Test tube multi-code detection method and device based on neural network, and model training method

Abstract

The invention belongs to the field of IVD image processing, and discloses a test tube multi-code detection method and device based on a neural network, and a model training method. According to the method, a plurality of bar codes in an image are positioned and segmented through a test tube multi-code detection model based on a neural network, so that the plurality of bar codes in a large image can be quickly and efficiently positioned and detected; the segmented image is processed and enhanced, so that the definition and the identification degree of the bar code in the small image are improved, and the decoding accuracy is improved; and a parallel decoding strategy is executed on all small images, so that the decoding efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of IVD image recognition, and in particular to a test tube multi-code detection method and device based on a neural network, and a model training method. Background Art

[0002] In vitro diagnosis, or IVD (In Vitro Diagnosis), refers to a diagnostic method that obtains clinical diagnostic information by testing samples such as human body fluids, cells and tissues in vitro, and then determines the disease or body function. It plays an important role in disease prevention, diagnosis, and treatment. Currently, more than 80% of clinical disease diagnosis can be completed by IVD. It includes sample pre-treatment, multi-row rapid sample injection, multi-turn turntable sample barcode high-speed reading, etc., which are applied to automated test lines, test tube sorting, blood bag management, coagulation, immunity, urine, biochemistry, luminescence platforms, etc.

[0003] Existing barcode recognition technology comes from the traditional logistics and supermarket field. In this scenario, one object has one code, which is relatively simple and the environment is not complicated. However, the barcode recognition environment in the IVD field faces the requirements of a large number of barcodes, a small code area, no missing codes, and fast speed. Among them, the most important is multi-barcode positioning detection. Traditional barcode recognition technology, especially single-type barcode positioning methods, faces the scenario of multiple barcodes in the same image, which is common in IVD detection, and is prone to missing or wrong codes. Summary of the invention

[0004] The purpose of the present invention is to provide a test tube multi-code detection method and device based on a neural network, and a model training method to improve the rapid recognition and detection technology of bar codes on test tubes including tube caps, tube bodies, etc.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] The test tube multi-code detection method based on neural network includes the following steps:

[0007] S1: Load the neural network-based test tube multi-code detection model and parameters;

[0008] S2: Acquire an image to be detected and preprocess the image, wherein the image to be detected includes multiple barcodes;

[0009] S3: inputting the preprocessed image into the test tube multi-code detection model to obtain coordinate information of the positions of multiple barcodes in the image to be detected;

[0010] S4: Cutting the image to be detected into small images containing a single barcode according to the coordinate information of the multiple barcodes;

[0011] S5: performing image enhancement processing on the multiple small images;

[0012] S6: decoding the small images obtained in step S5 after image enhancement processing respectively;

[0013] S7: Output the decoding results of the corresponding small images respectively.

[0014] In step S3, the barcode position coordinate information includes the barcode center origin and the barcode boundary.

[0015] In step S6: the small images are decoded in parallel, that is, all the small images are decoded at the same time.

[0016] In step S6, when the decoding of the small image fails, the barcode whose decoding fails is marked in the image to be detected and displayed.

[0017] A test tube multi-code detection device comprises an image acquisition device, a processor and a memory, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the method as claimed in any one of claims 1 to 4 are implemented.

[0018] The test tube multi-code detection model training method based on neural network includes the following steps:

[0019] A1: Collect sample images;

[0020] A2: Mark all barcode areas in the sample image and record the barcode positioning information;

[0021] A3: Build the original detection model;

[0022] A4: Input the sample image into the original detection model, compare the detection result with the annotation of the sample image, calculate the loss function, adjust the parameters of the original detection model and continue the cyclic training until the training is completed to obtain a test tube multi-code detection model based on a neural network.

[0023] Compared with the prior art, the beneficial effects of the present invention are that multiple barcodes in an image can be located and segmented by a test tube multi-code detection model based on a neural network, so that multiple barcodes in a large image can be located and detected quickly and efficiently; the clarity and recognition of the barcodes in the small image are improved by processing and enhancing the segmented image, thereby improving the decoding accuracy; and a parallel decoding strategy is executed on all small images to improve the decoding efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A flow chart of test tube multi-code detection according to an embodiment of the present invention;

[0025] Figure 2An image to be detected according to an embodiment of the present invention;

[0026] Figure 3 A barcode positioning image according to an embodiment of the present invention;

[0027] Figure 4 A partially segmented small image according to an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0029] Example 1

[0030] like Figure 1 As shown, the test tube multi-code detection method based on neural network includes the following steps:

[0031] S1: Load the neural network-based test tube multi-code detection model and parameters;

[0032] S2: Acquire an image to be detected and preprocess the image, wherein the image to be detected contains multiple barcodes. Figure 2 shown.

[0033] S3: Input the preprocessed image into the test tube multi-code detection model to obtain the coordinate information of the positions of multiple barcodes in the image to be detected, such as Figure 3 shown.

[0034] S4: According to the coordinate information of the multiple barcodes, the image to be detected is cut into small images containing a single barcode, such as Figure 4 shown.

[0035] S5: performing image enhancement processing on the multiple small images;

[0036] S6: decoding the small images obtained in step S5 after image enhancement processing respectively;

[0037] S7: Output the decoding results of the corresponding small images respectively.

[0038] In step S3, the barcode position coordinate information includes the barcode center origin and the barcode boundary.

[0039] In step S6: the small images are decoded in parallel, that is, all the small images are decoded at the same time.

[0040] In step S6, when the decoding of the small image fails, the barcode whose decoding fails is marked in the image to be detected and displayed.

[0041] Example 2

[0042] A test tube multi-code detection device includes an image acquisition device, a processor and a memory, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0043] S1: Load the neural network-based test tube multi-code detection model and parameters;

[0044] S2: Acquire an image to be detected and preprocess the image, wherein the image to be detected includes multiple barcodes;

[0045] S3: inputting the preprocessed image into the test tube multi-code detection model to obtain coordinate information of the positions of multiple barcodes in the image to be detected;

[0046] S4: Cutting the image to be detected into small images containing a single barcode according to the coordinate information of the multiple barcodes;

[0047] S5: performing image enhancement processing on the multiple small images;

[0048] S6: decoding the small images obtained in step S5 after image enhancement processing respectively;

[0049] S7: Output the decoding results of the corresponding small images respectively.

[0050] In step S3, the barcode position coordinate information includes the barcode center origin and the barcode boundary.

[0051] In step S6: the small images are decoded in parallel, that is, all the small images are decoded at the same time.

[0052] In step S6, when the decoding of the small image fails, the barcode whose decoding fails is marked in the image to be detected and displayed.

[0053] Example 3

[0054] The test tube multi-code detection model training method based on neural network includes the following steps:

[0055] A1: Collect sample images;

[0056] A2: Mark all barcode areas in the sample image and record the barcode positioning information;

[0057] A3: Build the original detection model;

[0058] A4: Input the sample image into the original detection model, compare the detection result with the annotation of the sample image, calculate the loss function, adjust the parameters of the original detection model and continue the cyclic training until the training is completed to obtain a test tube multi-code detection model based on a neural network.

[0059] Example 4

[0060] Pipe cap multi-code decoding method based on deep neural network vision: Input the whole image:

[0061] The system receives as input a large image containing multiple barcodes

[0062] Deep network reasoning barcode positioning, pre-positioning the barcodes in the input image through deep neural network vision. After training, the deep neural network can quickly and accurately locate the positions of multiple barcodes in the image.

[0063] The code area coordinates are returned, and the system obtains the barcode coordinate information located by the deep neural network to clarify the position of each barcode in the image.

[0064] Image processing enhancement: Based on the barcode positioning information, the system cuts out a small image area containing each barcode. Image processing enhancement is performed on the small image to improve the clarity and recognition of the barcode.

[0065] The decoding library decodes in parallel, and the small images enhanced by image processing are sent to the decoding library for parallel decoding at the same time. The parallel decoding strategy greatly improves the decoding speed, thereby completing the decoding of multiple barcodes in a short time.

[0066] Output code words, the decoding library outputs the decoding results, that is, the code words corresponding to multiple barcodes.

[0067] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention in the form of a ring-shaped light source. Therefore, the embodiments should be considered exemplary and non-restrictive in every sense, and the scope of the invention is defined by the appended claims rather than the above description, and it is intended that all changes within the meaning and range of equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

[0068] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. Test tube multi-code detection method based on neural network, It is characterized in that The following steps are involved: S1: Load the neural network-based test tube multi-code detection model and parameters; S2: Acquire an image to be detected and preprocess the image, wherein the image to be detected includes multiple barcodes; S3: inputting the preprocessed image into the test tube multi-code detection model to obtain coordinate information of the positions of multiple barcodes in the image to be detected; S4: Cutting the image to be detected into small images containing a single barcode according to the coordinate information of the multiple barcodes; S5: performing image enhancement processing on the multiple small images; S6: decoding the small images obtained in step S5 after image enhancement processing respectively; S7: Output the decoding results of the corresponding small images respectively.

2. The test tube multi-code detection method based on neural network according to claim 1, It is characterized in that In step S3, the barcode position coordinate information includes the barcode center origin and the barcode boundary.

3. The test tube multi-code detection method based on neural network according to claim 1, It is characterized in that In step S6: the small images are decoded in parallel, that is, all the small images are decoded at the same time.

4. The test tube multi-code detection method based on neural network according to claim 1, It is characterized in that In step S6, when the decoding of the small image fails, the barcode whose decoding fails is marked in the image to be detected and displayed.

5. A test tube multi-code detection device, including an image acquisition device, a processor and a memory, It is characterized in that The memory stores a computer program, and the processor implements the steps of the method according to any one of claims 1 to 4 when executing the computer program.

6. A neural network-based test tube multi-code detection model training method. It is characterized in that The following steps are involved: A1: Collect sample images; A2: Mark all barcode areas in the sample image and record the barcode positioning information; A3: Build the original detection model; A4: Input the sample image into the original detection model, compare the detection result with the annotation of the sample image, calculate the loss function, adjust the parameters of the original detection model and continue the cyclic training until the training is completed to obtain a test tube multi-code detection model based on a neural network.

Images