TWI567655B - Object of two - dimensional code discrimination method - Google Patents

Description

Two-dimensional code recognition method for objects

The invention relates to image recognition; in particular, to a two-dimensional code discrimination method for an object.

The two-dimensional code is a black and white rectangular data point indicating the encoded information. When the two-dimensional code is read, the scanning device scans the two-dimensional code and decodes it to obtain the information represented by the two-dimensional code. The application of the two-dimensional code is quite extensive, and the object using the two-dimensional code is commonly used with paper and display screens. The two-dimensional code is also widely used in manufacturing, for example, a corresponding two-dimensional code is set on the produced object to be used as a production history tracking of the object.

However, if the two-dimensional code pattern on the object is blurred or deformed, the decoding of the two-dimensional code may fail. The black plastic material is used as a column. After the two-dimensional code is laser-engraved on the object, the temperature generated during the laser engraving will soften the plastic material, resulting in the data point on the two-dimensional code graphic. Deformation occurs, so the decoding of the QR code on the object is prone to failure. Secondly, the object is black, not easy to reflect, and the difficulty of decoding is increased.

In view of this, the object of the present invention is to provide a two-dimensional code discrimination method for an object, which can effectively improve the probability of successful decoding of the two-dimensional code on the object.

In order to achieve the above object, the present invention provides a two-dimensional code discrimination method for an object, wherein the object has a two-dimensional code, and the method comprises the following steps:

A. capturing an image of the two-dimensional code of the object;

B. Perform gamma adjustment on the image of step A;

C. Smoothing, edge detection, and grayscale contrast enhancement of the image processed in step B;

D. Perform image segmentation on the image after step C grayscale contrast enhancement;

E. Perform image binarization processing, erosion treatment, and expansion processing on the image processed in step D;

F. Perform object analysis on the image after the step E expansion process to obtain a feature image;

G, performing erosion processing and expansion processing on the characteristic image to obtain a first image;

H, performing a Dot Matched Filter on the first image to obtain a second image;

I, after subtracting the image from the first image and the second image, performing grayscale contrast enhancement to obtain a third image;

J. Decoding the two-dimensional code for the third image.

The effect of the invention is that a series of image processing is performed on the original image of the object, which can effectively improve the lack of decoding caused by the two-dimensional code blur in the prior art, and improve the probability of successful decoding of the two-dimensional code on the object.

In order to explain the present invention more clearly, a preferred embodiment will be described in detail with reference to the drawings. Referring to FIG. 1 , an image processing system for applying a two-dimensional code identification method for an object according to a preferred embodiment of the present invention includes a camera 10 , an image processing device 20 , a reflector 30 and a light source . 40. The image viewing system is configured to recognize a two-dimensional code on the object of the lens barrel 50. In the embodiment, the lens barrel 50 is made of black plastic, and the two-dimensional code is laser. Engraved on the back of the lens barrel 50 for production history tracking.

The camera 10 is used to capture an image of the lens barrel 50. The image processing device 20 is a computer and is electrically connected to the camera 10 for receiving the image captured by the camera 10, performing image processing and discriminating the two-dimensional code. In this embodiment, the image processing algorithm library It is used by Matrox Imaging Library MIL 9.0. The reflector 30 has a first opening 302 and a second opening 304. The reflector 30 is gradually expanded from the first opening 302 toward the second opening 304. The inner surface of the reflector 30 is The curved surface is used to reflect light toward the second opening 304. The light source 40 is used to provide light. The light source is an annular light source in this embodiment, but is not limited thereto, and may be composed of a plurality of light emitting elements.

The two-dimensional code discrimination method of the object of the embodiment includes the following steps as shown in FIG. 2 and FIG. 3:

The camera 10, the reflector 30, and the light source 40 are mounted. The lens 102 of the camera 10 is placed in the first opening 302 of the reflector 30, and the light source 40 is disposed inside the reflector 30 and surrounds the second opening 304 to illuminate the light emitted by the light source 40. The inner surface of the reflector 30 is reflected by the inner surface of the reflector 30 and is emitted toward the second opening 304.

The lens barrel 50 is placed at a position corresponding to the second opening 304 of the reflector 30 such that the back of the lens barrel 50 faces the second opening 304. Thereby, the light emitted by the light source 40 can illuminate the corresponding area of the second opening 304 to obtain more light on the back of the lens barrel 50.

The image processing device 20 activates the camera 10 to capture the image of the lens barrel 50, and the captured image is transmitted to the image processing device 20. The original image is as shown in FIG. 4, and the image of the lens barrel 50 includes the second image. The image of the dimension code 52, the image of the two-dimensional code 52 is composed of a plurality of data points, each of which includes a plurality of pixels.

Next, the image processing device 20 performs the following steps of image processing:

The gamma value of the original image captured is adjusted to change the dynamic range of the original image, so that the area of the two-dimensional code 52 can be found, and the adjusted gamma value ranges from 1.1 to 2.2. In this embodiment, the gamma value is preferably 2.19, and the adjusted image is as shown in FIG. 5.

The global domain of the image after adjusting the gamma value is continuously smoothed (Smoothing), edge detected, and grayscale contrast enhanced. The gray scale contrast enhancement parameters are: the gray scale difference magnification of the image is 1.0~3.0, the gray scale difference of the image is -128~128, and the gray scale enhancement of the image is 0.01~0.2. Preferably, the image grayscale difference is 1.3, the image grayscale difference is -42, and the image grayscale enhancement is 0.032.

The image segmentation is performed on the grayscale contrast enhanced image. The purpose of the image segmentation is to segment the locally connected regions in the two-dimensional code 52 in the image to highlight the features of the two-dimensional code 52, which facilitates subsequent object analysis. In this embodiment, the image is segmented by a watershed algorithm. The parameters of the watershed algorithm are: Watershed Variation is 1~255, and Watershed Threshold is 1~255. Preferably, the Watershed Variation is 13, and the Watershed Threshold is 60-255. The image after image segmentation is shown in FIG. 6.

The image after the image segmentation process is sequentially subjected to image binarization processing, erosion processing, and expansion processing to remove some of the noise in the image.

Then, object analysis is performed on the image after the expansion process to obtain a feature image of the two-dimensional code 52. In this embodiment, the image is imaged by a binary large object analysis (blob analysis) algorithm. Analysis, the purpose of the image processed in the previous step may still be a little bit of residual, by object analysis to select the data points belonging to the QR code. The parameters of the binary target analysis algorithm are: Blob Area is 5000~25000, Blob FeretX range is 30~180, and Blob FeretY range is 30~180. The preferred Blob Area is 5000~12000, the Blob FeretX range is 50~150, and the Blob FeretY range is 50~150.

Performing erosion processing and expansion processing on the feature image to obtain a first image, and then performing a matched filter on the first image to obtain a second image, wherein the purpose of the matched filtering is to enhance the image. The feature of the data point of the two-dimensional code 52 suppresses the noise of the non-data point.

Finally, performing image subtraction processing, subtracting the second image from the first image to enhance contrast of the image, and performing grayscale contrast enhancement to obtain a third image. In this step, the grayscale contrast enhancement is mainly for image processing of the binary code 52. The grayscale contrast enhancement parameters are: the grayscale difference magnification of the image ranges from 1.0 to 3.0, and the grayscale difference of the image is -128 to 128. The image grayscale enhancement is 0.01~0.2. Preferably, the image grayscale difference is 1.3, the image grayscale difference is -42, and the image grayscale enhancement is 0.032. The area of the two-dimensional code 52 in the third image is as shown in FIG.

After the step of performing the image processing, the image processing device 20 performs decoding of the two-dimensional code 52 on the region of the two-dimensional code 52 in the third image, and determines whether the decoding is successful:

If so, the information included in the two-dimensional code 52 is obtained, and the identification method is ended.

If not, the image processing device 20 further performs the following steps (refer to FIG. 3) for the third image, wherein:

The image processing device 20 reduces the third image to a predetermined number of image level images by an image pyramid algorithm, and defines one of the image size levels to be a first image level. The image of the second largest image is the image of the second image level. In this embodiment, the length and width of the third image are each reduced by half by a Gaussian image pyramid to form an image of the first image level, and the length and width of the image of the second image level are the length and width of the image of the first image level. Half of this, pushed by this column. Since the image is reduced, the data point of the two-dimensional code 52 in the image of the first image level is reduced, so that the noise points in the image are also reduced and are not easily judged as belonging to the data point of the two-dimensional code 52.

Decoding the two-dimensional code 52 on the image of the first image level, and determining whether the decoding is successful:

If so, the information included in the two-dimensional code 52 is obtained, and the identification method is ended.

If no, proceed to the following steps, where:

Adjusting the spacing between the data points of the two-dimensional code 52 within an allowable range to form a fourth image, and continuing to decode the two-dimensional code 52 for the fourth image, and determining whether the decoding is successful:

If so, the information included in the two-dimensional code 52 is obtained, and the identification method is ended.

If not, decoding the two-dimensional code 52 on the image of the second image level, and determining whether the decoding is successful:

If so, the information included in the two-dimensional code 52 is obtained, and the identification method is ended.

If not, the distance between the data points of the two-dimensional code 52 is further adjusted, and then decoded. If the decoding is unsuccessful, the two-dimensional code 52 is decoded or adjusted at least once for the image of the next image level image. The decoding after the spacing of the data points of the two-dimensional code 52 is until the decoding is successful or the decoding of the data point of the two-dimensional code 52 is performed after the decoding of the image of the last image level is performed. If the number of layers in all image levels cannot be decoded after performing the above steps, it is determined that the two-dimensional code 52 cannot be recognized, and the identification method is ended.

According to the above description, the present invention can effectively improve the lack of decoding caused by the blurring or deformation of the two-dimensional code in the prior art by performing a series of image processing processes on the original image of the object, and improve the decoding of the two-dimensional code on the object. The chance. In addition, the present invention reflects the light emitted by the light source by the reflector, and can illuminate the light source on the two-dimensional code of the object to make the image of the two-dimensional code clearer and increase the chance of successful decoding.

The above is only a preferred embodiment of the present invention, and equivalent changes to the scope of the present invention and the scope of the patent application are intended to be included in the scope of the present invention.

[this invention]
10‧‧‧ camera
102‧‧‧ lens
20‧‧‧Image processing device
30‧‧‧Reflector
302‧‧‧ first opening
304‧‧‧second opening
40‧‧‧Light source
50‧‧‧Mirror tube
52‧‧‧ QR code

1 is a schematic diagram of an image viewing system according to a preferred embodiment of the present invention. 2 and 3 are flow charts of a two-dimensional code discrimination method for the object of the above preferred embodiment. Figure 4 shows the original image of the lens barrel. Fig. 5 is an image of a lens barrel adjusted by a gamma value. FIG. 6 is an image of a lens barrel after image segmentation by a water division algorithm. Figure 7 is an image of a region of the two-dimensional code in the third image.

Claims (17)

A two-dimensional code discriminating method for an object, wherein the object has a two-dimensional code, the method comprising the following steps: A: capturing an image of the two-dimensional code of the object; B, performing gamma value on the image of step A Adjusting; C. Smoothing the image processed in step B, edge detection, and grayscale contrast enhancement; D. Perform image segmentation on the image after step C grayscale contrast enhancement; E. After processing step D The image is sequentially subjected to image binarization processing, erosion processing, and expansion processing; F. Performing object analysis on the image after step E expansion processing to obtain a feature image; G, performing erosion processing and expansion processing on the characteristic image to Obtaining a first image; H, performing matched filtering on the first image to obtain a second image; I, subtracting the image from the first image and the second image, and performing grayscale contrast enhancement to obtain a third image; and J, decoding the two-dimensional code for the third image. The method of claim 2, wherein the step J comprises determining whether the decoding of the two-dimensional code is successful, and if not, performing the following steps: K. using an image pyramid algorithm, the first The three images are reduced to an image of a first image level; L. The image of the first image level is decoded by the two-dimensional code. The method of claim 2, wherein the reduced image of the first image level has a plurality of data points, the points constitute the two-dimensional code, and the step L comprises: Determining whether the decoding of the two-dimensional code is successful. If not, adjusting the spacing between the data points to form a fourth image, and continuing to decode the two-dimensional code for the fourth image. The method of claim 2, wherein the step K comprises using the image pyramid algorithm to reduce the third image to an image of a second image level, wherein the image size of the second image level is smaller than The image size of the first image level; the step L includes determining whether the decoding of the two-dimensional code is successful, and if not, decoding the two-dimensional code of the image of the second image level. The method for discriminating the two-dimensional code of the object according to claim 1, wherein in step D, the image is segmented by the watershed algorithm after the grayscale contrast enhancement. The method for determining the two-dimensional code of the object according to claim 5, wherein in the step D, the parameters of the watershed algorithm are: Watershed Variation is 1~255, and Watershed Threshold is 1~255. The method according to claim 6, wherein the watershed variation is 13 and the Watershed Threshold is 60 to 255. The two-dimensional code discriminating method for the object according to claim 1, wherein in step F, the object analysis is performed on the image after the step E is expanded by the binary main target analysis algorithm. The method for determining the two-dimensional code of the object according to claim 8, wherein in step F, the parameters of the binary main target analysis algorithm are: Blob Area is 5000~25000, Blob FeretX ranges from 30~180, Blob FeretY The range is 30~180. The method for discriminating the two-dimensional code of the object according to claim 9, wherein in the step F, the Blob Area is 5000 to 12000, the Blob FeretX is 50 to 150, and the Blob FeretY is 50 to 150. The method for identifying a two-dimensional code of the object according to claim 1, wherein in step B, the gamma value ranges from 1.1 to 2.2. A two-dimensional code discriminating method for an object according to claim 11, wherein in step B, the gamma value ranges from 2.19. The method for discriminating the two-dimensional code of the object according to claim 1, wherein in step C, the grayscale contrast enhancement parameter is: the grayscale difference magnification of the image ranges from 1.0 to 3.0, and the grayscale difference of the image is -128~ 128, the image gray scale enhancement is 0.01~0.2. The method for discriminating the two-dimensional code of the object according to claim 13, wherein in step C, the parameters of the grayscale contrast enhancement are: an image grayscale difference magnification of 1.3, an image grayscale difference of -42, and an image grayscale enhancement Is 0.032. The method for discriminating the two-dimensional code of the object according to claim 1, wherein in step I, the grayscale contrast enhancement parameter is: the grayscale difference magnification of the image ranges from 1.0 to 3.0, and the grayscale difference of the image is -128~ 128, the image gray scale enhancement is 0.01~0.2. The method for identifying a two-dimensional code of the object according to claim 15, wherein in step I, the grayscale contrast enhancement parameter is: an image grayscale difference magnification of 1.3, an image grayscale difference of -42, and an image grayscale enhancement. Is 0.032. The method of claim 2, wherein the step A includes: providing a camera, a reflector, and at least one light source, wherein the reflector has a first opening and a second Opening, and the reflector has an inner surface for reflecting light; placing the lens of the camera in the first opening of the reflector, and placing the light source inside the reflector, so that The light emitted by the light source is reflected by the inner surface of the reflector to be emitted to the second opening; the object is placed at a position corresponding to the second opening; and the image of the two-dimensional code of the object is captured by the camera.