OCR intelligent identification system and method

Abstract

The invention discloses an OCR (Optical Character Recognition) intelligent recognition system and method, and the system comprises a switch control subsystem which is used for transmitting a trigger signal to an upper computer when identification characters on a product need to be recognized; the upper computer is used for responding to the trigger signal, sending a light source control signal to the light source subsystem, sending a camera control signal to the vision subsystem, and operating an OCR (Optical Character Recognition) program to recognize identification characters in character images transmitted by the vision subsystem; the light source subsystem is used for providing uniform illumination for the vision subsystem in response to the light source control signal; and the visual subsystem is used for responding to the camera control signal, acquiring a character image of the to-be-identified product and sending the character image to the upper computer. By integrating the switch control subsystem, the upper computer, the light source subsystem and the visual subsystem, automatic identification is realized, the time and the cost of manual operation are reduced, the production efficiency is remarkably improved, and meanwhile, the accuracy and the stability of identification are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to an OCR intelligent recognition system and method. Background Art

[0002] In the current hybrid integrated circuit production process, product identification and traceability are important links to ensure product quality and production process management. To achieve this goal, manufacturers usually mark key information such as product batch number and serial number on the product shell by ink printing or laser engraving. However, these identification information currently mainly relies on manual recognition and cannot be quickly and accurately read by automated equipment such as barcode scanners, which greatly limits the improvement of production efficiency and the deepening of automation.

[0003] Although existing automatic optical inspection equipment (AOI) has been widely used in product quality defect detection, it is obviously insufficient in recognizing the identification characters on the circuit casing. Currently, there is a lack of recognition equipment specifically for such application scenarios on the market, resulting in manufacturers still facing challenges in character recognition. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the technical problem to be solved by the present invention is to provide an OCR intelligent recognition system and method capable of automatically recognizing identification characters on a circuit housing.

[0005] In order to solve the above technical problems, a technical solution adopted by the present invention is to provide an OCR intelligent recognition system, comprising:

[0006] The switch control subsystem is used to send a trigger signal to the host computer when the identification characters on the product need to be recognized;

[0007] The host computer is used to respond to the trigger signal, send a light source control signal to the light source subsystem, send a camera control signal to the visual subsystem, and run an OCR recognition program to recognize the identification characters in the character image transmitted by the visual subsystem;

[0008] A light source subsystem, for providing uniform illumination for the visual subsystem in response to a light source control signal;

[0009] The visual subsystem is used to respond to the camera control signal, collect the character image of the product to be identified and send it to the host computer.

[0010] Furthermore, the switch control subsystem includes:

[0011] A push button switch, which is used to be pressed when it is necessary to identify the identification characters on the product;

[0012] The network control IO module is used to send a trigger signal to the host computer when the button switch is pressed.

[0013] Furthermore, the light source subsystem includes:

[0014] Serial port server, used to convert serial port data into network data to realize communication between the host computer and the light source controller;

[0015] A light source controller, used to control the on / off state and brightness of the coaxial light source in response to a light source control signal;

[0016] The coaxial light source is used to provide uniform and stable lighting for the visual subsystem under the control of the light source controller.

[0017] Furthermore, the visual subsystem includes:

[0018] Switch, used to realize communication between the host computer and the industrial camera;

[0019] Industrial lens, used to image the characters on the product to be identified on the photosensitive surface of the industrial camera to achieve light beam transformation;

[0020] The industrial camera is used to respond to the camera control signal, collect the character image of the product to be identified through the industrial lens, convert it into a digital signal and transmit it to the host computer.

[0021] Furthermore, the host computer is also used to set an image recognition area according to the size of the identification characters on the product to be recognized and apply the parameters of the image recognition area to the industrial camera, and the industrial camera collects the character image in the image recognition area.

[0022] Furthermore, the host computer includes:

[0023] A control module, configured to send a light source control signal to the light source subsystem and a camera control signal to the vision subsystem in response to a trigger signal;

[0024] The recognition subsystem is used to run the OCR recognition program in response to the trigger signal to recognize the identification characters in the character image transmitted by the visual subsystem.

[0025] Furthermore, the identification subsystem includes:

[0026] An image processing module, used for preprocessing the character image transmitted by the visual subsystem;

[0027] The image recognition module is used to recognize the image obtained after preprocessing by the image processing module and extract the identification characters therein.

[0028] Furthermore, the image processing module includes:

[0029] The grayscale processing submodule is used to process the character image transmitted by the visual subsystem to obtain a grayscale image;

[0030] Binarization submodule, used to perform binarization processing on the grayscale image to obtain a black and white image;

[0031] The denoising submodule is used to perform denoising on black and white images to remove stains and scratches;

[0032] The image correction submodule is used to detect and correct the tilt angle of the image obtained after noise reduction processing to ensure that the text lines are horizontal or vertical.

[0033] Furthermore, it also includes an organic glass storage table for placing the product to be identified and a loading platform for installing the organic glass loading platform and the coaxial light source; the organic glass loading platform is arranged in the center of the loading platform, the side of the product to be identified with the identification characters is placed on the organic glass loading platform, the coaxial light source is invertedly installed below the loading platform and the coaxial light source is coaxially arranged with the organic glass loading platform, the industrial camera is arranged below the coaxial light source, and the coaxial light source illuminates the product to be identified from bottom to top.

[0034] In order to solve the above technical problems, another technical solution adopted by the present invention is to provide an OCR intelligent recognition method, comprising the following steps:

[0035] Send a trigger signal to the host computer when it is necessary to identify the identification characters on the product;

[0036] In response to the trigger signal, the host computer sends a light source control signal to the light source subsystem, and sends a camera control signal to the vision subsystem;

[0037] In response to the light source control signal, the light source subsystem provides uniform illumination for the visual subsystem;

[0038] In response to the camera control signal, the visual subsystem collects the character image of the product to be identified and sends it to the host computer;

[0039] The host computer recognizes the identification characters in the character image transmitted by the visual subsystem.

[0040] The OCR intelligent recognition system and method of the present invention have at least the following beneficial effects: the present invention constructs a complete automatic recognition process by integrating a switch control subsystem, a host computer, a light source subsystem and a visual subsystem. Compared with the traditional manual recognition method, the present invention realizes automatic recognition, reduces the time and cost of manual operation, significantly improves production efficiency, and improves the accuracy and stability of recognition; by adopting a light source design that irradiates from bottom to top, it is ensured that no matter what the height of the product to be recognized is, the distance from the recognized surface to the camera remains consistent, thereby completely eliminating the influence of the product height on the recognition accuracy; by setting the image recognition area, the recognition range is narrowed, the recognition time is significantly reduced, and in the case where there are interference items around the characters to be recognized, by setting the image recognition area, it can be ensured that only useful character information is recognized, avoiding recognition errors caused by interference items; by identifying product labels, numbers and other information through the technical solution of the present invention, automated production and quality control are realized, and the digitization and intelligence of production and manufacturing are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0042] Figure 1 This is a system block diagram of an implementation scheme of an OCR intelligent recognition system of the present invention.

[0043] Figure 2 for Figure 1 221 in the image processing module 221.

[0044] Figure 3 This is a recognition result diagram when the OCR intelligent recognition system of the present invention is not set with an image recognition area.

[0045] Figure 4 This is a recognition result diagram after the OCR intelligent recognition system of the present invention sets the image recognition area.

[0046] Figure 5 This is an abnormal recognition result diagram of the OCR intelligent recognition system of the present invention when no image recognition area is set and there are interference items.

[0047] Figure 6 The OCR intelligent recognition system of the present invention sets the image recognition area to exclude interference items.

[0048] Figure 7 It is a structural diagram of an implementation scheme of an OCR intelligent recognition system of the present invention.

[0049] Figure 8It is a schematic diagram of the internal structure of an OCR intelligent recognition system according to an embodiment of the present invention.

[0050] Fig. 9 It is a flow chart of an implementation method of the OCR intelligent recognition method of the present invention. DETAILED DESCRIPTION

[0051] The present invention will be further described below in conjunction with the accompanying drawings.

[0052] See also Figure 1 The OCR intelligent recognition system of the present invention can be used to recognize the identification characters on the circuit housing, and includes a switch control subsystem 100, a host computer 200, a light source subsystem 300 and a visual subsystem 400. The switch control subsystem 100 is used to send a trigger signal to the host computer 200 when it is necessary to recognize the identification characters on the product; the host computer 200 is used to respond to the trigger signal, send a light source control signal to the light source subsystem 300, and send a camera control signal to the visual subsystem 400; the light source subsystem 300 is used to respond to the light source control signal to provide uniform illumination for the visual subsystem 400; the visual subsystem 400 is used to respond to the camera control signal, collect the character image of the product to be recognized and send it to the host computer 200. The host computer 200 is also used to respond to the trigger signal, run the OCR recognition program, and recognize the identification characters in the character image after receiving the character image transmitted by the visual subsystem 400.

[0053] The switch control subsystem 100 includes a button switch 110 and a network control IO module 120. The button switch is used to be pressed when the identification characters on the product need to be identified; the network control IO module 120 is used to send a trigger signal to the host computer 200 when the button switch 110 is pressed. The host computer 200 includes a control module 210 and a recognition subsystem 220. The control module 210 is used to send a light source control signal to the light source subsystem 300 and a camera control signal to the visual subsystem 400 in response to the trigger signal; the recognition subsystem 220 is used to run the OCR recognition program in response to the trigger signal to recognize the identification characters in the character image transmitted by the visual subsystem 400. After the recognition is completed, the recognition result will be written into the database for easy query and tracing. In this embodiment, the host computer 200 uses a touch screen all-in-one machine, which can display the video interface and recognition results in real time. The operator can query historical data and set recognition parameters through the touch screen all-in-one machine.

[0054] The recognition subsystem 220 includes an image processing module 221 and an image recognition module 222. The image processing module 221 is used to pre-process the character image transmitted by the visual subsystem 400; the image recognition module 222 is used to recognize the image obtained after pre-processing by the image processing module 221 and extract the identification characters therein. Figure 2 , the image processing module 221 includes a grayscale processing submodule 221a, a binarization submodule 221b, a noise reduction submodule 221c and an image correction submodule 221d. The grayscale processing submodule 221a is used to perform grayscale processing on the character image transmitted by the visual subsystem 400 to obtain a grayscale image to reduce the complexity of subsequent processing. The binarization submodule 221b is used to perform binarization processing on the grayscale image to obtain a black and white image so as to more clearly identify the characters. The noise reduction submodule 221c is used to perform noise reduction processing on the black and white image to remove interference factors such as stains and scratches and improve the accuracy of character recognition. The image correction submodule 221d is used to detect and correct the tilt angle of the image obtained after the noise reduction processing to ensure that the text lines are horizontal or vertical for subsequent character recognition.

[0055] The image recognition module 222 uses an algorithm program developed based on the open source text recognition model PDDL eOCR to recognize characters in the preprocessed image. The program can set the image recognition area according to the size of the identification characters of the object to be recognized. After the setting is successful, the program automatically stores the parameter, and the setting is permanent. Setting the image recognition area has the following advantages: (1) It reduces the scope of program recognition and can significantly reduce the recognition time. Please refer to Figure 3 and Figure 4 After setting the image recognition area, the recognition time was reduced from 495ms to 273ms, which is about 45% lower. (2) If there are interference items around the characters to be recognized, there is a certain probability of causing recognition errors. Please refer to Figure 5 , there are pins around the characters to be recognized, and the projection of the pins is easily recognized as characters, see Figure 6 ,It can be seen that by setting the image recognition area, interference items can be effectively avoided and the accuracy of the recognition results can be improved.

[0056] The light source subsystem 300 includes a serial port server 310, a light source controller 320, and a coaxial light source 330. The serial port server 310 is used to convert serial port data into network data to realize communication between the host computer 200 and the light source controller 320; the light source controller 320 is used to control the on / off state and light source brightness of the coaxial light source 330 in response to the light source control signal; the coaxial light source 330 is used to provide uniform and stable lighting for the visual subsystem 400 under the control of the light source controller 320, reduce object reflection, and improve the accuracy and reproducibility of machine vision. In specific operation, the operator first inputs or selects the required light source brightness value through the interface of the host computer 200, and the host computer 200 sends this instruction to the serial port server 310 through the network as a light source control signal, and the serial port server 310 converts the network data back to the serial port data format and sends it to the light source controller 320, and the light source controller 320 adjusts the brightness of the coaxial light source 330 according to the received instruction. The operator can monitor the status and brightness of the light source in real time through the host computer 200, and make further adjustments as needed to obtain the best character image acquisition effect.

[0057] The visual subsystem 400 includes a switch 410, an industrial lens 420 and an industrial camera 430. The switch 410 is used to realize the communication between the host computer 200 and the industrial camera 430. The host computer 200 can send a camera control signal to the industrial camera 430 through the switch 410 and receive the character image collected by the industrial camera 430. The industrial lens 420 is used to image the characters on the product to be identified on the photosensitive surface of the industrial camera 430. The industrial lens 420 can realize the transformation and focusing of the light beam to ensure the clarity and resolution of the character image. The industrial camera 430 is used to respond to the camera control signal, collect the character image of the product to be identified through the industrial lens 420, convert it into a digital signal and transmit it to the host computer 200. In this embodiment, the host computer 200 sets the image recognition area according to the size of the identification character on the object to be identified and applies the parameters of the image recognition area to the industrial camera 430. The industrial camera 430 collects the character image within the image recognition area. The image recognition area defines the image range that the industrial camera 430 needs to collect to ensure that only useful character information is included. In specific operation, the host computer 200 sends the set image recognition area parameters to the industrial camera 430 through the switch 410. The industrial camera 430 adjusts its image acquisition area according to these parameters to ensure that only the character images within the image recognition area are acquired. Under the uniform illumination of the coaxial light source 330, the industrial camera 430 acquires the character images on the product to be identified through the industrial lens 420. The industrial camera 430 converts the acquired image into a digital signal and transmits it to the host computer 200 in real time through the switch 410.

[0058] See also Figure 7 and Figure 8 The OCR intelligent system of the present invention also includes a fixing bracket 1 for installing and fixing all hardware, an organic glass stage 2 for placing the product to be identified, a loading platform 3 for installing the organic glass stage 2 and the coaxial light source 330, and a protective plate 4 for preventing the light path from being affected by the ambient light source. The organic glass stage 2 is arranged in the center of the loading platform 3, and the side of the product to be identified with the identification characters is placed on the organic glass stage 2. The coaxial light source 330 is inverted and installed below the loading platform 3 and the coaxial light source 330 is coaxially arranged with the organic glass stage 2. The industrial camera 430 is arranged below the coaxial light source 330, and the coaxial light source 330 illuminates the product to be identified from bottom to top. The protective plate 4 is arranged below the loading platform 3 along the four sides of the loading platform 3 to wrap the entire light path, so that the light path is not affected by the external ambient light source, thereby improving the recognition accuracy.

[0059] The present invention also discloses an OCR intelligent recognition method, which is implemented based on the OCR intelligent recognition system of the above embodiment. Fig. 9 , is a flow chart of an implementation method of an OCR intelligent recognition method of the present invention. This implementation method specifically includes the following steps:

[0060] S1. Send a trigger signal to the host computer 200.

[0061] Specifically, when the identification characters on the product need to be identified, the button switch 110 is pressed. When the button switch 110 is pressed, the network control IO module 120 sends a trigger signal to the host computer 200.

[0062] S2. Sending light source control signal and camera control signal.

[0063] Specifically, in response to the trigger signal, the host computer 200 sends a light source control signal to the light source subsystem 300 , and sends a camera control signal to the vision subsystem 400 .

[0064] S3. Provide uniform illumination for the visual subsystem 400.

[0065] Specifically, in response to the light source control signal, the light source controller 320 controls the on / off state and brightness of the coaxial light source 330 to provide uniform and stable lighting for the vision subsystem 400, reduce object reflections, and improve the accuracy and reproducibility of machine vision.

[0066] S4. Collect the character image of the product to be identified and send it to the host computer 200.

[0067] Specifically, in response to the camera control signal, the industrial camera 430 collects the character image of the product to be identified through the industrial lens 420, converts it into a digital signal and transmits it to the host computer 200. In this embodiment, the host computer 200 sets the image recognition area according to the size of the identification character on the object to be identified and applies the parameters of the image recognition area to the industrial camera 430. The industrial camera 430 collects the character image within the image recognition area. The image recognition area defines the image range that the industrial camera 430 needs to collect to ensure that only useful character information is included.

[0068] S5. Recognize the identification characters in the character image.

[0069] Specifically, after receiving the character image transmitted by the visual subsystem 400, the upper computer 200 first performs grayscale processing on the character image to obtain a grayscale image; then, it performs binarization processing on the grayscale image to obtain a black and white image; secondly, it performs noise reduction processing on the black and white image to remove interference factors such as stains and scratches; then, it detects and corrects the inclination angle of the image obtained after the noise reduction processing to ensure that the text lines are horizontal or vertical; finally, the algorithm program developed based on the open source text recognition model PDdleOCR recognizes the characters in the image to obtain the identification characters on the product to be identified.

[0070] The present invention constructs a complete automatic recognition process by integrating the switch control subsystem, the host computer, the light source subsystem and the visual subsystem. Compared with the traditional manual recognition method, the present invention realizes automatic recognition, reduces the time and cost of manual operation, significantly improves production efficiency, and improves the accuracy and stability of recognition. By adopting the light source design of irradiating from bottom to top, it ensures that no matter what the height of the product to be recognized is, the distance from the recognized surface to the camera remains consistent, thereby completely eliminating the influence of the product height on the recognition accuracy. The coaxial light source of the present invention not only provides uniform and stable illumination, but also effectively avoids interference from external light due to the setting of the protective plate. The accuracy and stability of character recognition are further improved; by setting the image recognition area, the recognition range is narrowed, the recognition time is significantly reduced, and when there are interference items around the characters to be recognized, by setting the image recognition area, it can be ensured that only useful character information is recognized, avoiding recognition errors caused by interference items; the recognition range is wide, and characters within the range of 100*100mm can be recognized. Characters can be accurately recognized without training, with a recognition accuracy of more than 98% and a recognition speed of less than 1s / time; product labels, numbers and other information are recognized through the technical solution of the present invention, automated production and quality control are realized, and the digitization and intelligence of production and manufacturing are improved.

[0071] The above content only expresses the preferred embodiments of the present invention, and its description is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention patent. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be based on the attached claims.

Claims

1. An OCR intelligent recognition system, characterized in that: include: The switch control subsystem is used to send a trigger signal to the host computer when the identification characters on the product need to be recognized; The host computer is used to respond to the trigger signal, send a light source control signal to the light source subsystem, send a camera control signal to the visual subsystem, and run an OCR recognition program to recognize the identification characters in the character image transmitted by the visual subsystem; A light source subsystem, for providing uniform illumination for the visual subsystem in response to a light source control signal; The visual subsystem is used to respond to the camera control signal, collect the character image of the product to be identified and send it to the host computer.

2. The OCR intelligent recognition system according to claim 1, characterized in that: The switch control subsystem comprises: A push button switch, which is used to be pressed when it is necessary to identify the identification characters on the product; The network control IO module is used to send a trigger signal to the host computer when the button switch is pressed.

3. The OCR intelligent recognition system according to claim 1, characterized in that: The light source subsystem comprises: Serial port server, used to convert serial port data into network data to realize communication between the host computer and the light source controller; A light source controller, used to control the on / off state and brightness of the coaxial light source in response to a light source control signal; The coaxial light source is used to provide uniform and stable lighting for the visual subsystem under the control of the light source controller.

4. The OCR intelligent recognition system as claimed in claim 3, characterized in that: The visual subsystem includes: Switch, used to realize communication between the host computer and the industrial camera; Industrial lens, used to image the characters on the product to be identified on the photosensitive surface of the industrial camera to achieve light beam transformation; The industrial camera is used to respond to the camera control signal, collect the character image of the product to be identified through the industrial lens, convert it into a digital signal and transmit it to the host computer.

5. The OCR intelligent recognition system as claimed in claim 4, characterized in that: The host computer is also used to set an image recognition area according to the size of the identification characters on the product to be recognized and apply the parameters of the image recognition area to the industrial camera, and the industrial camera collects the character image in the image recognition area.

6. The OCR intelligent recognition system according to claim 1, characterized in that: The host computer comprises: A control module, configured to send a light source control signal to the light source subsystem and a camera control signal to the vision subsystem in response to a trigger signal; The recognition subsystem is used to run the OCR recognition program in response to the trigger signal to recognize the identification characters in the character image transmitted by the visual subsystem.

7. The OCR intelligent recognition system as claimed in claim 6, characterized in that: The identification subsystem comprises: An image processing module, used for preprocessing the character image transmitted by the visual subsystem; The image recognition module is used to recognize the image obtained after preprocessing by the image processing module and extract the identification characters therein.

8. The OCR intelligent recognition system as claimed in claim 7, characterized in that: The image processing module comprises: The grayscale processing submodule is used to process the character image transmitted by the visual subsystem to obtain a grayscale image; Binarization submodule, used to perform binarization processing on the grayscale image to obtain a black and white image; The denoising submodule is used to perform denoising on black and white images to remove stains and scratches; The image correction submodule is used to detect and correct the tilt angle of the image obtained after noise reduction processing to ensure that the text lines are horizontal or vertical.

9. The OCR intelligent recognition system as claimed in claim 4, characterized in that: It also includes an organic glass storage table for placing the product to be identified and a loading platform for installing the organic glass loading platform and the coaxial light source; the organic glass loading platform is arranged in the center of the loading platform, the side of the product to be identified with the identification characters is placed on the organic glass loading platform, the coaxial light source is invertedly installed below the loading platform and the coaxial light source is coaxially arranged with the organic glass loading platform, the industrial camera is arranged below the coaxial light source, and the coaxial light source illuminates the product to be identified from bottom to top.

10. An OCR intelligent recognition method, characterized in that: The following steps are involved: Send a trigger signal to the host computer when it is necessary to identify the identification characters on the product; In response to the trigger signal, the host computer sends a light source control signal to the light source subsystem, and sends a camera control signal to the vision subsystem; In response to the light source control signal, the light source subsystem provides uniform illumination for the visual subsystem; In response to the camera control signal, the visual subsystem collects the character image of the product to be identified and sends it to the host computer; The host computer recognizes the identification characters in the character image transmitted by the visual subsystem.

Images