CN103824257A

CN103824257A - Two-dimensional code image preprocessing method

Info

Publication number: CN103824257A
Application number: CN201210465531.9A
Authority: CN
Inventors: 周园
Original assignee: WUXI HANXING ELECTRONIC CO Ltd
Current assignee: WUXI HANXING ELECTRONIC CO Ltd
Priority date: 2012-11-16
Filing date: 2012-11-16
Publication date: 2014-05-28

Abstract

The invention provides a two-dimensional code image preprocessing method comprising the following steps: converting a color image into a gray image, detecting image quality, identifying noise model, carrying out adaptive filtering, segmenting input image data and carrying out binaryzation on each region according to a gray level threshold. The image processing speed and efficiency can be effectively improved for image classification and selection. Filter selection is targeted, the filtering effect is good, the anti-noise-interference ability is strong, and the method is applicable to a wide range of scenes. Processing is regional, the computational complexity can be effectively reduced, the processing quality can be improved, the processing quality can be improved, and the method is applicable to real-time signal processing. Adaptive filtering can be realized by the use of a parallel pipeline structure, the processing speed is high, and the method is applicable to real-time signal processing.

Description

A kind of image in 2 D code preprocess method

Technical field

The present invention relates to image processing method, be specifically related to a kind of image in 2 D code preprocess method.

Background technology

Along with the development of planar bar code technology, its application also becomes more and more extensive, from beverage label to books and reference materials, can see the trace of Quick Response Code from ticket to webpage everywhere.Under numerous application scenarioss, pattern in 2 D code may be by oil stain and dye discoloration, and identification equipment is taken in image with camera, and the image that identification equipment is taken may be by noise pollution, and image may be subject to instrument effect of jitter and thicken.Can recognition correct rate and the recognition speed that improve Quick Response Code be that can many Quick Response Codes apply the key that promote.

Quick Response Code identification generally can be divided into pre-service, and proofread and correct and four steps of decoding, as shown in figure mono-location.Pre-service converts coloured image to bianry image, and location is to determine the position of Quick Response Code in image, and correction is that the pattern that makes to tilt returns back to normal vertical state by mapping, and decoding is that pattern in 2 D code is identified to information extraction data.Pretreated quality directly affects result and the decoding accuracy in follow-up each stage, what traditional preprocess method had directly carries out binaryzation to image, what have carries out simple filtering to image, and antagonism external interference ability is poor, and application scenarios has significant limitation.

Summary of the invention

The object of the invention is to utilize color space conversion, image quality detects, noise pattern identification, and auto adapted filtering, the modules such as subregion binaryzation are carried out real-time pre-service to Quick Response Code; By Intelligent Measurement, filtering and subarea processing improve system rejection to disturbance ability targetedly,, accuracy, processing speed.

The technical scheme addressing the above problem is: a kind of image in 2 D code preprocess method,

Comprise the steps:

Step 1: color space conversion module receives view data, and converts coloured image to gray scale image;

Step 2: the acutance of image quality detection module detected image, with the threshold value comparison of existing sample statistics gained, according to comparative result by Images Classification;

Step 3: the classification of image being done according to step 2, have for all kinds of images of processing;

Step 4: noise pattern identification module is the kind to definite noise according to the statistical property of picture noise and noise sample aspect ratio;

Step 5: auto adapted filtering module selects corresponding wave filter to carry out filtering to the noise detecting, filter coefficient is done corresponding renewal according to the statistical property of view data, output data are sent into region binarization block;

Step 6: region binarization block is cut apart input image data, chooses gray threshold to each region, by gray threshold, binaryzation is carried out in each region, output binaryzation data.

The acutance of image quality detection module detected image, the threshold value comparison with existing sample statistics gained, is divided three classes image according to comparative result: First Kind Graph picture element is poor, after filtering, also cannot decode, and need abandon; Equations of The Second Kind figure picture element is good without filtering processing; The 3rd class image has noise or disturbs and need do filtering processing.

If image is First Kind Graph picture, treating apparatus abandons original image and again reads input data; If image is Equations of The Second Kind image, view data is input to region binarization block by treating apparatus; If image is the 3rd class image, view data is input to noise pattern identification module by treating apparatus;

The invention has the advantages that Images Classification is selected effectively to improve image treatment effeciency and speed; Wave filter is selected targetedly, good wave filtering effect, and anti-noise jamming ability is strong, is suitable for scene extensive; Subarea processing, can effectively reduce computational complexity, improves processing speed, improves Disposal quality, is applicable to real time signal processing; Auto adapted filtering available parallelism pipeline organization realizes, and processing speed is fast, is applicable to real time signal processing.

Below in conjunction with accompanying drawing, the present invention is described in further detail.

Accompanying drawing explanation

Fig. 1 is that image in 2 D code is processed structural representation;

Fig. 2 is Quick Response Code pre-processing structure schematic diagram of the present invention;

The normal QR image in 2 D code of Fig. 3;

Fig. 4 is containing the QR image in 2 D code of Gaussian noise shake;

Fig. 5 uses the binaryzation QR image in 2 D code of the inventive method processing.

Embodiment

In order to deepen the understanding of the present invention, below in conjunction with specific embodiment, the invention will be further described, and this embodiment only, for explaining the present invention, does not form limiting the scope of the present invention.

A kind of image in 2 D code preprocess method provided by the present invention,

Below take the QR image in 2 D code pre-service containing Gaussian noise shake as example accompanying drawings the specific embodiment of the present invention.

As Figure 1-5,

Step 1: color space conversion module receives rgb image data shown in Fig. 4, and converts coloured image to gray scale image, changes as follows:

I=0.2989*R+0.5870*G+0.1140*B

Step 2: the variance of image quality detection module detected image is as image sharpness coefficient, and with setting threshold comparison, this routine mean variance is 0.15, between [0.12 0.18], classifies as the 3rd class image; First Kind Graph picture element is poor, after filtering, also cannot decode, and need abandon; Equations of The Second Kind figure picture element is good without filtering processing;

Step 3: if image is First Kind Graph picture, treating apparatus abandons original image and again reads input data; If image is Equations of The Second Kind image, view data is input to region binarization block by treating apparatus; If image is the 3rd class image, view data is input to noise pattern identification module by treating apparatus;

Step 4: noise pattern identification module computed image intensity profile, search for known intensity profile and comprise two peak values, respectively at gray scale 0 and 255 places, detect gray scale [0 5 10 15] and [255 250 245 240] each point, all there is the spectral line that relative peak numerical value is higher, and spectral line value reduces gradually, hence one can see that image is containing Gaussian noise; It is 256 that numerical value is greater than 100 spectral line number, known image blurring with the contrast of sample statistics result, containing shaking interference.；

Step 5: auto adapted filtering module, according to recognition result, is estimated jitter distance, and angle and noise energy, carry out Wiener filtering to image;

Step 6: region binarization block is cut apart input image data, detects intensity profile to each region, and gray areas is divided into two parts, calculates each region expectation value, and the mean value of getting two expectation values is threshold value, by threshold value, binaryzation is carried out in each region.

The present invention utilizes color space conversion, and image quality detects, noise pattern identification, and auto adapted filtering, the modules such as subregion binaryzation are carried out real-time pre-service to Quick Response Code; By Intelligent Measurement, filtering and subarea processing improve system rejection to disturbance ability, accuracy, processing speed targetedly.

The foregoing is only preferred embodiment of the present invention, in order to limit the present invention, within the spirit and principles in the present invention not all, any modification of doing, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.

Claims

1. an image in 2 D code preprocess method, is characterized in that:

Comprise the steps:

Step 2: the acutance of image quality detection module detected image, with existing sample and threshold value comparison, according to comparative result by Images Classification;

2. image in 2 D code preprocess method according to claim 1, it is characterized in that: the acutance of image quality detection module detected image, with existing sample and threshold value comparison, is divided three classes image according to comparative result: First Kind Graph picture element is poor, after filtering, also cannot decode, need abandon; Equations of The Second Kind figure picture element is good without filtering processing; The 3rd class image has noise or disturbs and need do filtering processing.

3. image in 2 D code preprocess method according to claim 1, is characterized in that: if image is First Kind Graph picture, treating apparatus abandons original image and again reads input data; If image is Equations of The Second Kind image, view data is input to region binarization block by treating apparatus; If image is the 3rd class image, view data is input to noise pattern identification module by treating apparatus.