KR102614119B1 - Method and apparatus for detecting data matrix - Google Patents

Abstract

The present invention discloses a data matrix recognition method and apparatus. According to the present invention, it includes a processor and a memory connected to the processor, wherein the memory is configured to: (a) apply a first binarization threshold to an RGB image obtained by capturing a plurality of data matrices corresponding to a plurality of columns stored in a cryo box; Apply to generate a binarized image, (b) extract a point set included in the binarized image, and (c) remove adjacent points and concave surfaces from the extracted point set to create a first image corresponding to the shape of the data matrix. extracting a set of shapes, (d) estimating one or more data matrices by applying a preset rule to the extracted first shape set, (e) decoding the estimated one or more data matrices, and (f) the RGB Data storing program instructions executed by the processor to repeatedly perform steps (b) to (e) by applying a binarization threshold sequentially larger than the first binarization threshold to the image to generate a binarization image. A matrix recognition device is provided.

Description

{Method and apparatus for detecting data matrix}

The present invention relates to a data matrix recognition method and device, and more specifically, to a method and device that can obtain information about a plurality of samples at once.

Generally, samples such as microorganisms, antibodies, and bio-synthetic materials are stored in long columns, and a plurality of columns are stored in a cryo box having a plurality of grid-shaped storage holes.

In the past, in order to identify a sample, you had to open the column and perform sample analysis, but recently, a data matrix (2D code) for sample identification is attached to each column cover.

However, in the past, there was a cumbersome problem because each column had to be taken out of the cryo box and the data matrix had to be recognized individually.

Meanwhile, a method to recognize multiple objects stored in a specific space using a data matrix has been proposed, but since it is based on coordinates, there are limitations in applying it to a cryobox for storing small-sized samples.

Republic of Korea Open Patent KR 10-1713134

In order to solve the problems of the prior art described above, the present invention seeks to propose a data matrix recognition method and device that can identify a plurality of samples stored in a cryo box at once.

In order to achieve the above-mentioned object, according to one embodiment of the present invention, there is provided a data matrix recognition device, comprising: a processor; and a memory connected to the processor, wherein the memory (a) applies a first binarization threshold to an RGB image obtained by capturing a plurality of data matrices corresponding to a plurality of columns stored in the cryo box to generate a binarized image. (b) extracting a point set included in the binarized image, (c) extracting a first shape set corresponding to the shape of the data matrix by removing adjacent points and concave surfaces from the extracted point set, (d) estimating one or more data matrices by applying a preset rule to the extracted first shape set, (e) decoding the estimated one or more data matrices, and (f) performing the first binarization on the RGB image. A data matrix recognition device is provided that stores program instructions executed by the processor to repeatedly perform steps (b) to (e) by applying a binarization threshold that is sequentially larger than the threshold to generate a binarized image. .

The program instructions may resize the RGB image to a preset small size before generating the binarized image, and perform inter-pixel chrominance equalization by performing inter-pixel Gaussian blur on the resized RGB image.

The program instructions set minimum and maximum areas for estimating the one or more data matrices by dividing the area of the resized RGB image by a predetermined value, and set the minimum and maximum areas for the extracted first shape set. And one or more data matrices can be estimated by applying a preset horizontal/height ratio.

The program instructions may sequentially change the binarization threshold for generating the binarization image by 5 units in the range of 55 to 155.

The program instructions may pre-store shape coordinates corresponding to the data matrix decoded at the first binarization threshold and not perform estimation of the data matrix on the pre-stored shape coordinates at the next binarization threshold.

According to another aspect of the present invention, there is a method of performing data matrix recognition in a device including a processor and memory, wherein (a) a plurality of data matrices corresponding to a plurality of columns stored in a cryo box are captured in an RGB image. generating a binarized image by applying a first binarization threshold; (b) extracting a point set included in the binarized image; (c) extracting a first shape set corresponding to the shape of the data matrix by removing adjacent points and concave surfaces from the extracted point set; (d) estimating one or more data matrices by applying a preset rule to the extracted first shape set; (e) decoding the estimated one or more data matrices; and (f) generating a binarized image by applying a binarization threshold having a sequentially larger value than the first binarization threshold to the RGB image and repeating steps (b) to (e). A method is provided.

According to another aspect of the present invention, a computer readable program for performing the above method is provided.

According to the present invention, there is an advantage in that a plurality of data matrices included in a cryo box can be quickly and accurately recognized through RGB image resizing, Gaussian blur processing, and sequential binarization.

1 is a diagram illustrating the configuration of a data matrix recognition device according to an embodiment of the present invention.
Figure 2 is a flowchart showing a data matrix recognition process according to an embodiment of the present invention.
Figure 3 is a diagram showing an RGB image stored in another cryo box according to this embodiment and including a data matrix attached to a plurality of columns.
Figure 4 is a diagram showing the result of smoothing the color difference between pixels through Gaussian blur according to this embodiment.
Figure 5 shows a binarized image obtained by sequentially changing the threshold from 55 to 155 in 5 increments.
Figure 6 is a diagram showing the removal process of the decoded data matrix according to this embodiment.
Figure 7 sequentially shows the process of estimating a data matrix from an RGB image according to this embodiment.

Since the present invention can make various changes and have various embodiments, specific embodiments will be illustrated in the drawings and described in detail.

However, this is not intended to limit the present invention to specific embodiments, and should be understood to include all changes, equivalents, and substitutes included in the spirit and technical scope of the present invention.

The present invention relates to a method and device that can accurately and quickly recognize a data matrix attached to a plurality of columns.

When multiple columns are stored in a cryobox, it is not easy to recognize the data matrix attached to each cartridge at once due to the influence of column tilt, illumination, gradient, etc.

For this purpose, the present invention generates a binarized image by sequentially changing the binarization threshold for an image taken with a cryo box, extracts a shape corresponding to the data matrix for the binarized image generated at each threshold, and extracts the extracted shape. Perform decoding on .

1 is a diagram illustrating the configuration of a data matrix recognition device according to an embodiment of the present invention.

Referring to FIG. 1, the data matrix recognition device according to this embodiment may include a processor 100 and a memory 102.

Here, the processor 100 may include a central processing unit (CPU) capable of executing a computer program or another virtual machine.

Memory 102 may include a non-volatile storage device, such as a non-removable hard drive or a removable storage device. Removable storage devices may include compact flash units, USB memory sticks, etc. Memory 102 may also include volatile memory, such as various types of random access memory.

In this memory 102, program instructions executable by the processor 100 are stored.

Program instructions according to this embodiment include (a) generating a binarized image by applying a first binarization threshold to an RGB image obtained by capturing a plurality of data matrices corresponding to a plurality of columns stored in a cryo box, and (b) extracting a point set included in the binarized image, (c) extracting a first shape set corresponding to the shape of the data matrix by removing adjacent points and concave surfaces from the extracted point set, and (d) extracting the first shape set corresponding to the shape of the data matrix. 1 Estimating one or more data matrices by applying a preset rule to a set of shapes, (e) decoding the estimated one or more data matrices, and (f) assigning to the RGB image a value sequentially greater than the first binarization threshold. Steps (b) to (e) are repeated by generating a binarized image by applying a binarization threshold with .

Hereinafter, the data matrix recognition process according to this embodiment will be described in detail with reference to the drawings.

Figure 2 is a flowchart showing a data matrix recognition process according to an embodiment of the present invention.

The device according to this embodiment loads an image taken from the upper direction of the cryo box (step 200).

The image in step 200 is an RGB image stored in a cryo box as shown in FIG. 3 and including a data matrix attached to a plurality of columns.

To improve data matrix recognition speed, the loaded RGB image is resized (step 202).

For example, if the RGB image is 4320×5760 in size, step 202 can fix the width at 1000 and resize it to 1000×1333 by adjusting the height at the same ratio.

According to a preferred embodiment of the present invention, inter-pixel Gaussian Blur processing may be performed on the resized RGB image.

In RGB images taken in a relatively bright environment (high illumination environment), differences between pixels in data matrix color may increase, interfering with recognition.

As shown in FIG. 4, considering that the same color actually has different color differences in an RGB image, the device according to this embodiment can smooth the color difference between pixels through Gaussian blur, which does not affect image quality. This further increases the recognition rate of the data matrix.

A binarized image is generated by applying a first binarization threshold to the resized RGB image (step 204).

Here, the first binarization threshold is a value for converting each pixel of the resized RGB image to 0 or 1, and can be applied equally to R, G, and B, but is not necessarily limited thereto.

The device according to this embodiment decodes at least a portion of a plurality of data matrices by sequentially changing the threshold for binarization. Preferably, the binarization threshold according to this embodiment can be sequentially changed in predetermined units from the lower limit to the upper limit.

Preferably, the lower limit can be set to 55, the upper limit can be set to 155, and the binarization threshold can be sequentially increased by 5 units.

Here, the upper limit is set to 155 because if the data matrix is within RGB(155,155,155) to RGB(255,255,255), the difference cannot be recognized and binarization is difficult.

This is because sequentially changing the threshold for binarization may result in cases where the data matrix of a specific area is not recognized depending on the threshold.

Figure 5 shows a binarized image obtained by sequentially changing the threshold from 55 to 155 in 5 increments.

Referring to FIG. 5, as the threshold increases sequentially, the recognized data matrix changes. For example, at a low threshold, the data matrix located below the relatively low-brightness cryo box of FIG. 3 is recognized, but as the threshold increases, the data matrix is recognized. In Fig. 3, the data matrix located on the cryo box with relatively high brightness can be recognized.

Considering this, the device according to the present embodiment performs the data matrix recognition and decoding process starting from a low threshold, and when applying the next threshold, decoding is performed only on the newly estimated data matrix, excluding the previously decoded data matrix. It will be performed.

Referring again to FIG. 2, a shape set corresponding to the shape of the data matrix is extracted from the binarized image obtained by applying the first binarization threshold (step 206).

Step 206 includes the process of extracting a set of points through contouring, removing adjacent points through approxPolyDP, and removing concave surfaces (concave contour smoothing) through convexHull to extract a shape set corresponding to the shape of the data matrix. .

As shown in FIG. 3, the data matrix may be a square, and step 206 may be defined as a process of searching a set of squares through square approximation in a binarized image obtained by applying a predetermined binarization threshold.

After the shape corresponding to the data matrix is searched in step 206, the device according to this embodiment estimates the data matrix using a preset rule (step 208) and decodes the estimated data matrix (step 210).

If the data matrix is a rectangle, step 206 is a process of searching all rectangle sets recognized in the binarized image, and step 208 is a process of estimating the actual data matrix from the searched rectangle set.

According to this embodiment, the size and horizontal/vertical ratio of the minimum area (minArea) and maximum area (maxArea) for data matrix estimation may be set in advance.

The minimum area and maximum area may be set as a value obtained by dividing the size of the resized RGB image in step 202 by a predetermined value in consideration of the size of a typical data matrix.

If the size of the resized RGB image is A, the minimum area can be set to A/B and the maximum area can be set to A/C.

Here, B=5, C=2000, and in the above example, if the resized RGB image is 1000×1333, the minimum area may be 666 and the maximum area may be 266,000.

And, the width/height ratio can be set to 1.3.

The device according to this embodiment estimates a data matrix by estimating a rectangle that belongs to a preset minimum area and a maximum area among the rectangles extracted in step 206 and has a preset width/height ratio.

After decoding of the estimated data matrix is completed, the data matrix recognition device returns to step 204, applies the next threshold to binarize the RGB image, and repeatedly performs steps 206 to 210.

At this time, the estimation and decoding process for the data matrix decoded in the previous process is omitted, thereby increasing the computation speed.

More specifically, the data matrix recognition device according to the present embodiment pre-stores a rectangular area (shape coordinate) corresponding to the data matrix decoded at the first binarization threshold and pre-stores it at the next binarization threshold, as shown in FIG. 6. Prevent duplicate recognition of the stored rectangular area (decoded data matrix).

At this time, the shape coordinates of the decoded data matrix may be coordinates including a margin (for example, an area expanded by 10% of a rectangular area).

Figure 7 sequentially shows the process of estimating a data matrix from an RGB image according to this embodiment, including resizing the RGB image, extracting a point set from each binarized image, removing adjacent points and concave surfaces, and extracting a shape set corresponding to the data matrix. And the data matrix can be recognized at once through data matrix estimation through minArea/maxArea.

The above-described embodiments of the present invention have been disclosed for illustrative purposes, and those skilled in the art will be able to make various modifications, changes, and additions within the spirit and scope of the present invention, and such modifications, changes, and additions will be possible. should be regarded as falling within the scope of the patent claims below.

Claims (7)

A data matrix recognition device, comprising:
processor; and
Including a memory connected to the processor,
The memory is,
(a) Generating a binarized image by applying a first binarization threshold to an RGB image obtained by capturing a plurality of data matrices corresponding to a plurality of columns stored in a cryo box,
(b) extracting the point set included in the binarized image,
(c) extracting a first shape set corresponding to the shape of the data matrix by removing adjacent points and concave surfaces from the extracted point set,
(d) estimating one or more data matrices by applying preset rules to the extracted first shape set,
(e) decode the estimated one or more data matrices;
(f) generate a binarized image by applying a binarization threshold having a sequentially larger value than the first binarization threshold to the RGB image and repeat steps (b) to (e),
A data matrix recognition device that stores program instructions executed by the processor. According to paragraph 1,
The program commands are:
Before generating the binarized image, resize the RGB image to a preset small size,
A data matrix recognition device that performs inter-pixel color difference equalization by performing inter-pixel Gaussian blur on the resized RGB image. According to paragraph 2,
The program commands are:
Setting a minimum area and a maximum area for estimating the one or more data matrices by dividing the area of the resized RGB image by a predetermined value,
A data matrix recognition device for estimating one or more data matrices by applying the minimum area, maximum area, and preset horizontal/vertical ratio to the extracted first shape set. According to paragraph 1,
The program commands are:
A data matrix recognition device that sequentially changes the binarization threshold for generating the binarization image by 5 units in the range of 55 to 155. According to paragraph 1,
The program commands are:
A data matrix recognition device that stores shape coordinates corresponding to the data matrix decoded at the first binarization threshold in advance and does not perform estimation of the data matrix for the pre-stored shape coordinates at the next binarization threshold. A method of performing data matrix recognition in a device including a processor and memory, comprising:
(a) generating a binarized image by applying a first binarization threshold to an RGB image obtained by photographing a plurality of data matrices corresponding to a plurality of columns stored in a cryo box;
(b) extracting a point set included in the binarized image;
(c) extracting a first shape set corresponding to the shape of the data matrix by removing adjacent points and concave surfaces from the extracted point set;
(d) estimating one or more data matrices by applying a preset rule to the extracted first shape set;
(e) decoding the estimated one or more data matrices; and
(f) a data matrix recognition method comprising repeatedly performing steps (b) to (e) by applying a binarization threshold sequentially larger than the first binarization threshold to the RGB image to generate a binarization image. . A computer program stored in a computer-readable storage medium that performs the method according to claim 6.