变换两类方法。其中，滑动窗口法为目前主流的物体检测方法，滑动窗口法是通过使用训练好的模板在输入图像的多个尺度和多视角进行滑动扫描，通过确定最大相应位置找到目标物体的外接窗口。
现有技术的物体检测方法中所采用的滑动窗口法，在对数据进行分析扫描时，运算量较大，计算复杂度也相对较高，进而使得检测时间较长，其适用于物体结构较为复杂的检测。而对于识别结构较为固定的标识时，显然滑动窗口法并不适用。
因此，迫切需要本领域技术人员解决的技术问题就在于，如何提供一种能够快速、准确的识别出标准卡片上的特定标识的方法。
发明内容
本申请提供一种用于检测预定区域中特定标识图像的方法，以解决现有现有检测过程中计算复杂度高，运算量大的问题。
本申请提供一种用于检测预定区域中特定标识图像的方法，包括：
获取预定区域的图像，作为待检目标图像；
对所述待检目标图像进行二值化处理，获得二值化图像，及所述二值化图像的取反图像，二者均称为待检目标二值图像；
标定所述待检目标二值图像的连通域，并将标定的连通域中像素个数满足设定阈值区间的连通域所在区域作为候选连通域区域；
逐一获取所述候选连通域区域的图像特征，并将该候选连通域区域的图像特征与标准特定标识图像的相应图像特征进行比对，将比对结果相同或最接近的候选连通域区域视为特定标识图像所在位置。
优选地，所述获取所述预定区域或包含该预定区域的图像，作为待检目标图像，包括：
获取所述预定区域或包含该预定区域的图像的预览图像；
计算所述预览图像的区域范围；
将所述预览图像框定在所述区域范围内，获取所述预定区域或包含预定区 域的图像，将该框定的区域范围内的图像作为待检目标图像。
优选地，所述计算预览图像的区域范围是指：根据所述预定区域的实际尺寸参数及预览图像的分辨率参数，获得框定所述预览图像的虚拟边框参数，具体为：
根据所述预定区域或包含该预定区域图像的实际尺寸，判断所述预览图像高度乘以所述预定区域或包含该预定区域图像的实际宽度的值是否大于预览图像宽度乘以所述预定区域或包含该预定区域图像实际高度的值；
若是，则所述虚拟边框的宽度为：预定区域或包含该预定区域图像的实际宽度乘以预览图像高度除以预定区域或包含该预定区域的图像实际高度乘以虚拟边框与所述预览图像的比值；所述虚拟边框的高度为：预览图像高度乘以虚拟边框与所述预览图像的比值；
若否，则虚拟边框的宽度为：预览图像宽度乘以虚拟边框与所述预览图像的比值；虚拟边框的高度为：预定区域或包含该预定区域的图像的实际高度乘以预览图像宽度除以预定区域或包含该预定区域图像的实际宽度乘以虚拟边框与所述预览图像的比值；
所述虚拟边框上边界为显示屏幕高度与虚拟边框高度差的二分之一；所述虚拟边框下边界为显示屏幕高度与虚拟边框高度和的二分之一；所述虚拟边框的左边界为显示屏幕宽度与虚拟边框宽度差的二分之一；所述虚拟边框的右边界为显示屏幕宽度与虚拟边框宽度和的二分之一。
优选地，所述获取所述预定区域或包含该预定区域的图像，作为待检目标图像，包括：根据特定标识图像在所述预定区域或预定区域图像中的特定位置信息，截取包含特定标识图像的区域范围，作为待检目标图像。
优选地，所述特定位置信息为事先设定的特定标识图像所位于预定区域或包含预定区域图像上的位置。
优选地，所述特定位置是位于所述预定区域或包含预定区域的图像的右上角或右下角或左上角或左下角。
优选地，根据特定标识图像在所述预定区域或预定区域图像中的特定位置信息，截取包含特定标识图像的区域范围，作为待检目标图像，包括：
建立预定区域或包含预定区域的标准图像模板；
根据所述预定区域或包含预定区域的图像遍历所述标准图像模板，通过所 述标准图像模板与所述预定区域或包含预定区域的图像的对应关系，获取特定标识图像在预定区域或包含预定区域的图像中的位置信息；
根据该位置信息截取所述预定区域或包含预定区域的图像中包含特定标识的图像区域，作为待检目标图像。
优选地，所述阈值区间的上限值为所述获取的预定区域图像或包含预定区域图像的宽度、高度和上限系数的乘积；下限值为所述获取的预定区域图像或包含预定区域图像的宽度、高度和下限系数的乘积。
优选地，所述将该候选连通域区域的图像特征与所述特定标识图像的相应图像特征进行比对，包括：采用单独图像特征比对或者组合图像特征比对；其中，所述图像特征包括连通域的体积比、连通域外接椭圆长轴方向、连通域外轮廓形状和/或连通域外接矩形长宽比。
优选地，逐一获取所述候选连通域区域的图像特征，并将该候选连通域区域的图像特征与所述特定标识图像的相应图像特征进行比对，将比对结果相同或最接近的候选连通域区域视为所述特定标识图像所在区域，包括：
以所述连通域的体积比和连通域外接椭圆长轴方向作为组合图像特征对候选连通域区域进行筛选，具体是：
获取候选连通域区域的外接边框；
计算所述外接边框内所述候选连通域中前景色的像素与该外界边框面积的体积比；
判断该体积比是否大于预先设定的特定标识图像的阈值范围，若否，则舍弃该候选连通域区域；若是，对所述候选连通域区域进行填充；
计算填充后的所述候选连通域区域外接椭圆长轴的方向，并获得该方向与设定的参考轴的夹角；
判断该夹角是否在预定的阈值范围内，若是，则将该候选连通域区域视为所述特定标识图像所在区域；若否，则舍弃该候选连通域区域。
优选地，包括：判断所述方向与所述参考轴的夹角在预定的阈值范围内后，以所述连通域外轮廓形状作为单独条件对所述候选连通域区域再次进行筛选，具体为：
填充所述候选连通域区域，获得填充图像；
检测所述填充图像的边缘，获得所述填充图像的边缘轮廓图像；
分别获取所述边缘轮廓图像的上下边缘和左右边缘的采样像素点，并选取在设定阈值范围内的所述上下边缘采样像素点和左右边缘采样像素点；
拟合所述选取的采样像素点，获得所述边缘轮廓图像的线段组合信息；
判断所述线段组合信息及交点位置信息与标准特定标识图像的线段组合信息及交点位置信息是否相同；若是，则将候选连通域区域视为特定标识图像区域；若否，则输出未检测特定标识图像区域。
优选地，所述获取所述边缘轮廓图像的上下边缘和左右边缘的采样像素点包括：
以与所述填充图像有两个交点为条件，进行扫描，将该些交点作为采样像素点；
所述选取在设定阈值范围内的所述上下边缘采样像素点和左右边缘采样像素点包括：
判断获取的所述采样像素点的上下边缘之间的距离和左右边缘之间的距离是否在设定的阈值范围内，若是，则选取；若否，则舍弃。
优选地，所述阈值范围为maxDist±maxDist×5％。
优选地，包括：
根据预定区域或包含预定区域图像中特定标识的位置信息和各个交点位置信息，计算仿射变换矩阵；
根据所述仿射变换矩阵获取所述预定区域内的相关信息。
优选地，所述特定标识图像为银行卡联合组织或银行卡发卡组织的标志图形。
本申请另外提供一种用于检测预定区域中特定标识图像的装置，包括：
获取单元，用于获取所述预定区域或包含该预定区域的图像，作为待检目标图像；
二值化处理单元，用于对所述待检目标图像进行二值化处理，获得二值化图像，及所述二值化图像的取反图像，二者均称为待检目标二值图像；
连通域标定单元，用于标定所述待检目标二值图像的连通域，并将标定的连通域中像素个数满足设定阈值区间的连通域作为候选连通域；
图像特征比对单元，用于逐一获取所述候选连通域的图像特征，并将该候选连通域的图像特征与标准特定标识图像的相应图像特征进行比对，将比对结果相同或最接近的候选连通域视为特定标识图像所在位置。
优选地，所述获取单元包括：
预览图像获取单元，用于获取所述预定区域或包含该预定区域的图像的预览图像；
计算单元，用于计算所述预览图像的区域范围；
框定单元，用于将所述预览图像框定在所述区域范围内，获取所述预定区域或包含预定区域的图像，将该框定的区域范围内的图像作为待检目标图像。
优选地，所述计算单元是指：根据所述预定区域的实际尺寸参数及预览图像的分辨率参数，获得框定所述预览图像的虚拟边框参数，具体为：
根据所述预定区域或包含该预定区域图像的实际尺寸，判断所述预览图像高度乘以所述预定区域或包含该预定区域图像的实际宽度的值是否大于预览图像宽度乘以所述预定区域或包含该预定区域图像实际高度的值；
若是，则所述虚拟边框的宽度为：预定区域或包含该预定区域图像的实际宽度乘以预览图像高度除以预定区域或包含该预定区域的图像实际高度乘以虚拟边框与所述预览图像的比值；所述虚拟边框的高度为：预览图像高度乘以虚拟边框与所述预览图像的比值；
若否，则虚拟边框的宽度为：预览图像宽度乘以虚拟边框与所述预览图像的比值；虚拟边框的高度为：预定区域或包含该预定区域的图像的实际高度乘以预览图像宽度除以预定区域或包含该预定区域图像的实际宽度乘以虚拟边框与所述预览图像的比值；
所述虚拟边框上边界为显示屏幕高度与虚拟边框高度差的二分之一；所述虚拟边框下边界为显示屏幕高度与虚拟边框高度和的二分之一；所述虚拟边框的左边界为显示屏幕宽度与虚拟边框宽度差的二分之一；所述虚拟边框的右边界为显示屏幕宽度与虚拟边框宽度和的二分之一。
优选地，所述获取单元包括：
截取单元，用于根据特定标识图像在所述预定区域或预定区域图像中的特 定位置信息，截取包含特定标识图像的区域范围，作为待检目标图像。
优选地，所述特定位置信息为事先设定的特定标识图像所位于预定区域或包含预定区域图像上的位置。
优选地，所述特定位置是位于所述预定区域或包含预定区域的图像的右上角或右下角或左上角或左下角。
优选地，所述截取单元包括：
建立单元，用于建立预定区域或包含预定区域的标准图像模板；
遍历单元，用于根据所述预定区域或包含预定区域的图像遍历所述标准图像模板，通过所述标准图像模板与所述预定区域或包含预定区域的图像的对应关系，获取特定标识图像在预定区域或包含预定区域的图像中的位置信息；
待检目标图像确定单元，用于根据该位置信息截取所述预定区域或包含预定区域的图像中包含特定标识的图像区域，作为待检目标图像。
优选地，所述阈值区间的上限值为所述预定区域或包含预定区域的图像宽度、高度和上限系数的乘积；下限值为所述预定区域或包含预定区域的图像宽度、高度和下限系数的乘积。
优选地，所述将该候选连通域区域的图像特征与所述特定标识图像的相应图像特征进行比对，包括：采用单独图像特征比对或者组合图像特征比对；其中，所述图像特征包括连通域的体积比、连通域外接椭圆长轴方向、连通域外轮廓形状和/或连通域外接矩形长宽比。
优选地，包括：组合图像特征筛选单元，具体是：
外接边框获取单元，用于获取候选连通域区域的外接边框；
体积比计算单元，用于计算所述外接边框内所述候选连通域中前景色的像素与该外界边框面积的体积比；
体积比阈值判断单元，用于判断该体积比是否大于预先设定的特定标识图像的阈值范围；若否，则舍弃该候选连通域区域；若是，对所述候选连通域区域进行填充；
填充计算单元，用于计算填充后的所述候选连通域区域外接椭圆长轴的方向，并获得该方向与设定的参考轴的夹角；
夹角阈值判断单元，用于判断该夹角是否在预定的阈值范围内，若是，则 将该候选连通域区域视为所述特定标识图像所在区域；若否，则舍弃该候选连通域区域。
优选地，单独图像特征筛选单元，包括：
填充单元，用于填充所述候选连通域区域，获得填充图像；
边缘轮廓获取单元，用于根据检测所述填充图像的边缘，获得所述填充图像的边缘轮廓图像；
采样单元，用于分别获取所述边缘轮廓图像的上下边缘和左右边缘的采样像素点，并选取在设定阈值范围内的所述上下边缘采样像素点和左右边缘采样像素点；
采样像素点拟合单元，用于分别获取所述边缘轮廓图像的上下边缘和左右边缘的采样像素点，并选取在设定阈值范围内的所述上下边缘采样像素点和左右边缘采样像素点；
线段组合判断单元，用于判断所述线段组合信息及交点位置信息与标准特定标识图像的线段组合信息和交点位置信息是否相同；若是，则将候选连通域区域视为特定标识图像区域；若否，则输出未检测特定标识图像区域。
优选地，所述采样单元包括：
扫描单元，以与所述填充图像有两个交点为条件，进行扫描，将该些交点作为采样像素点；
距离阈值判断单元，用于判断获取的所述采样像素点的上下边缘之间的距离和左右边缘之间的距离是否在设定的阈值范围内，若是，则选取；若否，则舍弃。
优选地，所述阈值范围为maxDist±maxDist×5％。
优选地，包括：
仿射变换单元，用于根据预定区域或包含预定区域图像中特定标识的位置信息和各个交点位置信息，计算仿射变换矩阵；
相关信息获取单元，用于根据所述仿射变换矩阵获取相关信息。
优选地，所述特定标识图像为银行卡联合组织或银行卡发卡组织的标志图形。
与现有技术相比，本申请具有以下优点:
本申请通过获取预定区域的图像，作为待检目标图像，对该待检目标图像进行二值化处理，并标定二值化图像中的连通域作为候选连通域图像，在该些候选连通域图像中筛选出符合特定标识图像特征的连通域，进而确定检测到特定标识信息。由于本申请仅是对预定区域或者包含预定区域图像中的候选连通域图像的图像特征与标准特定标识图像特征进行比较，因此，在对数据进行扫描分析时，是在较小的区域内进行，所以该方式能够减小检测过程中的运算量，并且降低计算复杂度，能够根据特征比较快速完成检测。
附图说明
图1是本申请提供的一种用于检测预定区域中特定标识图像的方法流程图；
图2是本申请提供的一种用于检测预定区域中特定标识图像的方法中虚拟边框计算示意图；
图3是本申请提供的一种用于检测预定区域中特定标识图像的方法中虚拟边缘的示意图；
图4是本申请提供的一种用于检测预定区域中特定标识图像的方法中截取包含特定标识图像的区域范围的第一实施例示意图；
图5是本申请提供的一种用于检测预定区域中特定标识图像的方法中截取包含特定标识图像的区域范围的第二实施例流程图；
图6是本申请提供的一种用于检测预定区域中特定标识图像的方法中对待检目标图像进行二值化处理后的二值化图像示意图；
图7是本申请提供的一种用于检测预定区域中特定标识图像的方法中，对候选连通域区域进行第一次筛选的流程图；
图8是本申请提供的一种用于检测预定区域中特定标识图像的方法中，对候选联通区域进行第二次筛选的流程图；
图9是本申请提供的一种用于检测预定区域中特定标识图像的方法中，对银联标识上置和下置的检测示意图；
图10是本申请提供的一种用于检测预定区域中特定标识图像的装置结构框图。
具体实施方式
在下面的描述中阐述了很多具体细节以便于充分理解本申请。但是本申请 能够以很多不同于在此描述的其它方式来实施，本领域技术人员可以在不违背本申请内涵的情况下做类似推广，因此本申请不受下面公开的具体实施的限制。
本申请以银行卡卡面所在区域为预定区域，银行卡上的银联标识为特定标识，说明本申请的具体实现过程，但是应当说明，本申请的方法并不局限于对银行卡上的银联标识的识别，本申请的方法可以用于所有具有预定区域中特定已知标识的识别场景，特定已知标识可以是产品标识，企业logo等等，产品标识的预定区域可以是产品本身；企业logo的预定区域可以是包含有该logo的任何载体本身。
请参考图1所示，图1是本申请提供一种用于检测预定区域中特定标识图像方法的第一实施例流程图，在该实施例中，包括如下步骤：
步骤S100：获取预定区域的图像，作为待检目标图像；
步骤S200：对所述待检目标图像进行二值化处理，获得二值化图像，及所述二值化图像的取反图像，二者均称为待检目标二值图像；
步骤S300：标定所述待检目标二值图像的连通域，并将标定的连通域中像素个数满足设定阈值区间的连通域所在区域作为候选连通域区域；
步骤S400：逐一获取所述候选连通域区域的图像特征，并将该候选连通域区域的图像特征与标准特定标识图像的相应图像特征进行比对，将比对结果相同或最接近的候选连通域区域视为特定标识图像所在位置。
通过上述步骤，在对数据进行扫描分析时，是在较小的区域内进行，所以上述检测方法能够减小检测过程中的运算量，并且降低计算复杂度，能够根据特征比较快速完成检测。
下面对上述步骤的具体实现过程进行说明：
所述步骤S100：获取预定区域的图像，作为待检目标图像。
该步骤中的预定区域可以是银行卡卡面区域，也就是说获取银行卡卡面图像或者包含有银行卡卡面区域的图像，获取方式可以通过采用移动终端人为获取或者自动获取，例如：通过移动终端的摄像功能获取预定区域或包含该预定区域的图像。通过预定区域的选取可以减小检测扫描范围，提高检测速度的同时，降低运算量。
为更进一步提高检测效率，减少对非银联标识区域的检测，本申请可以在获取所述银行卡图像或者包含有银行卡区域的图像时，先通过获取的银行卡图 像或者包含银行卡卡面区域图像的预览图像，并根据标准银行卡的实际尺寸及预览图像的分辨率，获得框定所述预览图像的虚拟边框的长和宽；再根据该虚拟边框的长宽计算出虚拟边框的上下左右边界，具体如下：
参考图2所示，图2是本申请提供的一种用于检测预定区域中特定标识图像的方法中虚拟边框计算示意图。根据《银联卡业务运作规章》第三卷《卡片BIN号及标识规则》，可以获知银行卡卡片的标准尺寸参数，卡片宽度是85.60mm±0.30mm，卡片高度是53.98mm±0.30mm，本申请中忽略卡片误差范围0.30mm，即：卡片宽度CardWidthGdt＝85.6，卡片高度CardHeightGdt＝53.98；假定移动终端屏幕分辨率为长×高，即：ScreenWidth×PreviewHeight；预览图像的分辨率参数为宽×高，即：PreviewWidth×PreviewHeight；虚拟边框与预览图像的比值参数为DisplayRatio；在计算虚拟边框时，首先判断PreviewHeight(预览图像高度)×CardWidthGdt(卡片宽度)是否大于(不包含等于)PreviewWidth(预览图像宽度)×CardHeightGdt(卡片高度)，若是，则：
ScaledWidth＝CardWidthGdt×PreviewHeight/CardHeightGdt×DisplayRatio；
ScaledHeight＝PreviewHeight×DisplayRatio；
若否，则：
ScaledHeight＝CardHeightGdt×PreviewWidth/CardWidthGdt×DisplayRatio；
ScaledWidth＝PreviewWidth×DisplayRatio；
当计算得到虚拟边框的宽与高之后，可以根据移动终端屏幕的分辨率参数确定所述虚拟边框的上下左右边界，具体如下：
上边界Top＝ScreenHeight/2-ScaledHeight/2；
下边界Bottom＝ScreenHeight/2+ScaledHeight/2；
左边界Left＝ScreenWidth/2-ScaledWidth/2；
右边界Right＝ScreenWidth/2+ScaledWidth/2。
参考图3，图是3本申请提供的一种用于检测预定区域中特定标识图像的方法中虚拟边缘的示意图。
所述虚拟边框覆盖在所述银行卡图像或者包含银行卡图像区域中银行卡的上方，将虚拟边框所覆盖的范围内区域作为待检目标图像，从而达到检测范围减小化。也可以在虚拟边框的覆盖范围内根据标准银行中银联标识的特定位置 信息，截取包含银联标识图像的区域范围，将该区域范围作为待检目标图像，从而进一步达到检测范围的更小化，以提高检测速度，减少运算量。
截取包含银联标识图像的区域范围，本申请通过两种实现方式进行说明。
实现方式一：
参考图4所示，图4是本申请提供的一种用于检测预定区域中特定标识图像的方法中截取包含特定标识图像的区域范围的第一实施例示意图。根据《银联卡业务运作规章》第三卷《卡片BIN号及标识规则》可获知，银行卡上银联标识的位置信息是事先设定的位置，银联标识区位于卡片正面右侧，分为下置格式和上置格式两种。
当银联标识采用下置格式时，须位于卡片的右下角，具体要求是：银联标识的宽度22mm，高度15mm，银联标识右边沿到卡的右边沿的距离是2mm，银联标识下边沿到卡的下边沿的距离是2mm。
当银联标识采用上置格式时，须位于卡片的右上角，具体要求是：银联标识的宽度22mm，高度15mm，银联标识右边沿到卡的右边沿的距离是2mm，银联标识上边沿到卡的上边沿的距离是2mm。
根据上述对银联标识的位置，在截取银行卡图像时，可以仅对银行卡图像或者包含银行卡区域的图像的右上角和右下角进行处理，例如：假定银行卡图像或者包含银行卡区域的图像的分辨率(宽×高)为：Width×Height，则需要截取的图像分别为[Width/2，Width]×[0，Height]和[Width/2，Width]×[Height/2，Height]。
实现方式二：
请参考图5所示，图5提供的一种用于检测预定区域中特定标识图像的方法中截取包含特定标识图像的区域范围的第二实施例流程图；包括：
步骤S101：建立银行卡图像的标准图像模板；
步骤S102：根据所述银行卡图像或包含银行卡区域的图像遍历所述标准图像模板，通过所述标准图像模板与所述银行卡图像或包含银行卡区域的图像的对应关系，获取银联标识图像在所述银行卡图像或包含银行卡区域图像中的位置信息；
步骤S103：根据该位置信息截取包含有银联标识图像的区域范围，将该截取的银联标识图像的区域范围作为待检目标图像。
可以理解的是，本申请也可以在该银行卡图像或包含有银行卡区域的图像的基础上，根据银联标识的特定位置信息或者根据银行卡标准图像模板，直接截取包含银联标识的图像区域，将该图像区域作为待检目标图像；也可以在虚拟边框的基础上仅截取所述虚拟边框范围内包含银联标识图像区域的部分，舍弃虚拟边框以外的图像信息，进而达到提高检测的效率。
所述步骤S200：对所述待检目标图像进行二值化处理，获得二值化图像，及所述二值化图像的取反图像，二者均称为待检目标二值图像。
参考图6所示，图6是本申请提供的一种用于检测预定区域中特定标识图像的方法中对待检目标图像进行二值化处理后的二值化图像示意图；
在该步骤中，将待检目标图像，也就是对截取的包含有银联标识图像的区域范围的图像进行二值化处理。图像二值化处理，是指将图像上的像素点的灰度值设置为0或255，也就是将整个图像呈现出明显的只有黑和白的视觉效果；即将256个亮度等级的灰度图像通过适当的阈值选取而获得仍然可以反映图像整体和局部特征的二值化图像。要进行二值图像的处理与分析，首先要把原图像灰度化，之后对灰度图像进行二值化，得到二值化图像，进而有利于再对图像做进一步处理时，图像的集合性质只与像素值为0或255的点的位置有关，不再涉及像素的多级值，使处理变得简单，而且数据的处理和压缩量减小。
在该步骤中二值化处理过程中阈值的选取可以采用Otsu方法，将灰度图像分割成背景和目标(前景)两部分。所述Otsu方法又称为最大类间方差法，是一种自适应的阈值确定的方法，其按照图像的灰度特定，将图像分割成背景与目标两部分，背景和目标之间的类间方差越大，说明构成图像背景和目标差别越大。通过前景像素点数占图像的比例和平均灰度值；背景像素点数占图像的比例和平均灰度值，计算出前景和背景图像的方差，选取方差最大时所对应的灰度值即为全局阈值，根据该全局阈值对图像进行二值化变换。
由于在对二值化图像分析过程中，是要对二值化图像中白色像素(目标像素)进行标记，而包含有银联标识的图像作为待检测目标图像被二值化处理之后，银联标识可能会被二值化为前景(1)或背景(0)，因此，在得到二值化图像后，还可以再对该二值化图像进行取反，将二值化图像和取反图像都作为待检目标二值图像。
二值图像分析最重要的方法就是连通区域标记，它是所有二值图像分析的 基础，它通过对二值图像中白色像素(目标)的标记，让每个单独的连通区域形成一个被标识的块，进一步的我们就可以获取这些块的轮廓、外接矩形、质心、不变矩等几何参数。
二值图像的连通域标记处理操作就是从白色像素(通常二值图像用“1”)和黑色像素(通常用“0”来表示)组合的一幅点阵图像中，将互相邻接(4邻接或8邻接)的具有像素值“1”的像素集合提取出来。二值图像中包含黑(像素值为0)白(像素值为1或255)两种颜色，分别作为目标颜色和背景颜色，标记算法只对目标像素进行标记。
所述步骤S300：标定所述待检目标二值图像的连通域，并将标定的连通域中像素个数满足设定阈值区间的连通域作为候选连通域。
连通域标定是所有二值图像分析的基础，二值图像的连通域标记处理操作就是从白色像素(通常二值图像用“1”)和黑色像素(通常用“0”来表示)组合的一幅点阵图像中，将互相邻接的具有像素值“1”的像素集合提取出来，也就是说，二值图像中包含黑(像素值为0)白(像素值为1或255)两种颜色，分别作为目标颜色和背景颜色，在连通域标定时，标定算法只对目标像素(白色像素)进行标定。
连通域标定算法中，常见的邻接关系有两种：4邻接与8邻接。4邻接一共4个点，即上、下、左和右；8邻接的点一共有8个，包括了对角线位置的点，即：上、下、左、右、左上和右上。将上、下、左、右或者上、下、左、右、左上、右上中彼此连通的点所构成集合标定为连通域。
该步骤将截取的包含有银联标识图像的区域范围经过二值化后，对二值化图像采用上述连通域标定方式，标定二值化图像中的连通域，也就是说，标定像素值为1的所有连通域。对所标定的每个独立的连通域进行阈值判断，若该连通域的像素个数在给定的阈值区间内，则保留该连通域，否则对待检目标二值图像中的取反图像进行连通域标定，并对取反图像中所标定的连通域像素个数进行阈值判断，若连通域像素个数在阈值区间内，则保留该连通域，若否则舍弃该连通域。
需要说明的是，对于银联标识图像所形成的连通域，其连通域的像素个数会在一个合理范围内，该合理范围可以设定属于该银联标识图像连通域的阈值区间。
上述银联标识连通域的阈值区间具体计算可以参考如下：
阈值下限LowThreshhold＝Width×Height×下限系数；
阈值上限HighThreshhold＝Width×Height×上限系数；
其中，Width和Height分别为所获取的预定区域图像或包含预定区域的图像的宽和高，也就是说，Width和Height可以是待检目标图像的宽和高，下限系数和上限取值可以为经验值，分别为0.011和0.043。
当检测的特定标识信息非银联标识时，也可以根据该特定标识连通域的像素个数所处于的合理范围，来设定一个阈值区间。
所述步骤S400：逐一获取所述候选连通域区域的图像特征，并将该候选连通域区域的图像特征与标准特定标识图像的相应图像特征进行比对，将比对结构相同或者接近的候选连通域区域视为特定标识图像所在位置。
该步骤的具体实现过程是，对于候选的银联标识连通域区域的图像特征和标准银联标识图像，将二者的相应的图像特征进行比对，具体可以采用单独图像特征比对或者组合图像特征比对；其中，所述图像特征包括：包括连通域的体积比、连通域外接椭圆长轴方向、连通域外轮廓形状和/或连通域外接矩形长宽比等，也可以理解的是上述特征属于图像的形状特征，而形状特征还可以包括距离特征、图像凹凸特征、长短特征等等。
该步骤中，可以以所述连通域的体积比和连通域外界椭圆长轴方向作为第一组合图像特征，对候选银联标识图像的连通域区域进行筛选。
请参考图7所示，图7是本申请提供一种用于检测预定区域中特定标识图像的方法中对候选连通域区域进行第一次筛选的流程图。
具体步骤包括：
步骤S701：获取候选连通域区域的外接边框；
在该步骤中可以通过扫描的方式，从候选银联标识连通域图像的最外侧边缘向内扫描，检测到前景色(白色像素)后，停止扫描，获得外接边框，该外接边框为矩形。
步骤S702：计算所述外界边框内所述候选连通域中前景色的像素与该外界边框面积的体积比；
该步骤可以通过统计矩形外接边框内前景色(白色)的像素个数，用白色像素个数除以矩形外接边框的面积，得到连通域体积比。
步骤S703：判断该体积比是否大于预先设定的特定标识图像的阈值范围， 若否，则舍弃该候选连通域区域；若是，对所述候选连通域区域进行填充；
在该步骤中，可以根据事先设定标准银联标识连通域体积比的阈值Tsolid，判断候选银联标识连通域体积比是否与标准的银联标识连通域体积比的图像特征相符。其中阈值Tsolid可以通过标准银联标识图像的信息获得，例如：从各个角度获取标准银联标识图像，并计算其与外接矩阵的体积比，从而选取阈值Tsolid，通过候选银联标识连通域体积比与Tsolid的比较，确定候选连通域是否为银联标识图像。
当候选银联标识连通域体积比大于阈值Tsolid，该候选银联标识连通域为银联标识连通域，因为，在候选银联标识连通域中可能存在与标准银联标识连通域体积比图像特征相同的非银联标识连通域，所以，本申请在此基础上还要结合另一图像特征，即：连通域椭圆长轴方向的图像特征，进一步确定该候选银联标识。当确定候选银联标识连通域体积比大于阈值Tsolid之后，进入连通域椭圆长轴方向的图像特征判断，如下述步骤S704。
步骤S704：对所述候选连通域区域进行填充；
该步骤中，对银联标识候选连通域区域以前景色(白色)进行填充。
步骤S705：计算填充后的所述银联标识候选连通域区域的方向；
在该步骤中，可以通过采用椭圆拟合的方式，计算出连通域的方向。在该实施例中，连通域的方向以在所述连通域上获得的外接椭圆的长轴方向为连通域方向。所述椭圆拟合，是通过对候选连通域进行边缘检测，记录边缘上的点，对该些点进行椭圆拟合。
当计算出椭圆拟合长轴方向之后，以连通域的左下角为原点建立坐标轴，该坐标轴的X轴定义为参考轴，进而椭圆长轴与X轴之间形成夹角。同样地，对于标准银联标识图像，也可以计算出椭圆长轴与X轴之间形成的夹角。
步骤S706：判断该夹角是否在预定的阈值范围内，若是，则将该银联标识的候选连通域区域视为所述银联标识图像所在区域；若否，则舍弃该银联标识的候选连通域区域；
该步骤中，夹角的阈值范围可以根据上述对标准银联标识图像进行椭圆拟合，并计算该椭圆长轴方向与参考轴之间的夹角，因为，在获取银联标识待检目标图像，可能存在获取角度等因素的影响，使得候选连通域图像角度也会受到影响，所以本申请基于标准银联标识中椭圆长轴与参考轴之间的夹角，设定 夹角的阈值范围，以避免由于获取角度的问题而影响判断，此处的参考轴也为X轴。上述的参考轴的设定可以根据实际判断需要设定，即：也可以为Y轴，能够满足连通域外接椭圆长轴方向特征判断需求即可。
需要说明的是，当特定标识为存在两个明显的外接边缘，即：内层边缘和外层边缘连通域时；还可以在上述基础上再次对银联标识候选连通域图像进行二次体积比判断，但该二次体积比判断时，其阈值采用的并非上述Tsolid，而是针对此次判断设定的另一阈值Tsolid’，检测判断过程如上述步骤S701-步骤S706。需要说明的是，内层连通域的像素值与外层连通域的像素值是不同的，所以采用阈值也不同。
经过上述步骤已经可以确定银联标识候选连通域图像为银联标识图像，为更进一步确定，本申请在上述步骤基础上判断所述椭圆长轴与所述参考轴的夹角在预定的阈值范围内的基础上，以所述连通域外轮廓形状作为单独条件对所述候选连通域区域进行第二次筛选。
具体参考图8所示，图8是本申请提供的一种用于检测预定区域中特定标识图像的方法中，以连通域外轮廓形状作为单独条件对所述候选连通域区域进行第二次筛选的流程图。
具体包括如下步骤：
步骤S801：填充所述候选连通域区域，获得填充图像；
该步骤可以采用步骤S704中的填充方式，对候选连通域区域进行填充。
步骤S802：检测所述填充图像的边缘，获得所述填充图像的边缘轮廓图像；
该步骤具体实现，当候选连通域区域以白色进行填充后，白色区域为候选连通域区域，通过采用边缘检测方法，获得填充图像的边缘轮廓图像。所述边缘检测方法是图像处理中常用的一种技术，其可以捕获图像中物体的重要特征，常用的边缘检测方法有：查分边缘检测、梯度边缘检测、Roberts边缘检测算子、Sobel边缘检测算子、Prewitt边缘检测算子和Laplace边缘检测算子等等。本申请中采用Sobel边缘检测算子检测所述填充图像的边缘，所述Sobel边缘检测算子主要是根据像素点上下、左右邻点灰度加权差，在边缘处达到极值现象检测边缘，其对噪声具有平滑作用，能够提供较为精确的边缘方向信息。Sobel算子有两个，一个是检测水平边缘的；另一个是检测垂直边缘的。与Prewitt算子相比，Sobel算子对于象素的位置的影响做了加权，可以降低边缘模糊程度，因 此效果更好。当然，也可以根据不同的需求采用其他边缘检测算子来检测填充图像的边缘。
步骤S803：分别获取所述边缘轮廓图像的上下边缘和左右边缘的采样像素点，并选取在设定阈值范围内的所述上下边缘采样像素点和左右边缘采样像素点；
在该步骤中，采用扫描的方式采样边缘轮廓图像中的上下边缘和左右边缘的采样像素点。在以银联标识为实施例的情况下，填充图像的边缘轮廓为矩形图像，分别对矩形图像的上下边缘和左右边缘分别进行扫面，对上下边缘和左右边缘的扫描方式可以采用相同方式，下面以上下边缘举例说明。
对矩形图像的边缘进行逐列扫描，选取扫描线与所述边缘图像有两个交点的像素点，并对该些像素点之间的距离进行计算，统计所有像素点之间的最大值maxDist的像素点作为采样像素点。由于矩形图像的边缘可能存在锯齿，因此，设定最大距离的阈值范围对采样像素点进行选取，阈值范围可以设定在maxDist±maxDist×5％，选取在该阈值范围内的像素点作为采样像素点，该阈值范围以外的像素点则舍弃。从而得到矩形图像上下边缘的采样像素点，同理，可以获得左右边缘的采样像素点。
步骤S804：拟合所述选取的采样像素点，获得所述边缘轮廓图像的线段组合信息。
通过步骤S803选取的候选连通域矩形图像四条边的采样像素点，步骤S604可以采用多项式曲线拟合方法，对所选取的采样像素点进行拟合，从而得到四条线段，并通过该四条线段组合可以获得四个顶点；步骤S804的具体实现方式是针对银行卡银联标识进行举例说明，可以理解的是，如果其他标识可以采用不同的拟合方式或者计算方法，进行采样像素点的拟合。
步骤S805：判断所述线段组合信息与标准特定标识图像的线段组合信息是否相同，若是，则将候选连通域区域视为特定标识图像区域；若否，则输出未检测特定标识图像区域。
该步骤中，对于在步骤S804中获得的线段组合，即：四条线段以及根据该四条线段的组合获得四个顶点，将该四条线段与四个顶点与标准银行卡银联标识的图像进行比较，相同则该银联标识候选连通域区域即为银联标识图像。
需要说明的是，经过以上步骤，候选银联标识被检测到，系统可以输出银联标识的四个角点(参考图9)。同时，若检测得到的组合信息，例如：顶点信息多于或少于4个，则可以判断该候选区域非银联图标。若检测到的银联标识候选区域为空，则提示用户未检测到银联标识，若检测到的银联标识候选区域个数多于一个，则将所有候选区域提供给用户，用于用户自主选择。
上述是以银联标识为例对检测特定标识进行的说明，对于银联标识而言，其具有的图像特征是形状特征，包括有平行四边形、四个顶点等属性；当特定标识为文字或产品标识或企业logo等时，可以通过其本申请具有的图像特征进行检测，例如：各个发卡银行的logo，以招商银行的logo为例进行说明，请参考图3，招商银行的logo图像的图像特征可以是圆形、字母M以及字母M的各个顶点、三角形或者是凹凸属性等等，可以通过该些特征属性采用上述检测方法，对该logo进行检测。
由此可见，本申请提供的检测特定标识的方法并不限于上述银联标识。
当检测到银联标识后可以根据检测到的银联标识的四个顶点信息以及标准银联信息的顶点信息，计算银联标识图像的仿射变换矩阵；根据该矩阵可以计算获得银行卡上其他信息。仿射变换矩阵可以采用如下方式获得：
假定标准的银联标识的四个顶点坐标为[x1,y1]，[x2,y2]，[x3,y3]和[x4,y4]，而在检测到候选连通域图像中的4个顶点的坐标为[X1,Y1]，[X2,Y2]，[X3,Y3]和[X4,Y4]，根据4个点的对应关系就可以计算仿射变换(projective)矩阵M，对于检测图像的每个像素点可以计算射影变换后的图像，此图像所在的视觉就是正视角图像。根据银行卡中的规定，可以获知银行卡上任何信息，例如：需要获知银行卡上的卡号信息时，将卡号的像素点乘以矩阵M即可以获得卡号的正是正视角图像，从而的准确的卡号信息。
以上公开了本申请一种用于检测预定区域中特定标识图像的方法实施例，与所述方法实施例相对应，本申请还公开了一种用于检测预定区域中特定标识图像装置实施例，请参看图10，其为本申请一种用于检测预定区域中特定标识图像的装置的实施例结构示意图。由于装置实施例基本相似于方法实施例，所以描述得比较简单，相关之处参见方法实施例的部分说明即可。下述描述的装置实施例仅仅是示意性的。
如图10所示，本申请提供的一种用于检测预定区域中特定标识图像的装置，包括：
获取单元A100，用于获取所述预定区域或包含该预定区域的图像，作为待检目标图像；
二值化处理单元A200，用于对所述待检目标图像进行二值化处理，获得二值化图像，及所述二值化图像的取反图像，二者均称为待检目标二值图像；
连通域标定单元A300，用于标定所述待检目标二值图像的连通域，并将标定的连通域中像素个数满足设定阈值区间的连通域作为候选连通域；
图像特征比对单元A400，用于逐一获取所述候选连通域的图像特征，并将该候选连通域的图像特征与标准特定标识图像的相应图像特征进行比对，将比对结果相同或最接近的候选连通域视为特定标识图像所在位置。
其中，该获取单元A100包括：
预览图像获取单元，用于获取所述预定区域或包含该预定区域的图像的预览图像；
计算单元，用于计算所述预览图像的区域范围；该计算单元是指，根据所述预定区域的实际尺寸参数及预览图像的分辨率参数，获得框定所述预览图像的虚拟边框参数，具体为：
根据所述预定区域或包含该预定区域图像的实际尺寸，判断所述预览图像高度乘以所述预定区域或包含该预定区域图像的实际宽度的值是否大于预览图像宽度乘以所述预定区域或包含该预定区域图像实际高度的值；
若是，则所述虚拟边框的宽度为：预定区域或包含该预定区域图像的实际宽度乘以预览图像高度除以预定区域或包含该预定区域的图像实际高度乘以虚拟边框与所述预览图像的比值；所述虚拟边框的高度为：预览图像高度乘以虚拟边框与所述预览图像的比值；
若否，则虚拟边框的宽度为：预览图像宽度乘以虚拟边框与所述预览图像的比值；虚拟边框的高度为：预定区域或包含该预定区域的图像的实际高度乘以预览图像宽度除以预定区域或包含该预定区域图像的实际宽度乘以虚拟边框与所述预览图像的比值；
所述虚拟边框上边界为显示屏幕高度与虚拟边框高度差的二分之一；所述虚拟边框下边界为显示屏幕高度与虚拟边框高度和的二分之一；所述虚拟边框的左边界为显示屏幕宽度与虚拟边框宽度差的二分之一；所述虚拟边框的右边界为显示屏幕宽度与虚拟边框宽度和的二分之一。
框定单元，用于将所述预览图像框定在所述区域范围内，获取所述预定区域或包含预定区域的图像，将该框定的区域范围内的图像作为待检目标图像。
所述获取单元A100还包括：
截取单元，用于根据特定标识图像在所述预定区域或预定区域图像中的特定位置信息，截取包含特定标识图像的区域范围，作为待检目标图像。其中，截取单元中的特定位置信息为事先设定的特定标识图像所位于预定区域或包含预定区域图像上的位置。在以银行卡的银联标识为特定标识的情况下，所述特定位置可以根据《银联卡业务运作规章》第三卷《卡片BIN号及标识规则》中的记载可以获知，银联标识的特定位置位于银行卡卡面区域的右上角或者右下角。可以理解的是，当特定标识信息非银联标识，还有可以位于左上角或左下角等。
截取单元还可以通过下述方式实现：
建立单元，用于建立预定区域或包含预定区域的标准图像模板；
遍历单元，用于根据所述预定区域或包含预定区域的图像遍历所述标准图像模板，通过所述标准图像模板与所述预定区域或包含预定区域的图像的对应关系，获取特定标识图像在预定区域或包含预定区域的图像中的位置信息；
待检目标图像确定单元，用于根据该位置信息截取所述预定区域或包含预定区域的图像中包含特定标识的图像区域，作为待检目标图像。
连通域标定单元A300中的阈值区间的上限值为所述预定区域或包含预定区域的图像宽度、高度和上限系数的乘积；下限值为所述预定区域或包含预定区域的图像宽度、高度和下限系数的乘积。
图像特征比对单元A400，包括组合图像特征筛选单元或单独图像特征筛选单元，通过采用单独图像特征比对或者组合图像特征比对；其中，所述图像特征包括连通域的体积比、连通域外接椭圆长轴方向、连通域外轮廓形状和/或连通域外接矩形长宽比。
组合图像特征筛选单元，具体是：
外接边框获取单元，用于获取候选连通域区域的外接边框；
体积比计算单元，用于计算所述外接边框内所述候选连通域中前景色的像素与该外界边框面积的体积比；
体积比阈值判断单元，用于判断该体积比是否大于预先设定的特定标识图 像的阈值范围；若否，则舍弃该候选连通域区域；若是，对所述候选连通域区域进行填充；
填充计算单元，用于计算填充后的所述候选连通域区域外接椭圆长轴的方向，并获得该方向与设定的参考轴的夹角；
夹角阈值判断单元，用于判断该夹角是否在预定的阈值范围内，若是，则将该候选连通域区域视为所述特定标识图像所在区域；若否，则舍弃该候选连通域区域。
单独图像特征筛选单元，包括：
填充单元，用于填充所述候选连通域区域，获得填充图像；
边缘轮廓获取单元，用于根据检测所述填充图像的边缘，获得所述填充图像的边缘轮廓图像；
采样单元，用于分别获取所述边缘轮廓图像的上下边缘和左右边缘的采样像素点，并选取在设定阈值范围内的所述上下边缘采样像素点和左右边缘采样像素点；
采样像素点拟合单元，用于分别获取所述边缘轮廓图像的上下边缘和左右边缘的采样像素点，并选取在设定阈值范围内的所述上下边缘采样像素点和左右边缘采样像素点；
线段组合判断单元，用于判断所述线段组合信息及交点位置信息与标准特定标识图像的线段组合信息和交点位置信息是否相同；若是，则将候选连通域区域视为特定标识图像区域；若否，则输出未检测特定标识图像区域。
所述采样单元包括：
扫描单元，以与所述填充图像有两个交点为条件，进行扫描，将该些交点作为采样像素点；
距离阈值判断单元，用于判断获取的所述采样像素点的上下边缘之间的距离和左右边缘之间的距离是否在设定的阈值范围内，若是，则选取；若否，则舍弃。在该距离阈值判断单元中，阈值范围可以是maxDist±maxDist×5％，其中maxDist为两个采样像素点的最大距离。
本申请还包括仿射变换单元，用于根据预定区域或包含预定区域图像中特定标识的位置信息和各个交点位置信息，计算仿射变换矩阵；以及相关信息获取单元，用于根据所述仿射变换矩阵获取相关信息。
本申请虽然以较佳实施例公开如上，但其并不是用来限定本申请，任何本领域技术人员在不脱离本申请的精神和范围内，都可以做出可能的变动和修改，因此本申请的保护范围应当以本申请权利要求所界定的范围为准。
在一个典型的配置中，计算设备包括一个或多个处理器(CPU)、输入/输出接口、网络接口和内存。
内存可能包括计算机可读介质中的非永久性存储器，随机存取存储器(RAM)和/或非易失性内存等形式，如只读存储器(ROM)或闪存(flash RAM)。内存是计算机可读介质的示例。
1、计算机可读介质包括永久性和非永久性、可移动和非可移动媒体可以由任何方法或技术来实现信息存储。信息可以是计算机可读指令、数据结构、程序的模块或其他数据。计算机的存储介质的例子包括，但不限于相变内存(PRAM)、静态随机存取存储器(SRAM)、动态随机存取存储器(DRAM)、其他类型的随机存取存储器(RAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、快闪记忆体或其他内存技术、只读光盘只读存储器(CD-ROM)、数字多功能光盘(DVD)或其他光学存储、磁盒式磁带，磁带磁磁盘存储或其他磁性存储设备或任何其他非传输介质，可用于存储可以被计算设备访问的信息。按照本文中的界定，计算机可读介质不包括非暂存电脑可读媒体(transitory media)，如调制的数据信号和载波。
2、本领域技术人员应明白，本申请的实施例可提供为方法、系统或计算机程序产品。因此，本申请可采用完全硬件实施例、完全软件实施例或结合软件和硬件方面的实施例的形式。而且，本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。

Claims (30)

  1. 一种用于检测预定区域中特定标识图像的方法，其特征在于，包括：
    获取预定区域的图像，作为待检目标图像；
    对所述待检目标图像进行二值化处理，获得二值化图像，及所述二值化图像的取反图像，二者均称为待检目标二值图像；
    标定所述待检目标二值图像的连通域，并将标定的连通域中像素个数满足设定阈值区间的连通域所在区域作为候选连通域区域；
    逐一获取所述候选连通域区域的图像特征，并将该候选连通域区域的图像特征与标准特定标识图像的相应图像特征进行比对，将比对结果相同或最接近的候选连通域区域视为特定标识图像所在位置。
  2. 根据权利要求1所述的用于检测预定区域中特定标识图像的方法，其特征在于，所述获取所述预定区域图像，作为待检目标图像，包括：
    获取所述预定区域或包含该预定区域的图像的预览图像；
    计算所述预览图像的区域范围；
    将所述预览图像框定在所述区域范围内，获取所述预定区域或包含预定区域的图像，将该框定的区域范围内的图像作为待检目标图像。
  3. 根据权利要求2所述的用于检测预定区域中特定标识图像的方法，其特征在于，所述计算预览图像的区域范围是指：根据所述预定区域的实际尺寸参数及预览图像的分辨率参数，获得框定所述预览图像的虚拟边框参数，具体为：
    根据所述预定区域或包含该预定区域图像的实际尺寸，判断所述预览图像高度乘以所述预定区域或包含该预定区域图像的实际宽度的值是否大于预览图像宽度乘以所述预定区域或包含该预定区域图像实际高度的值；
    若是，则所述虚拟边框的宽度为：预定区域或包含该预定区域图像的实际宽度乘以预览图像高度除以预定区域或包含该预定区域的图像实际高度乘以虚拟边框与所述预览图像的比值；所述虚拟边框的高度为：预览图像高度乘以虚拟边框与所述预览图像的比值；
    若否，则虚拟边框的宽度为：预览图像宽度乘以虚拟边框与所述预览图像的比值；虚拟边框的高度为：预定区域或包含该预定区域的图像的实际高度乘以预览图像宽度除以预定区域或包含该预定区域图像的实际宽度乘以虚拟边框 与所述预览图像的比值；
    所述虚拟边框上边界为显示屏幕高度与虚拟边框高度差的二分之一；所述虚拟边框下边界为显示屏幕高度与虚拟边框高度和的二分之一；所述虚拟边框的左边界为显示屏幕宽度与虚拟边框宽度差的二分之一；所述虚拟边框的右边界为显示屏幕宽度与虚拟边框宽度和的二分之一。
  4. 根据权利要求1-3任意一项所述的用于检测预定区域中特定标识图像的方法，其特征在于，所述获取所述预定区域或包含该预定区域的图像，作为待检目标图像，包括：
    根据特定标识图像在所述预定区域或预定区域图像中的特定位置信息，截取包含特定标识图像的区域范围，作为待检目标图像。
  5. 根据权利要求4所述的用于检测预定区域中特定标识图像的方法，其特征在于，所述特定位置信息为事先设定的特定标识图像所位于预定区域或包含预定区域图像上的位置。
  6. 根据权利要求5所述的用于检测预定区域中特定标识图像的方法，其特征在于：所述特定位置是位于所述预定区域或包含预定区域的图像的右上角或右下角或左上角或左下角。
  7. 根据权利要求4所述的用于检测预定区域中特定标识图像的方法，其特征在于，根据特定标识图像在所述预定区域或预定区域图像中的特定位置信息，截取包含特定标识图像的区域范围，作为待检目标图像，包括：
    建立预定区域或包含预定区域的标准图像模板；
    根据所述预定区域或包含预定区域的图像遍历所述标准图像模板，通过所述标准图像模板与所述预定区域或包含预定区域的图像的对应关系，获取特定标识图像在预定区域或包含预定区域的图像中的位置信息；
    根据该位置信息截取所述预定区域或包含预定区域的图像中包含特定标识的图像区域，作为待检目标图像。
  8. 根据权利要求1所述的用于检测预定区域中特定标识图像的方法，其特征在于：
    所述阈值区间的上限值为所述获取的预定区域图像或包含预定区域图像的宽度、高度和上限系数的乘积；下限值为所述获取的预定区域图像或包含预定区域图像的宽度、高度和下限系数的乘积。
  9. 根据权利要求1所述的用于检测预定区域中特定标识图像的方法，其特征在于，所述将该候选连通域区域的图像特征与所述特定标识图像的相应图像特征进行比对，包括：采用单独图像特征比对或者组合图像特征比对；其中，所述图像特征包括连通域的体积比、连通域外接椭圆长轴方向、连通域外轮廓形状和/或连通域外接矩形长宽比。
  10. 根据权利要求9所述的用于检测预定区域中特定标识图像的方法，其特征在于，逐一获取所述候选连通域区域的图像特征，并将该候选连通域区域的图像特征与所述特定标识图像的相应图像特征进行比对，将比对结果相同或最接近的候选连通域区域视为所述特定标识图像所在区域，包括：
    以所述连通域的体积比和连通域外接椭圆长轴方向作为组合图像特征对候选连通域区域进行筛选，具体是：
    获取候选连通域区域的外接边框；
    计算所述外接边框内所述候选连通域中前景色的像素与该外界边框面积的体积比；
    判断该体积比是否大于预先设定的特定标识图像的阈值范围，若否，则舍弃该候选连通域区域；若是，对所述候选连通域区域进行填充；
    计算填充后的所述候选连通域区域外接椭圆长轴的方向，并获得该方向与设定的参考轴的夹角；
    判断该夹角是否在预定的阈值范围内，若是，则将该候选连通域区域视为所述特定标识图像所在区域；若否，则舍弃该候选连通域区域。
  11. 根据权利要求10所述的用于检测预定区域中特定标识图像的方法，其特征在于，包括：判断所述方向与所述参考轴的夹角在预定的阈值范围内后，以所述连通域外轮廓形状作为单独条件对所述候选连通域区域再次进行筛选，具体为：
    填充所述候选连通域区域，获得填充图像；
    检测所述填充图像的边缘，获得所述填充图像的边缘轮廓图像；
    分别获取所述边缘轮廓图像的上下边缘和左右边缘的采样像素点，并选取在设定阈值范围内的所述上下边缘采样像素点和左右边缘采样像素点；
    拟合所述选取的采样像素点，获得所述边缘轮廓图像的线段组合信息；
    判断所述线段组合信息及交点位置信息与标准特定标识图像的线段组合信息及交点位置信息是否相同；若是，则将候选连通域区域视为特定标识图像区域；若否，则输出未检测特定标识图像区域。
  12. 根据权利要求11所述的用于检测预定区域中特定标识图像的方法，其特征在于，所述获取所述边缘轮廓图像的上下边缘和左右边缘的采样像素点包括：
    以与所述填充图像有两个交点为条件，进行扫描，将该些交点作为采样像素点；
    所述选取在设定阈值范围内的所述上下边缘采样像素点和左右边缘采样像素点包括：
    判断获取的所述采样像素点的上下边缘之间的距离和左右边缘之间的距离是否在设定的阈值范围内，若是，则选取；若否，则舍弃。
  13. 根据权利要求12所述的用于检测预定区域中特定标识图像的方法，其特征在于，所述阈值范围为maxDist±maxDist×5％。
  14. 根据权利要求11所述的检测标准卡片中特定标识图像的方法，其特征在于，包括：
    根据预定区域或包含预定区域图像中特定标识的位置信息和各个交点位置信息，计算仿射变换矩阵；
    根据所述仿射变换矩阵获取所述预定区域内的相关信息。
  15. 根据权利要求1-3、5-14中任意一项所述的用于检测预定区域中特定标识图像方法，其特征在于：所述特定标识图像为银行卡联合组织或银行卡发卡组织的标志图形。
  16. 一种用于检测预定区域中特定标识图像的装置，其特征在于，包括：
    获取单元，用于获取所述预定区域的图像，作为待检目标图像；
    二值化处理单元，用于对所述待检目标图像进行二值化处理，获得二值化图像，及所述二值化图像的取反图像，二者均称为待检目标二值图像；
    连通域标定单元，用于标定所述待检目标二值图像的连通域，并将标定的连通域中像素个数满足设定阈值区间的连通域作为候选连通域；
    图像特征比对单元，用于逐一获取所述候选连通域的图像特征，并将该候 选连通域的图像特征与标准特定标识图像的相应图像特征进行比对，将比对结果相同或最接近的候选连通域视为特定标识图像所在位置。
  17. 根据权利要求16所述的用于检测预定区域中特定标识图像的装置，其特征在于，所述获取单元包括：
    预览图像获取单元，用于获取所述预定区域或包含该预定区域的图像的预览图像；
    计算单元，用于计算所述预览图像的区域范围；
    框定单元，用于将所述预览图像框定在所述区域范围内，获取所述预定区域或包含预定区域的图像，将该框定的区域范围内的图像作为待检目标图像。
  18. 根据权利要求17所述的用于检测预定区域中特定标识图像的装置，其特征在于，所述计算单元是指：根据所述预定区域的实际尺寸参数及预览图像的分辨率参数，获得框定所述预览图像的虚拟边框参数，具体为：
    根据所述预定区域或包含该预定区域图像的实际尺寸，判断所述预览图像高度乘以所述预定区域或包含该预定区域图像的实际宽度的值是否大于预览图像宽度乘以所述预定区域或包含该预定区域图像实际高度的值；
    若是，则所述虚拟边框的宽度为：预定区域或包含该预定区域图像的实际宽度乘以预览图像高度除以预定区域或包含该预定区域的图像实际高度乘以虚拟边框与所述预览图像的比值；所述虚拟边框的高度为：预览图像高度乘以虚拟边框与所述预览图像的比值；
    若否，则虚拟边框的宽度为：预览图像宽度乘以虚拟边框与所述预览图像的比值；虚拟边框的高度为：预定区域或包含该预定区域的图像的实际高度乘以预览图像宽度除以预定区域或包含该预定区域图像的实际宽度乘以虚拟边框与所述预览图像的比值；
    所述虚拟边框上边界为显示屏幕高度与虚拟边框高度差的二分之一；所述虚拟边框下边界为显示屏幕高度与虚拟边框高度和的二分之一；所述虚拟边框的左边界为显示屏幕宽度与虚拟边框宽度差的二分之一；所述虚拟边框的右边界为显示屏幕宽度与虚拟边框宽度和的二分之一。
  19. 根据权利要求16-18任意一项所述的用于检测预定区域中特定标识图像的装置，其特征在于，所述获取单元包括：
    截取单元，用于根据特定标识图像在所述预定区域或预定区域图像中的特 定位置信息，截取包含特定标识图像的区域范围，作为待检目标图像。
  20. 根据权利要求19所述的用于检测预定区域中特定标识图像的装置，其特征在于，所述特定位置信息为事先设定的特定标识图像所位于预定区域或包含预定区域图像上的位置。
  21. 根据权利要求20所述的用于检测预定区域中特定标识图像的装置，其特征在于，所述特定位置是位于所述预定区域或包含预定区域的图像的右上角或右下角或左上角或左下角。
  22. 根据权利要求19所述的用于检测预定区域中特定标识图像的装置，其特征在于，所述截取单元包括：
    建立单元，用于建立预定区域或包含预定区域的标准图像模板；
    遍历单元，用于根据所述预定区域或包含预定区域的图像遍历所述标准图像模板，通过所述标准图像模板与所述预定区域或包含预定区域的图像的对应关系，获取特定标识图像在预定区域或包含预定区域的图像中的位置信息；
    待检目标图像确定单元，用于根据该位置信息截取所述预定区域或包含预定区域的图像中包含特定标识的图像区域，作为待检目标图像。
  23. 根据权利要求16所述的用于检测预定区域中特定标识图像的装置，其特征在于，所述阈值区间的上限值为所述预定区域或包含预定区域的图像宽度、高度和上限系数的乘积；下限值为所述预定区域或包含预定区域的图像宽度、高度和下限系数的乘积。
  24. 根据权利要求16所述的用于检测预定区域中特定标识图像的装置，其特征在于：所述将该候选连通域区域的图像特征与所述特定标识图像的相应图像特征进行比对，包括：采用单独图像特征比对或者组合图像特征比对；其中，所述图像特征包括连通域的体积比、连通域外接椭圆长轴方向、连通域外轮廓形状和/或连通域外接矩形长宽比。
  25. 根据权利要求24所述的用于检测预定区域中特定标识图像的装置，其特征在于，包括：组合图像特征筛选单元，具体是：
    外接边框获取单元，用于获取候选连通域区域的外接边框；
    体积比计算单元，用于计算所述外接边框内所述候选连通域中前景色的像素与该外界边框面积的体积比；
    体积比阈值判断单元，用于判断该体积比是否大于预先设定的特定标识图 像的阈值范围；若否，则舍弃该候选连通域区域；若是，对所述候选连通域区域进行填充；
    填充计算单元，用于计算填充后的所述候选连通域区域外接椭圆长轴的方向，并获得该方向与设定的参考轴的夹角；
    夹角阈值判断单元，用于判断该夹角是否在预定的阈值范围内，若是，则将该候选连通域区域视为所述特定标识图像所在区域；若否，则舍弃该候选连通域区域。
  26. 根据权利要求25所述的用于检测预定区域中特定标识图像的装置，其特征在于，单独图像特征筛选单元，包括：
    填充单元，用于填充所述候选连通域区域，获得填充图像；
    边缘轮廓获取单元，用于根据检测所述填充图像的边缘，获得所述填充图像的边缘轮廓图像；
    采样单元，用于分别获取所述边缘轮廓图像的上下边缘和左右边缘的采样像素点，并选取在设定阈值范围内的所述上下边缘采样像素点和左右边缘采样像素点；
    采样像素点拟合单元，用于分别获取所述边缘轮廓图像的上下边缘和左右边缘的采样像素点，并选取在设定阈值范围内的所述上下边缘采样像素点和左右边缘采样像素点；
    线段组合判断单元，用于判断所述线段组合信息及交点位置信息与标准特定标识图像的线段组合信息和交点位置信息是否相同；若是，则将候选连通域区域视为特定标识图像区域；若否，则输出未检测特定标识图像区域。
  27. 根据权利要求26所述的用于检测预定区域中特定标识图像的装置，其特征在于，所述采样单元包括：
    扫描单元，以与所述填充图像有两个交点为条件，进行扫描，将该些交点作为采样像素点；
    距离阈值判断单元，用于判断获取的所述采样像素点的上下边缘之间的距离和左右边缘之间的距离是否在设定的阈值范围内，若是，则选取；若否，则舍弃。
  28. 根据权利要求27所述的用于检测预定区域中特定标识图像的装置，其特征在于，所述阈值范围为maxDist±maxDist×5％。
  29. 根据权利要求26所述的用于检测预定区域中特定标识图像的装置，其特征在于，包括：
    仿射变换单元，用于根据预定区域或包含预定区域图像中特定标识的位置信息和各个交点位置信息，计算仿射变换矩阵；
    相关信息获取单元，用于根据所述仿射变换矩阵获取相关信息。
  30. 根据权利要求16-18、20-29中任意一项所述的用于检测预定区域中特定标识图像的装置，其特征在于，所述特定标识图像为银行卡联合组织或银行卡发卡组织的标志图形。
PCT/CN2015/083277 2014-07-29 2015-07-03 一种用于检测预定区域中特定标识图像的方法及装置 Ceased WO2016015548A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201410367807.9 2014-07-29
CN201410367807.9A CN105303189B (zh) 2014-07-29 2014-07-29 一种用于检测预定区域中特定标识图像的方法及装置

Publications (1)

Publication Number Publication Date
WO2016015548A1 true WO2016015548A1 (zh) 2016-02-04

Family

ID=55180375

Family Applications (2)

Application Number Title Priority Date Filing Date
PCT/CN2015/083277 Ceased WO2016015548A1 (zh) 2014-07-29 2015-07-03 一种用于检测预定区域中特定标识图像的方法及装置
PCT/US2015/042606 Ceased WO2016018987A1 (en) 2014-07-29 2015-07-29 Detecting specified image identifiers on objects

Family Applications After (1)

Application Number Title Priority Date Filing Date
PCT/US2015/042606 Ceased WO2016018987A1 (en) 2014-07-29 2015-07-29 Detecting specified image identifiers on objects

Country Status (4)

Country Link
US (3) US9799119B2 (zh)
CN (1) CN105303189B (zh)
TW (1) TWI655586B (zh)
WO (2) WO2016015548A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108776140A (zh) * 2018-06-08 2018-11-09 福建新大陆电脑股份有限公司 一种基于机器视觉的印刷品瑕疵检测方法及系统
CN110390666A (zh) * 2019-06-14 2019-10-29 平安科技（深圳）有限公司 道路损伤检测方法、装置、计算机设备及存储介质
CN114862795A (zh) * 2022-04-29 2022-08-05 吉林大学 物体表面温度测量系统、测量方法、智能终端及存储介质
CN116993731A (zh) * 2023-09-27 2023-11-03 山东济矿鲁能煤电股份有限公司阳城煤矿 基于图像的盾构机刀头缺陷检测方法

Families Citing this family (67)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2921989B1 (en) * 2014-03-17 2024-12-04 Université de Genève Method for object recognition and/or verification on portable devices
CN105303189B (zh) 2014-07-29 2019-08-20 阿里巴巴集团控股有限公司 一种用于检测预定区域中特定标识图像的方法及装置
JP6511980B2 (ja) * 2015-06-19 2019-05-15 カシオ計算機株式会社 画像処理装置、画像処理方法及びプログラム
CN106127751B (zh) * 2016-06-20 2020-04-14 北京小米移动软件有限公司 图像检测方法、装置以及系统
CN106127826B (zh) * 2016-06-27 2019-01-22 安徽慧视金瞳科技有限公司 一种用于投影交互系统的连通域标记方法
CN106228546B (zh) * 2016-07-13 2019-01-29 广州视源电子科技股份有限公司 一种板卡的检测方法及装置
CN107657251A (zh) * 2016-07-26 2018-02-02 阿里巴巴集团控股有限公司 确定身份证件显示表面的设备及方法，图像识别方法
JP2018055496A (ja) * 2016-09-29 2018-04-05 日本電産サンキョー株式会社 媒体認識装置および媒体認識方法
CN108073927B (zh) * 2016-11-18 2020-05-15 深圳怡化电脑股份有限公司 一种字符识别方法及装置
CN106770321B (zh) * 2016-12-08 2019-07-19 广东工业大学 一种基于多阈值区间的塑料零件缺陷检测方法
CN106709394B (zh) * 2016-12-12 2019-07-05 北京慧眼智行科技有限公司 一种图像处理方法及装置
US10515319B2 (en) * 2016-12-16 2019-12-24 Fetch Robotics, Inc. System and method for computing a probability that an object comprises a target
WO2018112788A1 (zh) * 2016-12-21 2018-06-28 华为技术有限公司 图像处理方法及设备
US10438089B2 (en) * 2017-01-11 2019-10-08 Hendricks Corp. Pte. Ltd. Logo detection video analytics
CN106875408B (zh) * 2017-02-27 2020-03-17 网易（杭州）网络有限公司 用于截图的方法、装置及终端设备
CN108510468B (zh) * 2017-02-28 2020-11-20 深圳市朗驰欣创科技股份有限公司 一种识别电磁阀开关状态的方法和装置
CN107240185B (zh) * 2017-06-23 2019-09-20 深圳怡化电脑股份有限公司 一种冠字号识别方法、装置、设备及存储介质
JP6977425B2 (ja) * 2017-09-14 2021-12-08 セイコーエプソン株式会社 画像処理装置、二値画像生産方法および画像処理プログラム
CN107657621B (zh) * 2017-10-20 2021-04-13 南京林业大学 基于线性区域生长的二维激光点云序列实时分割方法
US11074472B2 (en) 2017-11-14 2021-07-27 Symbol Technologies, Llc Methods and apparatus for detecting and recognizing graphical character representations in image data using symmetrically-located blank areas
CN108022239B (zh) * 2017-12-05 2020-03-10 郑州轻工业学院 一种基于机器视觉的起泡葡萄酒褐变过程检测装置及方法
CN108460388B (zh) * 2018-01-18 2022-03-29 深圳市易成自动驾驶技术有限公司 定位标志的检测方法、装置及计算机可读存储介质
CN108764327B (zh) * 2018-05-23 2022-04-05 北京五八信息技术有限公司 图像模板检测方法、装置、计算设备及可读存储介质
CN109641351B (zh) * 2018-06-08 2021-11-26 深圳蓝胖子机器人有限公司 一种物体特征的识别方法、视觉识别装置及机器人
CN109410215A (zh) * 2018-08-02 2019-03-01 北京三快在线科技有限公司 图像处理方法、装置、电子设备及计算机可读介质
CN109389150B (zh) * 2018-08-28 2022-04-05 东软集团股份有限公司 图像一致性比对方法、装置、存储介质及电子设备
CN109409272B (zh) * 2018-10-17 2021-06-18 北京空间技术研制试验中心 基于机器视觉的电缆验收系统和方法
CN109600599A (zh) * 2018-10-29 2019-04-09 上海神添实业有限公司 一种快速定位目标的立体视觉处理装置及其处理方法
CN109711415A (zh) * 2018-11-13 2019-05-03 平安科技（深圳）有限公司 证件轮廓确定方法、装置及存储介质、服务器
US11436366B2 (en) 2019-01-21 2022-09-06 Bitdefender IPR Management Ltd. Parental control systems and methods for detecting an exposure of confidential information
KR102226843B1 (ko) * 2019-02-26 2021-03-12 주식회사 핀그램 오브젝트 검출 시스템 및 그 방법
CN110063753B (zh) * 2019-03-18 2022-03-22 深圳蓝韵医学影像有限公司 超声图像中测量胎儿头围的方法与装置
CN109949331B (zh) * 2019-04-17 2021-02-02 合肥泰禾智能科技集团股份有限公司 货箱边缘检测方法及装置
CN110147838B (zh) * 2019-05-20 2021-07-02 苏州微创关节医疗科技有限公司 一种产品规格录入、检测方法及系统
US10796199B1 (en) * 2019-05-29 2020-10-06 Alibaba Group Holding Limited Image recognition and authentication
CN110830716B (zh) * 2019-11-08 2021-08-27 南京星云数字技术有限公司 图像生成方法、装置、计算机设备和存储介质
CN110929738A (zh) * 2019-11-19 2020-03-27 上海眼控科技股份有限公司 证卡边缘检测方法、装置、设备及可读存储介质
CN112966538A (zh) * 2019-11-27 2021-06-15 深圳怡化电脑股份有限公司 一种票据检测方法、装置、计算机设备及存储介质
CN113048905B (zh) * 2019-12-27 2022-08-19 上海微电子装备（集团）股份有限公司 对准标记图像制作方法、对准标记测量方法及测量装置
US11508169B2 (en) * 2020-01-08 2022-11-22 Palo Alto Research Center Incorporated System and method for synthetic image generation with localized editing
CN111242116B (zh) * 2020-01-21 2022-11-08 上海万物新生环保科技集团有限公司 屏幕定位方法及装置
CN111292310B (zh) * 2020-02-14 2024-01-26 生物岛实验室 获取生物组织图像的方法、装置、电子设备及存储介质
CN111724445B (zh) * 2020-05-08 2023-05-26 华中科技大学 一种大视野小尺寸识别码的识别方法及系统
CN111738127B (zh) * 2020-06-17 2023-08-25 安徽淘云科技股份有限公司 实体书就位检测方法、装置、电子设备和存储介质
CN114022403A (zh) * 2020-07-16 2022-02-08 京东方科技集团股份有限公司 检测显示面板不良的方法及装置
CN111898602B (zh) * 2020-08-10 2024-04-16 赞同科技股份有限公司 一种图像中的凭证号码区域识别方法、装置及设备
CN111931771B (zh) * 2020-09-16 2021-01-01 深圳壹账通智能科技有限公司 票据内容识别方法、装置、介质及电子设备
CN112183517B (zh) * 2020-09-22 2023-08-11 平安科技（深圳）有限公司 证卡边缘检测方法、设备及存储介质
CN112215862B (zh) * 2020-10-12 2024-01-26 虎博网络技术(上海)有限公司 静态标识检测方法、装置、终端设备和可读存储介质
CN112668676A (zh) * 2020-12-12 2021-04-16 浙江大华技术股份有限公司 包裹的安检图像处理方法、电子装置和存储介质
CN113538574B (zh) * 2021-01-04 2022-09-27 腾讯科技（深圳）有限公司 位姿定位方法、装置、设备及计算机可读存储介质
US11921831B2 (en) * 2021-03-12 2024-03-05 Intellivision Technologies Corp Enrollment system with continuous learning and confirmation
CN113221754A (zh) * 2021-05-14 2021-08-06 深圳前海百递网络有限公司 快递单图像检测方法、装置、计算机设备和存储介质
CN113436201B (zh) * 2021-05-26 2023-01-06 清华大学 物品指纹区域选取方法及装置
CN113538339A (zh) * 2021-06-18 2021-10-22 杭州类脑科技有限公司 一种基于医学影像的长短径估算方法
TWI816150B (zh) * 2021-07-01 2023-09-21 鴻海精密工業股份有限公司 目標對象檢測方法、電子設備及儲存介質
CN113361643A (zh) * 2021-07-02 2021-09-07 人民中科(济南)智能技术有限公司 基于深度学习的通用标志识别方法、系统、设备及存储介质
CN113901935B (zh) * 2021-10-15 2025-12-05 丰图科技(深圳)有限公司 杆状目标识别方法、装置、服务器及存储介质
CN114359376B (zh) * 2021-12-27 2025-11-21 浙江华睿科技股份有限公司 包裹定位方法、装置、电子设备和存储介质
CN114596330B (zh) * 2022-01-26 2025-08-19 广州盖盟达工业品有限公司 一种基于动态视觉的商品定位方法、装置及存储介质
CN114627280B (zh) * 2022-03-24 2025-07-15 北京十六进制科技有限公司 一种图像歪斜纠正方法、系统、终端和可读存储介质
CN114744721B (zh) * 2022-04-28 2025-06-13 优必康(青岛)科技有限公司 一种机器人的充电控制方法、终端设备及存储介质
TWI794132B (zh) * 2022-09-19 2023-02-21 威盛電子股份有限公司 偵測誤判物件的系統
CN115375686B (zh) * 2022-10-25 2023-01-24 山东鲁玻玻璃科技有限公司 一种基于图像处理的玻璃边缘瑕疵检测方法
CN116030086B (zh) * 2023-01-04 2025-08-12 中山大学 一种基于信息融合的目标检测方法及装置
CN117036691A (zh) * 2023-07-19 2023-11-10 珠海读书郎软件科技有限公司 一种试卷卷面遮挡识别方法、存储介质及设备
KR20250058420A (ko) * 2023-10-23 2025-04-30 현대모비스 주식회사 Avps 마커 기반 측위를 위한 avps 마커의 인식 방법 및 장치

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101576958A (zh) * 2009-06-15 2009-11-11 西安交通大学 一种基于形状特征的带边框矩形印章图形提取方法
CN102054168A (zh) * 2010-12-23 2011-05-11 武汉大学苏州研究院 一种有价票据圆形印鉴识别方法
CN102496025A (zh) * 2011-12-07 2012-06-13 方正国际软件有限公司 一种文档中含有人物肖像的图像检测方法及系统
US20140161353A1 (en) * 2009-03-30 2014-06-12 The Neat Company, Inc. d/b/a Neatreceipts, Inc. Table grid detection and separation

Family Cites Families (43)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5436970A (en) 1993-02-18 1995-07-25 Eastman Kodak Company Method and apparatus for transaction card verification
US5546462A (en) 1993-04-09 1996-08-13 Washington University Method and apparatus for fingerprinting and authenticating various magnetic media
US20020009208A1 (en) 1995-08-09 2002-01-24 Adnan Alattar Authentication of physical and electronic media objects using digital watermarks
US6961442B2 (en) 2001-03-09 2005-11-01 Digimarc Corporation Watermarking a carrier on which an image will be placed or projected
US7555139B2 (en) 1995-05-08 2009-06-30 Digimarc Corporation Secure documents with hidden signals, and related methods and systems
US7770013B2 (en) 1995-07-27 2010-08-03 Digimarc Corporation Digital authentication with digital and analog documents
US6788800B1 (en) 2000-07-25 2004-09-07 Digimarc Corporation Authenticating objects using embedded data
US6819775B2 (en) 1996-07-05 2004-11-16 ECOLE POLYTECHNIQUE FéDéRALE DE LAUSANNE Authentication of documents and valuable articles by using moire intensity profiles
US8496168B1 (en) * 1998-04-17 2013-07-30 Diebold Self-Service Systems Division Of Diebold, Incorporated Banking system controlled responsive to data bearing records
US6487301B1 (en) 1998-04-30 2002-11-26 Mediasec Technologies Llc Digital authentication with digital and analog documents
US6785815B1 (en) 1999-06-08 2004-08-31 Intertrust Technologies Corp. Methods and systems for encoding and protecting data using digital signature and watermarking techniques
WO2001080169A1 (en) 2000-04-17 2001-10-25 Digimarc Corporation Authentication of physical and electronic media objects using digital watermarks
US7246239B2 (en) 2001-01-24 2007-07-17 Digimarc Corporation Digital watermarks for checking authenticity of printed objects
US7213757B2 (en) 2001-08-31 2007-05-08 Digimarc Corporation Emerging security features for identification documents
US7537170B2 (en) 2001-08-31 2009-05-26 Digimarc Corporation Machine-readable security features for printed objects
EP1477940B1 (en) 2003-04-30 2007-08-01 Hewlett-Packard Development Company, L.P. Authentication method and system
JP2005005768A (ja) 2003-06-09 2005-01-06 Seiko Epson Corp 画像処理装置、画像処理方法及びプログラム
WO2007087194A2 (en) 2006-01-20 2007-08-02 Glenbrook Associates, Inc. System and method for the automated processing of physical objects
JP4764808B2 (ja) * 2006-12-07 2011-09-07 キヤノン株式会社 画像ファイル管理装置及び画像ファイル管理方法
US8019115B2 (en) 2007-04-25 2011-09-13 Graphic Security Systems Corp. Object authentication using a portable digital image acquisition device
CN101751572A (zh) * 2008-12-05 2010-06-23 比亚迪股份有限公司 一种图案检测方法、装置、设备及系统
CN102194118B (zh) * 2010-03-02 2013-04-10 方正国际软件(北京)有限公司 一种从图像中抽取信息的方法及装置
US8285074B2 (en) * 2010-09-01 2012-10-09 Palo Alto Research Center Incorporated Finding low variance regions in document images for generating image anchor templates for content anchoring, data extraction, and document classification
US9436816B2 (en) * 2010-12-16 2016-09-06 Microsoft Technology Licensing, Llc Supplementing biometric identification with device identification
US9443298B2 (en) 2012-03-02 2016-09-13 Authentect, Inc. Digital fingerprinting object authentication and anti-counterfeiting system
CN102426647B (zh) * 2011-10-28 2014-03-12 Tcl集团股份有限公司 一种台标识别的方法、装置
US8910861B2 (en) * 2012-04-05 2014-12-16 Bank Of America Corporation Automatic teller machine (“ATM”) including a user-accessible USB port
US9489754B2 (en) 2012-06-06 2016-11-08 Apple Inc. Annotation of map geometry vertices
CN102779279B (zh) * 2012-06-14 2014-09-03 南京航空航天大学 搅拌摩擦焊弧形纹间距的图像识别方法
CN102799653A (zh) * 2012-06-29 2012-11-28 中国科学院自动化研究所 一种基于空间连通域预定位的商标检测方法
CN102819728A (zh) * 2012-07-17 2012-12-12 中国航天科工集团第三研究院第八三五七研究所 一种基于分类模板匹配的交通标志检测方法
US8705836B2 (en) * 2012-08-06 2014-04-22 A2iA S.A. Systems and methods for recognizing information in objects using a mobile device
CN102855472A (zh) * 2012-08-14 2013-01-02 杨端端 一种名片识别中的自动拍摄方法
US20140053067A1 (en) 2012-08-17 2014-02-20 Kenneth C. Tkatchuk Method and Apparatus for Sequentially Displaying a Plurality of Images Including Selective Asynchronous Matching of a Subset of the Images
CN103679164A (zh) * 2012-09-21 2014-03-26 阿里巴巴集团控股有限公司 一种基于移动终端的标志识别处理方法和系统
JP6102156B2 (ja) 2012-09-28 2017-03-29 オムロン株式会社 画像処理システムおよび画像処理方法
US9349152B2 (en) * 2012-10-02 2016-05-24 Simetricka Systems Ltd. Image identifiers and methods and systems of presenting image identifiers
US20150006360A1 (en) * 2013-06-28 2015-01-01 Google Inc. Threshold Confidence Levels for Extracted Card Data
US9330310B2 (en) * 2013-11-21 2016-05-03 Tencent Technology (Shenzhen) Company Limited Methods and devices for obtaining card information
CN103679147A (zh) * 2013-12-05 2014-03-26 广州绿怡信息科技有限公司 手机型号的识别方法与装置
US9191403B2 (en) * 2014-01-07 2015-11-17 Fair Isaac Corporation Cyber security adaptive analytics threat monitoring system and method
US9311639B2 (en) * 2014-02-11 2016-04-12 Digimarc Corporation Methods, apparatus and arrangements for device to device communication
CN105303189B (zh) * 2014-07-29 2019-08-20 阿里巴巴集团控股有限公司 一种用于检测预定区域中特定标识图像的方法及装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140161353A1 (en) * 2009-03-30 2014-06-12 The Neat Company, Inc. d/b/a Neatreceipts, Inc. Table grid detection and separation
CN101576958A (zh) * 2009-06-15 2009-11-11 西安交通大学 一种基于形状特征的带边框矩形印章图形提取方法
CN102054168A (zh) * 2010-12-23 2011-05-11 武汉大学苏州研究院 一种有价票据圆形印鉴识别方法
CN102496025A (zh) * 2011-12-07 2012-06-13 方正国际软件有限公司 一种文档中含有人物肖像的图像检测方法及系统

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108776140A (zh) * 2018-06-08 2018-11-09 福建新大陆电脑股份有限公司 一种基于机器视觉的印刷品瑕疵检测方法及系统
CN108776140B (zh) * 2018-06-08 2024-02-02 新大陆数字技术股份有限公司 一种基于机器视觉的印刷品瑕疵检测方法及系统
CN110390666A (zh) * 2019-06-14 2019-10-29 平安科技（深圳）有限公司 道路损伤检测方法、装置、计算机设备及存储介质
CN110390666B (zh) * 2019-06-14 2023-06-27 平安科技（深圳）有限公司 道路损伤检测方法、装置、计算机设备及存储介质
CN114862795A (zh) * 2022-04-29 2022-08-05 吉林大学 物体表面温度测量系统、测量方法、智能终端及存储介质
CN116993731A (zh) * 2023-09-27 2023-11-03 山东济矿鲁能煤电股份有限公司阳城煤矿 基于图像的盾构机刀头缺陷检测方法
CN116993731B (zh) * 2023-09-27 2023-12-19 山东济矿鲁能煤电股份有限公司阳城煤矿 基于图像的盾构机刀头缺陷检测方法

Also Published As

Publication number Publication date
CN105303189B (zh) 2019-08-20
US10360689B2 (en) 2019-07-23
US10885644B2 (en) 2021-01-05
CN105303189A (zh) 2016-02-03
TW201604797A (zh) 2016-02-01
TWI655586B (zh) 2019-04-01
WO2016018987A1 (en) 2016-02-04
US20180005390A1 (en) 2018-01-04
US20190295267A1 (en) 2019-09-26
US9799119B2 (en) 2017-10-24
US20160034783A1 (en) 2016-02-04
HK1218582A1 (zh) 2017-02-24

Similar Documents

Publication Publication Date Title
CN105303189B (zh) 一种用于检测预定区域中特定标识图像的方法及装置
EP3620981B1 (en) Object detection method, device, apparatus and computer-readable storage medium
TWI654567B (zh) Method and apparatus for extracting specific information from standard cards
US10586321B2 (en) Automatic detection, counting, and measurement of lumber boards using a handheld device
CN108230376B (zh) 遥感图像处理方法、装置和电子设备
US8811751B1 (en) Method and system for correcting projective distortions with elimination steps on multiple levels
CN107545207A (zh) 基于图像处理的dm二维码识别方法及装置
US8897600B1 (en) Method and system for determining vanishing point candidates for projective correction
CN108280450A (zh) 一种基于车道线的高速公路路面检测方法
US8913836B1 (en) Method and system for correcting projective distortions using eigenpoints
CN111160395A (zh) 图像识别方法、装置、电子设备和存储介质
WO2021195873A1 (zh) 识别sfr测试卡图像中感兴趣区域的方法及装置、介质
CN104574312A (zh) 标靶图像求圆心的方法和装置
JP2016162420A (ja) 注目領域検出装置、注目領域検出方法及びプログラム
CN115830049A (zh) 角点检测的方法和装置
US11216905B2 (en) Automatic detection, counting, and measurement of lumber boards using a handheld device
CN110263778A (zh) 一种基于图像识别的抄表方法及装置
CN110135288B (zh) 一种电子证照的快速核对方法及装置
CN112052859A (zh) 一种自由场景下的车牌精准定位方法及装置
CN114742142B (zh) 液晶显示屏的检测方法及液晶显示屏检测设备
CN106910196A (zh) 一种图像检测方法及装置
EP3872707A1 (en) Automatic detection, counting, and measurement of lumber boards using a handheld device
CN115809999A (zh) 器件上目标对象检测方法、装置、电子设备及存储介质
HK1218582B (zh) 一种用於检测预定区域中特定标识图像的方法及装置
CN120912095B (zh) 一种物流跟踪系统及方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15826547

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15826547

Country of ref document: EP

Kind code of ref document: A1