CN102254144A - Robust method for extracting two-dimensional code area in image - Google Patents

Robust method for extracting two-dimensional code area in image Download PDF

Info

Publication number
CN102254144A
CN102254144A CN 201110193354 CN201110193354A CN102254144A CN 102254144 A CN102254144 A CN 102254144A CN 201110193354 CN201110193354 CN 201110193354 CN 201110193354 A CN201110193354 A CN 201110193354A CN 102254144 A CN102254144 A CN 102254144A
Authority
CN
Grant status
Application
Patent type
Prior art keywords
image
dimensional code
block
region
pyramid
Prior art date
Application number
CN 201110193354
Other languages
Chinese (zh)
Inventor
唐鹏
杜涛
王俊峰
陈懿
高琳
Original Assignee
四川大学
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

Links

Abstract

A robust method for extracting a two-dimensional code area in an image is disclosed, which comprises the following steps of: firstly, building a multi-scale Gaussian image pyramid according to an original image; secondly, partitioning each layer in the image pyramid; thirdly, binarizing all the image blocks, respectively, in order to obtain the binarization result of each layer of image; fourthly, fusing the binary images under a plurality of scales to obtain an image partition result; and finally, searching for the two-dimensional code area in unit of the image block and extracting the two-dimensional code area in the image by analyzing and calculating the convex hull of a characteristic point set through a connecting body. The method disclosed by the invention is capable of handling the common complex cases such as uneven illumination, background interference and the like, and has good robustness; as a search policy from coarse to fine under a plurality of scales is utilized, the method is simple and quick, and thereby is capable of taking both of the instantaneity and accuracy of handling into account.

Description

一种鲁棒的图像中二维码区域提取方法 Two-dimensional code region extraction method for robust image

技术领域 FIELD

[0001] 本发明属于计算机数字图像处理方法,特别是一种图像中准确分割提取二维码区域的方法。 [0001] The present invention pertains to a computer digital image processing method, particularly to a method of extracting an image segmentation accurate two-dimensional code area.

背景技术 Background technique

[0002] 二维条码(2-dimensional bar code)是指在一维条码的基础上扩展出另一维具有可读性的条码。 [0002] a two-dimensional bar code (2-dimensional bar code) refers to the other dimension extension readable bar code on the basis of the one-dimensional bar. 一维条码的宽度记载着数据,而其长度没有记载数据。 Width of one-dimensional bar recorded data, and its data length is not described. 二维条码的长度、 宽度均记载着数据。 Dimensional bar code length, width data are recorded. 二维条码有一维条码没有的“定位点”和“容错机制”。 There is a two-dimensional bar-dimensional bar code not "fix" and "fault tolerance." 容错机制在即使没有辨识到全部的条码、或是说条码有污损时,也可以正确地还原条码上的信息。 Fault tolerance even when there is no recognition to all of the bar code, bar code or that have defaced, you can correctly restore the information on the bar code. 在许多种类的二维条码中,常用的码制有:Data Matrix, QR Code, P DF417,汉信码等。 In many types of two-dimensional bar code, common code system has: Data Matrix, QR Code, P DF417, Chinese letter codes. 二维条码具有储存量大、保密性高、追踪性高、抗损性强、备援性大、成本便宜等特性,这些特性特别适用于表单、安全保密、追踪、证照、存货盘点、资料备援等方面。 Two-dimensional bar code has a large storage capacity, high security, high tracking resistance, resistance to damage and strong, redundancy large, cheap cost and other properties, which is especially suitable for the form, security, tracking, license, stock inventory, information available aid and so on.

[0003] 通过手机或相机摄像头拍摄包含二维码的图像,利用数字图像处理技术进行识另|J,是国内外对二维条码的主要研究方向。 [0003] photographed by a camera or cell phone camera image contains a two-dimensional code, a knowledge of the other digital image processing technology | J, is home to two-dimensional bar code in the main research directions. 对图像中二维码区域的提取,是二维码图像识别中最关键的步骤之一,这其中涉及到图像二值化方法和二维码区域定位。 The extracted two-dimensional code area in the image, two-dimensional code image recognition is one of the most critical step, which involves image binarization method and the two-dimensional code area positioned. 针对图像的二值化问题,目前主要分为全局阈值方法和局部阈值方法。 For the binary image problem, currently divided into global threshold method and local thresholding method. 全局阈值中最简单(也是最通用) 的方法是将在某一个灰度值以下的像素全部设为黑色,而其他所有的像素设为白色,阈值的选取是该方法的关键,在《快速响应矩阵码QR Code》(中华人民共和国国家标准GB/T 18284-2000)中给出的参考译码算法中,所选的阈值是在图像灰度值最大值与最小值之间的中值,但如果图像中光照不均勻,将难以区分出二维码区域与图像的背景区域。 Global threshold simplest (and most common) method is to set all pixels in the black gradation value of one, while all other pixels to white, threshold selection method is key to the rapid response " Referring matrix code decoding algorithm QR code "(PRC national standard GB / T 18284-2000) are given, the selected threshold is a value between the maximum and minimum values ​​of the gray image, but If the image uneven illumination, it is difficult to distinguish the two-dimensional code region and the background region of the image. 局部阈值方法就是按照一定的规则将整幅图像划分为多个窗口,对这些窗口中的每一个窗口,再按照全局阈值的方法进行二值化处理。 Local threshold value method in accordance with certain rules is the whole image into a plurality of windows, each of these windows a window, and then performs binarization processing method according to the global threshold. 局部阈值方法在处理光照不均或背景十分复杂的图像时效果较好,二值化图像中存在分块效应。 Local thresholding method is better in image processing or background illumination unevenness is very complex, the presence of blocking effects in the binarized image. 在二维码区域定位方面,《快速响应矩阵码QR Code》提供的参考方法是,在整幅图像中搜索QR码的寻像图形,根据寻像图形中各元素相当宽度的比例是1 : 1 : 3 : 1 : 1的特点,在水平和垂直方向进行探测,找出三个寻像图形,进而确定出二维码的方位。 Positioning the two-dimensional code area, the reference method "quick response code QR Code" is provided, the QR code searches the entire image in the finder pattern The pattern elements viewfinder considerable width ratio is 1: 1 : 3: 1: 1 features, in horizontal and vertical directions to detect, identify three finder pattern, and to determine the orientation of the two-dimensional code. 由于这种方法要对整个图像进行搜索,因此在分辨率较大的图像中探测很花费时间,此外复杂背景的干扰也使得探测会明显降低二维码的识别速度。 Since this method is to search the entire image, thus detecting very time resolution of the larger image in addition complex background interference makes detection will significantly reduce the two-dimensional code recognition speed.

发明内容 SUMMARY

[0004] 本发明的目的是提供一种鲁棒的图像中二维码区域提取方法,为二维码的识别提供准确有效的依据。 [0004] The object of the present invention is to provide a robust method of two-dimensional code region extraction image, provide accurate and effective basis for the identification of the two-dimensional code.

[0005] 本发明的目的是这样实现的:一种鲁棒的图像中二维码区域提取方法,包括以下步骤: [0005] The object of the present invention is implemented as follows: an image region extracting method of a robust two-dimensional code, comprising the steps of:

[0006] 1. 1)对采集的二维码图像进行灰度化处理,建立一种多尺度的高斯图像金字塔; [0006] 1.1) on the two-dimensional code image captured gradation processing, to create a multi-scale Gaussian image pyramid;

[0007] 1. 2)对金字塔中的每一层进行分块,根据每个分块图像的局部灰度分布信息,采用最大类间方差方法对每个分块图像进行二值化;[0008] 1.3)根据图像金字塔中每一层的二值化结果,对多个尺度上的二值图像进行融合,得到最终的二值图像; [0007] 1.2) of each layer of the pyramid is divided into blocks according to the local intensity distribution information of each image block, the method variance binarized image for each block maximum classes; [0008 ] 1.3) the image pyramid binarization result of each layer of the binary image on a plurality of scales are fused to give a final binary image;

[0009] 1.4)在图像金字塔的最底层,即原始尺度下,逐行扫描每个分块二值图像;扫描过程中,若发现相邻像素的像素值不同,则称为一次跳变,计算每个分块二值图像中出现的跳变次数;根据二维码具有丰富纹理的特点,将跳变次数大于某个阈值的分块设为二维码候选块,候选块作为后续二维码搜索区域; [0009] 1.4) the lowest level of the image pyramid, i.e. the original scale, progressive scan each block binary image; scanning process, if found different neighboring pixel values, is called a transition, calculating frequency hopping occurs in each block image binarization; two-dimensional code having a rich texture characteristics, the hop number becomes greater than a certain threshold, the block is defined as a two-dimensional code candidate block as the candidate block subsequent two-dimensional code the search area;

[0010] 1. 5)以分块为单位进行连通体分析,找出所有在图像上相邻接的候选块,建立多个连通体区域,每个连通体区域则作为候选二维码区域;根据候选二维码区域的形状特征, 确定包含有二维码的连通体区域; [0010] 1.5) in divided block units connected component analysis to identify all of the candidate blocks in the adjacent image regions to establish a plurality of connected components, each connected component of the regions as two-dimensional code candidate region; the shape of the two-dimensional code candidate region characteristic determining region comprising vias two-dimensional code;

[0011] 1.6)以图像块为单位,选取包含有二维码的连通体区域的内边缘块和外边缘块, 从而生成边缘带;通过Harris角点检测得到边缘带中的特征角点,建立特征角点集;采用Sklansky算法计算出每个角点集的外接凸包;提取每个外接凸包所对应的图像区域,得到最终的二维码区域。 [0011] 1.6) in image block units, comprising selecting a block and the inner edge of the outer edge of the two-dimensional code block connected component region, thereby generating the edge band; the feature points obtained through the edge band Harris corner detection, to establish set of feature points; Sklansky algorithm employed to calculate the corner point of each set of external convex hull; extracting an image region corresponding to each of the convex hull external obtain a final two-dimensional code area.

[0012] 上述步骤1. 2)中对图像金字塔中的每一层进行分块的步骤如下: Step [0012] Step 1.2 above) in an image pyramid for each layer block as follows:

[0013] 2. 1)按照设定的比例系数,将最底层的原始图像,划分成适当大小的图像块;其中比例系数为分块的宽度与图像宽度的比值; [0013] 2.1) in proportion to the coefficient setting, the bottom of the original image, the image is divided into blocks of appropriate size; wherein the scale factor of the image width ratio of the width of the block;

[0014] 2. 2)对金字塔上一层图像进行分块,分块的比例系数取上一层比例系数的两倍; [0014] 2.2) on one pyramid image block, the block scale factor is taken twice the scale factor of one;

[0015] 2. 3)重复步骤2. 2),直至到图像金字塔最顶层。 [0015] 2.3) repeating steps 2.2), until the top level of the image pyramid.

[0016] 上述步骤1.2)中对图像二值化处理的方法是:利用最大类间方差算法,计算出图像金字塔每层图像中的每个分块的分割阈值,然后使用阈值对相应的分块进行二值化处理,从而得到每层的二值图像;通过融合不同尺度下的二值图像,得到最终的二值化结果; 融合的实现方法是先将所有尺度下的二值图像,使用内插生成与原始图像等分辨率的二值图像,图像对齐后按像素取中值作为所需的二值化结果。 The method [0016] The step 1.2) to the image binarization process is: using Otsu algorithm to calculate the segmentation threshold for each block of the image pyramid image of each layer, and then using a threshold value for the corresponding block binarization processing, thereby obtaining a binary image of each layer; by fusing binary image at different scales, to obtain the final binarization result; to achieve fusion of the binary image is first of all scales, the use of and the like for interpolating the original image resolution binary image, the image taken aligned pixel value as the binarization result desired.

[0017] 上述步骤1. 5)中所述的连通体区域是相邻候选块之间的八连通区域,每个连通体区域作为一个候选二维码区域,依据该区域外接矩形的长宽比规则,确定包含有二维码的连通体区域。 [0017] Step 1.5 above) in the region of eight vias connected region between adjacent candidate blocks, each connected component area as a two-dimensional code candidate region, based on the aspect ratio of the circumscribed rectangular region rules that determine connected components include two-dimensional code region.

[0018] 当图像中存在复杂背景或是光照不均勻的情况时,图像整体与局部之间的统计特性差异较大,无法用单一的阈值进行二值化,将二维码区域与背景区分开。 [0018] When the complex background or uneven illumination situation is present in the image, the statistical properties of the entire image between the local differences, not a single binarizing threshold value, the two-dimensional code region separated from the background . 本发明采用分治法的思想对原始图像进行分块处理,同时设计一种多尺度分割的策略,减小分块二值化引起的分块效应。 Thought the present invention uses a divide and conquer method the original image into blocks, while the design strategy for multiscale segmentation, block reduced block effect caused by binarization. 再在二值图像中搜索包含二维码的图像块,通过连通体分析以及计算特征点外接凸包,同时提取出图像中多个二维码。 Then searches the binary image comprises two-dimensional code image block by connected component analysis and calculation of the convex hull external feature points, and extracts an image in a plurality of two-dimensional code.

[0019] 本发明的有益效果主要有一下几点:(1)本发明提供的技术方案对于光照不均、 复杂背景等常见的干扰情况具有较好的鲁棒性;(2)具有较好的普遍适用性,而不局限于某一类型的二维码,只要二维码模块分布比较均勻,都能够应用本发明的方案进行区域提取;(3)采用多尺度下由粗到精的搜索策略,方法简便快捷,能够兼顾处理的实时性和准确性;(4)由于技术方案中是并行地提取图像中多个二维码,因此本发明特别适用于具有并行处理能力的软硬件系统。 [0019] Advantageous effects of the invention are mainly the following points: (1) the present invention provides a technical solution for the common interference illumination unevenness, etc. with complex background is robust; (2) has a good universal applicability, without being limited to a certain type of two-dimensional code, as long as the two-dimensional code module is distributed more evenly, possible to apply the present invention region extraction; (3) under the multi-scale coarse to fine search strategy , simple and efficient method, both real-time and accuracy can be processed; (4) the aspect of the image is extracted a plurality of parallel two-dimensional code, the present invention is therefore particularly suitable for parallel processing system having a hardware and software capabilities.

附图说明[0020] 图1是本发明所述方法的系统示意框图。 BRIEF DESCRIPTION [0020] FIG. 1 is a schematic block diagram of the system of the present invention method.

[0021] 图2是本发明在图像金字塔中进行图像分块的示意图(图中,上方为原图,下方为三个尺度下图像的分块)。 [0021] FIG. 2 is a schematic diagram of the image block of the present invention in an image pyramid (FIG., The top of the original image, to below the block image in three dimensions).

[0022] 图3是本发明方法在光照不均勻情况下的二值化结果图。 [0022] FIG. 3 is a method of the present invention, FIG binarization result in uneven lighting conditions.

[0023] 图4是本发明中分块搜索二维码区域示意图。 [0023] FIG. 4 is a schematic view of a search region of the two-dimensional code carved block of the present invention.

[0024] 图5是本发明提取二维码区域的结果图。 [0024] FIG. 5 is a two-dimensional code region extraction result of the present invention FIG.

具体实施方式 Detailed ways

[0025] 下面结合附图具体描述本发明的实施方式。 [0025] The following detailed description of embodiments of the present invention in conjunction with the accompanying drawings.

[0026] 参照图1,在图像中分割提取二维码区域的方法步骤如下: [0026] Referring to FIG 1, the divided two-dimensional code in the image region extraction steps are as follows:

[0027] 步骤一:读取包含有二维码的图像,将其转化成256级的灰度图像。 [0027] Step a: read the two-dimensional code with an image converted into gray scale image 256 is. 然后建立(如图2中下方示出的三个图像分块)的图像金字塔,层数设置范围为3〜5层,建立方法是从原始尺度出发,利用高斯函数生成的核进行滤波,再进行隔行隔列的降采样,得到上一层的图像; And establishing (three image block shown below in FIG. 2) of the image pyramid, the number of layers set in the range of three to five layers, starting from the original method of establishing scale, nuclear generated Gaussian function filtering, then downsampling the interlaced every other column to obtain a layer of an image;

[0028] 步骤二:对金字塔中的每一层图像进行分块,根据每个分块图像的局部灰度分布信息,采用最大类间方差方法对每个分块图像进行二值化。 [0028] Step two: for each layer of the pyramid image block, according to the local distribution information of each gradation patch image using Otsu method for each image block is binarized. 具体实现步骤如下所述: The specific steps are as follows:

[0029] (1)设置图像分块宽度(高度)与图像宽度(高度)的比值,即比例系数RO(设置R的范围0. 125〜0. 05),图像中二维码尺寸越大,R的设置值越大。 [0029] (1) Set the image block width (height) of the image width (height) ratio, i.e. the ratio of the coefficient RO (range R set 0. 125~0. 05), the larger the image size of two-dimensional code, the larger the value of R. 将金字塔最底层的图像按照比例R进行分块,如图2(b)中的虚线框所示; The bottom of the pyramid image into blocks according to the ratio R, as shown in dashed box 2 (b) in FIG;

[0030] (2)对金字塔的上一层图像进行分块,分块的比例系数Rl取为上一层的两倍,即Rl = 2 X RO ; [0030] (2) the upper layer of the image pyramid is divided into blocks, the block scale factor Rl taken twice the upper hierarchy, i.e. Rl = 2 X RO;

[0031] (3)重复上一步,直至图像金字塔的最顶层; [0031] (3) Repeat the previous step until the top level of the image pyramid;

[0032] (4)利用最大类间方差算法,计算出图像金字塔每层图像中的每个分块的分割阈值,然后使用阈值对相应的分块进行二值化处理,得到每层的二值图像B0,Bi,…,Βη,η为金字塔的层数; [0032] (4) using the Otsu algorithm to calculate a threshold value for each block division image pyramid image of each layer, and then using a threshold value corresponding to the block is binarized to give each binary image B0, Bi, ..., Βη, η is a pyramid layers;

[0033] 步骤三:通过多尺度融合方法减少图像分块所引起分块效应,实施方法是先将步骤二获得的二值图像Β0,Β1,…,Βη,通过内插使之都达到原始图像的分辨率,得到图像10, II,...,In ;然后采用取像素中值的方式融合所有尺度下得到的二值图像10,II, ... In, 生成融合图像Ia,即图像Ia在图像坐标(X,y)处的像素值Ia(x, y)为IO (x, y),Il (χ, y), ...In (χ, y)的中值。 [0033] Step Three: Multi-scale by reducing the image block fusion blocking effect caused by the implementation of the method is a binary image obtained in step two first Β0, Β1, ..., Βη, by interpolation of the original image so as to have reached resolution, to obtain the image 10, II, ..., in; and using pixel values ​​in a manner taking the binary image obtained by fusion at all scales 10, II, ... in, generating a fusion image Ia, i.e., the image Ia the image coordinates (X, y) pixel value, Il (χ, y), ... in (χ, y) is the value of the IO (x, y) Ia (x, y). 如图3所示,为多尺度融合后的效果图,从图中可以看出,仅在图像边缘部分有少量的分块效应; 3, is a view of the effect of the multi-scale integration, it can be seen from the figure, only a small amount of blocking effects in the image edge portion;

[0034] 步骤四:按照图像金字塔最底层的分块方式,在图像χ轴方向上逐行扫描Ia的每个图像块。 [0034] Step Four: the bottom block in accordance with the image pyramid manner, χ axis direction of the image on each image block progressive Ia. 扫描过程中,若发现相邻像素的像素值不同,则记为一次跳变,统计每个块中出现的跳变次数。 During the scan, if found different neighboring pixel values, is referred to as a jump, hop count variable number appearing in each block. 设置跳变次数的阈值,根据二维码具有丰富纹理的特点,将跳变次数大于阈值的块标记为二维码候选块,候选块作为后续二维码搜索区域; Setting a threshold number of transitions, the two-dimensional code having a rich texture characteristics, the hop number becomes greater than a threshold value, the block flag is a two-dimensional code candidate block as the candidate block search subsequent two-dimensional code area;

[0035] 步骤五:以图像块为单位进行连通体标记,找出所有在图像上相邻接的候选块,实施方法是逐行扫描属于候选块的图像块,若当前块与其8连通的邻接块(在左、左上、上、 右上四个方向上判断)均为候选块,则将它们标记为同一标签值,扫描完成后具有相同标签值的块构成一个连通体区域。 [0035] Step Five: an image block units connected component marking, to find all candidate blocks adjacent in the image, an image block is to implement a progressive scan belonging to the candidate block, if the current block and its adjacent communication 8 block (left, upper left, upper, upper right determines the four directions) are candidate blocks, they are labeled as the same label value, the scan is complete a block having the same tag value constituting a connected component region. 由于二维码通常为正方形,当拍摄视角不是很偏斜的情况下,二维码在图像上仍接近与正方形。 As the two-dimensional code generally square, in the case where the shooting is not very skewed perspective, two-dimensional code in the image is still close to the square. 考察连通体的外接矩形的长宽比LWR,若0. 3 < LffR < 0. 7,则该连通体为二维码区域,否则为背景; Investigation vias LWR of the aspect ratio of the circumscribed rectangle, if 0. 3 <LffR <0. 7, even the whole body is the two-dimensional code area, or the background;

[0036] 步骤六:在每个二维码连通体的边缘带中检测出Harris角点特征。 [0036] Step Six: Harris corner point features detected in each of the two-dimensional code with an edge of vias. 如图4中的实线框所示,考虑到二维码边角部分的像素可能会落在连通体之外,在设置的边缘带中对连通体进行扩张。 As shown in FIG. 4 in solid boxes, considering the corner portion of the pixel two-dimensional code connected component may fall outside of the edge band is provided to expand the vias. 边缘带包含两个部分:即内边缘和外边缘。 Edge strip consists of two parts: the inner and outer edges. 内边缘是由连通体中与非候选块相邻接的图像块构成,外边缘则由与该连通体相邻接的非候选块组成。 The inner edge of the image block is constituted by vias with non-candidate block adjacent to the non-contact candidate block adjacent to the outer edge by vias composition. 外边缘中包含二维码部分边角部分的像素。 A pixel portion comprising an outer edge corner part of two-dimensional code. 根据边缘带中检测到的角点,建立特征角点集,采用Sklansky 算法计算出角点集的外接凸包,输出凸包所对应的图像区域。 The corner points detected edge band, establishing set of feature points, using an external corner Sklansky algorithm convex hull point set, the output image corresponding to the convex hull region. 如图5所示,为提取出的二维码区域结果图,其中虚线标记为外接凸包。 5, the two-dimensional code to extract the results of FIG region, wherein the dotted line labeled as the external convex hull.

Claims (4)

  1. 1. 一种鲁棒的图像中二维码区域提取方法,其特征是:包括以下步骤:1.1)对采集的二维码图像进行灰度化处理,建立一种多尺度的高斯图像金字塔;1. 2)对金字塔中的每一层进行分块,根据每个分块图像的局部灰度分布信息,采用最大类间方差方法对每个分块图像进行二值化;1. 3)根据图像金字塔中每一层的二值化结果,对多个尺度上的二值图像进行融合,得到最终的二值图像;1.4)在图像金字塔的最底层,即原始尺度下,逐行扫描每个分块二值图像;扫描过程中,若发现相邻像素的像素值不同,则称为一次跳变,计算每个分块二值图像中出现的跳变次数;根据二维码具有丰富纹理的特点,将跳变次数大于某个阈值的分块设为二维码候选块,候选块作为后续二维码搜索区域;1. 5)以分块为单位进行连通体分析,找出所有在图像上相邻接的候选块,建 1. A Robust two-dimensional code region extraction method for an image, characterized in that: comprising the steps of: 1.1) a two-dimensional code image captured gradation processing, a Gaussian image pyramid to establish a multi-scale; 1 . 2) of each layer of the pyramid is divided into blocks according to the local distribution information of each gradation patch image using Otsu method for each image block is binarized;. 13) the image pyramid binarization result of each layer, the binary image on a plurality of scales are fused to give a final binary image; 1.4) in the bottom of the image pyramid, i.e. the original scale, each progressive scan points binary image block; scanning process, if found adjacent pixel values ​​are different, then called a transition, calculating the number of hops across the image change each binary block; rich texture features having a two-dimensional code , the jump becomes greater than a certain threshold number of dimensional code candidate block to block, as a follow-up two-dimensional code candidate block search area;. 15) to perform connected component analysis block as a unit, to find all the images candidate block adjoining built 多个连通体区域,每个连通体区域则作为候选二维码区域;根据候选二维码区域的形状特征,确定包含有二维码的连通体区域;1.6)以图像块为单位,选取包含有二维码的连通体区域的内边缘块和外边缘块,从而生成边缘带;通过Harris角点检测得到边缘带中的特征角点,建立特征角点集;采用Sklansky算法计算出每个角点集的外接凸包;提取每个外接凸包所对应的图像区域,得到最终的二维码区域。 A plurality of connected component regions, each region of the connected component as two-dimensional code candidate region; wherein the shape of the two-dimensional code candidate region, the region is determined connected component comprising a two-dimensional code; 1.6) in image block units, comprising selecting two-dimensional code block and an inner edge of the outer edge block connected component region, thereby generating the edge band; Harris corner obtained by detecting feature points in the edge band, establishing set of feature points; Sklansky algorithm employed to calculate the angle for each external set of convex hull points; each of the extracted image region corresponding to the convex hull external obtain a final two-dimensional code area.
  2. 2.根据权利要求1所述的一种鲁棒的图像中二维码区域提取方法,其特征是:所述步骤1. 2)中对图像金字塔中的每一层进行分块的步骤如下:2. 1)按照设定的比例系数,将最底层的原始图像,划分成适当大小的图像块;其中比例系数为分块的宽度与图像宽度的比值;2. 2)对金字塔上一层图像进行分块,分块的比例系数取上一层比例系数的两倍;2. 3)重复步骤2. 2),直至到图像金字塔最顶层。 Image region extracting method of a robust two-dimensional code of claim 2 according to claim 1, wherein: said step 1.2) in step into blocks each layer of the image pyramid is as follows: 2.1) setting the scale factor, the bottom of the original image, the image is divided into blocks of appropriate size; wherein the scale factor of the image width ratio of the width of the block;. 22) on the layer of the image pyramid divided into blocks, the block scale factor is taken twice the scale factor of one;. 23) repeating steps 2.2), until the top level of the image pyramid.
  3. 3.根据权利要求1所述的一种鲁棒的图像中二维码区域提取方法,其特征是:所述步骤1. 2)中对图像二值化处理的方法是:利用最大类间方差算法,计算出图像金字塔每层图像中的每个分块的分割阈值,然后使用阈值对相应的分块进行二值化处理,从而得到每层的二值图像;通过融合不同尺度下的二值图像,得到最终的二值化结果;融合的实现方法是先将所有尺度下的二值图像,使用内插生成与原始图像等分辨率的二值图像,图像对齐后按像素取中值作为所需的二值化结果。 The region extraction method, a robust two-dimensional code image according to claim 1, characterized in that: the method step 1.2) to the image binarization process is: the use Otsu algorithm to calculate a threshold value for each block division image pyramid image of each layer, and then use the corresponding sub-threshold binarization processing block, whereby each binary image; values ​​by fusing two different scales image, to obtain the final binarization result; fusion implemented binary image is first binarized image at all scales, using interpolation to generate the original image resolution and the like, by taking the pixel values ​​as the image alignment binarization result desired.
  4. 4.根据权利要求1所述的一种鲁棒的图像中二维码区域提取方法,其特征是:所述步骤1. 5)中所述的连通体区域是相邻候选块之间的八连通区域,每个连通体区域作为一个候选二维码区域,依据该区域外接矩形的长宽比规则,确定包含有二维码的连通体区域。 The region extraction method, a robust two-dimensional code image according to claim 1, wherein: said step 1.5) in the region of eight vias between adjacent candidate blocks communication regions, each region connected component as a two-dimensional code candidate region, based on the aspect ratio of the circumscribed rectangular area rule, comprising determining two-dimensional code connected component regions.
CN 201110193354 2011-07-12 2011-07-12 Robust method for extracting two-dimensional code area in image CN102254144A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 201110193354 CN102254144A (en) 2011-07-12 2011-07-12 Robust method for extracting two-dimensional code area in image

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 201110193354 CN102254144A (en) 2011-07-12 2011-07-12 Robust method for extracting two-dimensional code area in image

Publications (1)

Publication Number Publication Date
CN102254144A true true CN102254144A (en) 2011-11-23

Family

ID=44981400

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 201110193354 CN102254144A (en) 2011-07-12 2011-07-12 Robust method for extracting two-dimensional code area in image

Country Status (1)

Country Link
CN (1) CN102254144A (en)

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102693409A (en) * 2012-05-18 2012-09-26 四川大学 Method for quickly identifying two-dimension code system type in images
CN103235949A (en) * 2013-04-12 2013-08-07 北京大学 Method and device for detecting points of interest in images
CN103625649A (en) * 2013-12-06 2014-03-12 北京工商大学 Aircraft autonomous landing region judging method based on self adaptive region division and window communication
CN104050476A (en) * 2014-06-23 2014-09-17 北京理工大学 Method for selecting target aiming point on tail section based on convex hull calculation
CN104363400A (en) * 2014-10-29 2015-02-18 广东欧珀移动通信有限公司 Method and device for scanning local two-dimensional codes
CN104463067A (en) * 2014-12-04 2015-03-25 四川大学 Method for extracting macro blocks of Grid Matrix two-dimensional bar code
CN104899589A (en) * 2015-05-12 2015-09-09 广州中大数码科技有限公司 Method for realizing two-dimensional bar code preprocessing by using threshold binarization algorithm
CN105225236A (en) * 2015-09-21 2016-01-06 中国科学院半导体研究所 Binary image connected region parallel detection method and system
CN105574897A (en) * 2015-12-07 2016-05-11 中国科学院合肥物质科学研究院 Crop growth situation monitoring Internet of Things system based on visual inspection
CN105574898A (en) * 2015-12-07 2016-05-11 中国科学院合肥物质科学研究院 Method and system for monitoring plant lodging situation based on image detection
CN106778723A (en) * 2016-11-28 2017-05-31 华中科技大学 Surface image extraction method of wind turbine blade in complex background environment
CN107181938A (en) * 2016-03-11 2017-09-19 深圳超多维光电子有限公司 Image display method and equipment, and image analysis method, apparatus and system

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101093544A (en) * 2007-06-14 2007-12-26 中兴通讯股份有限公司 Method for correcting pattern in 2D code of perspective quick response matrix

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101093544A (en) * 2007-06-14 2007-12-26 中兴通讯股份有限公司 Method for correcting pattern in 2D code of perspective quick response matrix

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
林志毅等: "基于场景分类及灰度跳变的车牌定位方法", 《交通科技》, no. 2, 30 April 2006 (2006-04-30) *
潘道远等: "基于多方法融合的文本定位算法的研究", 《计算机应用与软件》, vol. 27, no. 6, 30 June 2010 (2010-06-30) *
蔡文婷: "移动端二维条码图像增强及应用研究", 《浙江工业大学硕士学位论文》, 15 June 2009 (2009-06-15) *
郝衎等: "基于图像颜色直方图及纹理特征提取的兴趣点凸包检索方法", 《微型电脑应用》, vol. 27, no. 6, 20 June 2011 (2011-06-20) *

Cited By (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102693409B (en) 2012-05-18 2014-04-09 四川大学 Method for quickly identifying two-dimension code system type in images
CN102693409A (en) * 2012-05-18 2012-09-26 四川大学 Method for quickly identifying two-dimension code system type in images
CN103235949A (en) * 2013-04-12 2013-08-07 北京大学 Method and device for detecting points of interest in images
CN103235949B (en) * 2013-04-12 2016-02-10 北京大学 The image detecting method and apparatus for interest point
WO2014166377A1 (en) * 2013-04-12 2014-10-16 北京大学 Image interest point detection method and device
US9779324B2 (en) 2013-04-12 2017-10-03 Peking University Method and device for detecting interest points in image
CN103625649A (en) * 2013-12-06 2014-03-12 北京工商大学 Aircraft autonomous landing region judging method based on self adaptive region division and window communication
CN104050476A (en) * 2014-06-23 2014-09-17 北京理工大学 Method for selecting target aiming point on tail section based on convex hull calculation
CN104363400A (en) * 2014-10-29 2015-02-18 广东欧珀移动通信有限公司 Method and device for scanning local two-dimensional codes
CN104463067B (en) * 2014-12-04 2017-03-22 四川大学 One kind Grid Matrix two-dimensional bar macro module extraction method
CN104463067A (en) * 2014-12-04 2015-03-25 四川大学 Method for extracting macro blocks of Grid Matrix two-dimensional bar code
CN104899589A (en) * 2015-05-12 2015-09-09 广州中大数码科技有限公司 Method for realizing two-dimensional bar code preprocessing by using threshold binarization algorithm
CN104899589B (en) * 2015-05-12 2018-10-12 广州中大数码科技有限公司 Using one kind of threshold binarization algorithm implemented method of two-dimensional bar pretreatment
CN105225236A (en) * 2015-09-21 2016-01-06 中国科学院半导体研究所 Binary image connected region parallel detection method and system
CN105574898A (en) * 2015-12-07 2016-05-11 中国科学院合肥物质科学研究院 Method and system for monitoring plant lodging situation based on image detection
CN105574897A (en) * 2015-12-07 2016-05-11 中国科学院合肥物质科学研究院 Crop growth situation monitoring Internet of Things system based on visual inspection
CN107181938A (en) * 2016-03-11 2017-09-19 深圳超多维光电子有限公司 Image display method and equipment, and image analysis method, apparatus and system
CN106778723A (en) * 2016-11-28 2017-05-31 华中科技大学 Surface image extraction method of wind turbine blade in complex background environment

Similar Documents

Publication Publication Date Title
Deng et al. Color image segmentation
Yi et al. Text string detection from natural scenes by structure-based partition and grouping
Zhao et al. A closed-form solution to retinex with nonlocal texture constraints
Lalonde et al. Detecting ground shadows in outdoor consumer photographs
Shivakumara et al. A laplacian approach to multi-oriented text detection in video
Pan et al. Detecting image region duplication using SIFT features
Shi et al. Scene text detection using graph model built upon maximally stable extremal regions
Zhang et al. Text extraction from natural scene image: A survey
Jung et al. A stroke filter and its application to text localization
Wang et al. Character location in scene images from digital camera
Yi et al. Text extraction from scene images by character appearance and structure modeling
CN101739561A (en) TV station logo training method and identification method
Clark et al. Rectifying perspective views of text in 3D scenes using vanishing points
US20110290882A1 (en) Qr code detection
Lu et al. Salient object detection using concavity context
Zhao et al. Detecting digital image splicing in chroma spaces
US20130129216A1 (en) Text Detection Using Multi-Layer Connected Components With Histograms
CN104156704A (en) Novel license plate identification method and system
CN101593277A (en) Method and device for automatically positioning text region in complicated color image
Wang et al. A sliding window technique for efficient license plate localization based on discrete wavelet transform
İlsever et al. Two-dimensional change detection methods: Remote sensing applications
Jaberi et al. Accurate and robust localization of duplicated region in copy–move image forgery
Park Shape-resolving local thresholding for object detection
CN102013008A (en) Smoke detection method based on support vector machine and device
Wang et al. AprilTag 2: Efficient and robust fiducial detection.

Legal Events

Date Code Title Description
C06 Publication
C10 Entry into substantive examination
C12 Rejection of a patent application after its publication