A medical mask defect detection device and method based on computer vision

Abstract

The present invention relates to a medical mask defect detection device and method based on computer vision. The detection device includes: an appearance component, a mask moving and transporting component, a non-defective mask collecting component, a defective mask discharging component, an image acquisition component, an industrial computer, an electrical control unit, a light source control unit, and a display component; the detection method is based on the device and AI computer vision, and uses a semantic segmentation model to detect various components of the mask, and then each component is sent to its own AI classifier to identify whether each component is a defective component, and then the results of each classifier are summarized to draw a conclusion: whether the target mask contains defects, and which part has which kind of defects. The medical mask defect detection device and method described in the present invention improve the efficiency and accuracy of mask detection, and meet the needs of existing factory mask defect detection. The modular design of the device can be quickly connected to the mask machine, and the detection efficiency is high.

Description

Technical Field

[0001] The present invention relates to computer image processing and medical protective product appearance quality detection technology, and in particular to a medical mask defect detection device and method based on computer vision. Background Art

[0002] The coverage rate of quality inspection equipment on the production lines of many domestic medical protective equipment manufacturers is still very low, and the common inspection method is still manual inspection. Manual inspection is inefficient and costly, and personal subjective factors have a great impact on the inspection results. The penalties for defective products entering the market in the medical protection field are becoming more and more severe, and the instability of manual inspection not only brings more complaints and fines to companies, but also brings immeasurable harm and losses to society. The introduction of intelligent inspection methods for medical protective equipment has become an urgent need in the industry.

[0003] Mask defects are diverse and difficult to quantify, and traditional visual algorithms are difficult to cope with such application scenarios. Therefore, the present invention is based on a computer vision algorithm based on deep learning, simulating the human neural network system, extracting high-order features of masks, and accurately locating the location and type of defects. At the same time, the present invention also takes advantage of the low resource consumption, fast speed, and stable results of traditional algorithms, solves the identification problem of some defects in a targeted manner, reduces the resource overhead of deep learning, and organically combines with deep learning algorithms to improve the comprehensive detection performance of the present invention. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a medical mask defect detection device based on computer vision, comprising:

[0005] Appearance component 1, appearance component 1 is a rectangular structure, divided into an upper and lower structure, the lower structure of the appearance component 1 is installed with an industrial computer 6, the industrial computer 6 is connected to an electrical control unit 7 and a light source control unit 8 through matching cables, the upper structure of the appearance component 1 is installed with a mask moving and transporting component 2, a non-defective mask collecting component 3 is provided at the rear of the mask moving and transporting component 2, a defective mask discharging component 4 is provided below the mask moving and transporting component 2, and a matching No. 1 image acquisition component 5.1, No. 2 image acquisition component 5.2, and No. 3 image acquisition component 5.3 are provided at the front of the mask moving and transporting component 2.

[0006] In one embodiment of the present invention, the mask moving and transporting component 2 includes three positive pressure nozzles 11, each of which is mounted on a matching reciprocating linear module, and a matching fixture is installed on the reciprocating linear module.

[0007] In one embodiment of the present invention, a rotatable angle clamp 12 is also mounted on the frame of the mask moving and transporting component 2 .

[0008] In one embodiment of the present invention, the defect-free mask collecting component 3 is located directly below the third positive pressure suction nozzle 11, and the defect-free mask collecting component 3 includes a defect-free mask collecting box and a matching push rod mechanism.

[0009] In one embodiment of the present invention, the defective mask discharging component 4 is installed below the mask shooting position corresponding to the third image acquisition component 5.3.

[0010] In one embodiment of the present invention, the first image acquisition component 5.1, the second image acquisition component 5.2, and the third image acquisition component 5.3 are respectively equipped with a high-angle ring light source, a strip light source, an industrial camera, and an industrial fixed-focus lens.

[0011] In one embodiment of the present invention, the cameras of the second image acquisition component 5.2 and the third image acquisition component 5.3 are of the same model, are installed relative to each other and staggered, and the spacing is the width of the camera field of view.

[0012] Another aspect of the present invention further provides a medical mask defect detection method based on computer vision, based on the design of the medical mask defect detection device according to claims 1-7, comprising the following steps:

[0013] Step S1: A mask A is delivered by the upstream mask machine, and the rotatable angle clamp 12 in the mask moving and transporting component 2 clamps the lower part of the nose bridge strip on the mask cloth. The electrical control unit 7 triggers the image acquisition component No. 1 to collect pictures. This picture mainly represents the side position of the nose bridge strip and whether it fits the mask cloth surface. The picture is acquired and processed by the industrial computer 6 to determine whether the nose bridge strip meets the standard and whether the nose bridge strip fits the mask cloth surface tightly, and the result is saved as a1;

[0014] Step S2: The mask A clamped by the rotatable angle clamp 12 in the mask moving and transporting component 2 is rotated and placed at the first positive pressure suction nozzle 11, and the clamp 12 immediately rotates back to the original position. The electrical control unit 7 triggers the image acquisition component No. 2 to collect pictures and process to obtain the image information of the first surface of the mask A, including the position, shape, quantity, and contamination characteristics of the nose bridge strips on both sides of the mask, ultrasonic welds, ultrasonic welding points, ear straps, rubber rings, cloth surfaces and other components. The picture is acquired and processed by the industrial computer 6, and the judgment result is saved as a2. At the same time, step S1 is performed synchronously to process the mask B delivered by the upstream mask machine to obtain the result b1;

[0015] Step S3: Mask A moves forward once through the reciprocating linear module in the mask moving and transporting component 2, and the distance is the width of a camera field of view. Mask A is moved to the second positive pressure suction nozzle 11, and the electrical control unit 7 triggers the image acquisition component No. 3 to collect pictures and process to obtain the image information of the second side of mask A. The picture features are the same as the first side. The picture is acquired and processed by the industrial computer 6, and the judgment result is saved as a3. At the same time, step 2 is performed synchronously to obtain the result b2 of mask B. The mask C delivered by the upstream mask machine is synchronously performed according to step 1 to obtain the result c1; the reciprocating linear module in the mask moving and transporting component 2 returns to its position backward;

[0016] Step S4: combine the results of a1, a2, and a3, that is, perform an OR operation, and the combined result is recorded as a; if a is judged to be defective, the suction nozzle blows the mask A at the current position to the defective mask discharge component 4; otherwise, the reciprocating linear module in the mobile conveying component 2 moves forward once, and the mask A is moved to the third positive pressure suction nozzle and placed in the receiving box of the non-defective mask collecting component 3; at the same time, step S3 and step S2 are performed synchronously to obtain results b3 and c2, and step S1 simultaneously processes the mask D delivered by the upstream mask machine to obtain result d1; the reciprocating linear module in the mask moving conveying component 2 returns to its position backward;

[0017] Step S5: Repeat the above steps to process the masks delivered from the upstream mask machine in turn. When the number of masks in the receiving box of the defect-free mask collecting component 3 reaches the preset value, the push rod mechanism is executed once to push all the masks to the designated station or belt conveyor line.

[0018] In one embodiment of the present invention, the industrial computer 6 is integrated with detection software, which is used to process the current mask image, determine whether the mask in the image has defects and send the result to the electrical control unit 7, while saving the process data.

[0019] The above technical solution of the present invention has the following advantages over the prior art: the medical mask defect detection and device method based on computer vision technology described in the present invention can be applied to the mask production line to identify the vast majority of defect items, almost covering all the defect items that must be inspected, thereby replacing manual inspection, improving the efficiency and accuracy of mask inspection, and meeting the needs of existing factory mask defect detection. The modular design of this device can be quickly connected to the mask machine, has high detection efficiency, and is fully adapted to the production speed of the current mask machines on the market. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to make the contents of the present invention more clearly understood, the present invention is further described in detail below based on specific embodiments of the present invention in conjunction with the accompanying drawings.

[0021] Figure 1It is an overall layout diagram of a medical mask defect detection device of the present invention.

[0022] Figure 2 The present invention is a left view of the arrangement of the upper actuator of a medical mask defect detection device.

[0023] Figure 3 The present invention is a right view of the arrangement of the upper actuator of a medical mask defect detection device.

[0024] Figure 4 It is a structural schematic diagram of a mask moving and transporting component of a medical mask defect detection device of the present invention.

[0025] Figure 5 It is a structural schematic diagram of a positive pressure nozzle of a medical mask defect detection device of the present invention.

[0026] Figure 6 It is a structural schematic diagram of a rotatable angle clamp of a medical mask defect detection device of the present invention.

[0027] Figure 7 It is a structural schematic diagram of a No. 1 image acquisition component of a medical mask defect detection device of the present invention.

[0028] Figure 8 It is a structural schematic diagram of a No. 2 image acquisition component and a No. 3 image acquisition component of a medical mask defect detection device of the present invention.

[0029] As shown in the figure: 1. Appearance parts, 2. Mask moving and transporting parts, 3. Non-defective mask collection parts, 4. Defective mask discharge parts, 5.1. Image acquisition part No. 1, 5.2. Image acquisition part No. 2, 5.3. Image acquisition part No. 3, 6. Industrial computer, 7. Electrical control unit, 8. Light source control unit, 9. Display parts, 10. Positive pressure nozzle, 11. Rotatable angle clamp. DETAILED DESCRIPTION

[0030] like Figure 1 , Figure 2 , Figure 3 As shown, this embodiment specifically discloses the appearance layout of a medical mask defect detection device based on computer vision technology. The overall size of the device is: 1100*620*1600 (length*width*height) mm 3The main components include: appearance component 1, mask moving and transporting component 2, non-defective mask collection component 3, defective mask discharge component 4, image acquisition component No. 1 5.1, image acquisition component No. 2 5.2, image acquisition component No. 3 5.3, industrial computer 6, electrical control unit 7, light source control unit 8, display component 9, positive pressure nozzle 10, rotatable angle clamp 11 and detection software, etc. Among them: appearance component 1 includes start and stop button, reset button, emergency stop button, three-color light, network interface, USB interface, serial interface, power interface, compressed air interface, ground wire, Foma casters and other accessories.

[0031] In one embodiment of the present invention, Figure 1 As shown, the image acquisition component No. 1 5.1 and the material receiving box in this embodiment exceed the overall size range of the equipment after installation, which is determined by the installation space on site of the embodiment.

[0032] In one embodiment of the present invention, Figure 4 , Figure 5 , Figure 6 As shown, this embodiment specifically discloses a mask moving and handling component of a medical mask defect detection device based on computer vision technology. It mainly includes: a rotatable clamping jaw 11, a servo motor, a linear guide rail, a T-shaped beam, three equidistantly installed fixed clamping jaws and three positive pressure suction nozzles 10 and a bracket. The main working points of the mask moving and handling component and its various parts are as follows:

[0033] 1) The rotatable angle clamp is used to remove the mask delivered by the mask machine. When the mask machine clamps the mask and sends it to the specified position, the rotatable angle clamp is located at the grasping position to clamp the mask. At this time, the No. 1 image acquisition component 5.1 captures the current picture.

[0034] 2) The function of the positive pressure nozzles No. 1 and No. 2 is to fix the shooting position of the mask. They are fixedly installed on the work surface, 250mm apart, and 350mm high relative to the work surface. The nozzles are installed in the opposite direction, and the corresponding shooting components are the No. 2 image acquisition component 5.2 and the No. 3 image acquisition component 5.3 (the No. 2 image acquisition component 5.2 and the No. 3 image acquisition component 5.3 are installed opposite to each other, so that the left and right sides of adjacent masks can be photographed respectively). When the No. 1 image acquisition component 5.1 takes pictures almost at the same time, the masks on the No. 1 and No. 2 positive pressure nozzles are also taken pictures by the No. 2 image acquisition component 5.2 and the No. 3 image acquisition component 5.3. No. 3 nozzle is located above the defect-free mask collection box, and the other end is connected to a rotating cylinder. When there is a mask here, the rotating cylinder drives the No. 3 positive pressure nozzle to rotate downward 90°, and the mask is stacked horizontally in the collection box. The air valve of the No. 3 nozzle is closed, and then it is rotated upward 90° to return to its original position.

[0035] 3) After the image acquisition components 5.1, 5.2 and 5.3 are finished, the images are quickly acquired by the industrial computer and the results are determined. At this time, the device enters the process of forward transporting the mask: the rotatable angle clamp is rotated 90°, and the mask is placed at the No. 1 positive pressure suction nozzle. The No. 1 positive pressure suction nozzle air valve opens to suck the mask, and at the same time, the clamps on the T-shaped crossbeam are closed to clamp the masks at the corresponding positions respectively. The suction nozzle air valve is closed, and the motor drives the linear guide rail and the crossbeam to move forward the distance of one station (250mm in this embodiment, the same below), and then the suction nozzle air valve opens to suck the mask from the previous station.

[0036] 4) After the transportation is completed, the electrical system immediately determines whether the air valve of nozzle No. 2 is closed again based on the data sent by the industrial computer. The judgment is based on: the mask at the position of nozzle No. 2, after the results of the image analysis at shooting positions 1, 2, and 3, if the result is determined to be a defective mask, the air valve of nozzle No. 2 is closed, and the mask naturally falls into the defective mask discharge component, otherwise the mask continues to be sucked by nozzle No. 2, and when it is transported next time, it is transported to nozzle No. 3 and horizontally stacked in the collection box for non-defective masks.

[0037] In one embodiment of the present invention, Figure 1 , 2 As shown in , 3, 4, 5, and 6, this embodiment specifically discloses a defective mask discharging component of a medical mask defect detection device based on computer vision technology. The discharging component is located directly below the No. 2 positive pressure suction nozzle, and the opening size is 180mm*180mm. When the mask is judged to be defective, the No. 2 positive pressure suction nozzle air valve will be closed, and the mask will naturally fall into the opening of the discharging component. The opening is connected to the conveyor through a slideway and can be transported to a designated location.

[0038] In one embodiment of the present invention, Figure 1 , 2 As shown in Figures 3, 4, 5, and 6, this embodiment specifically discloses a non-defective mask receiving component of a medical mask defect detection device based on computer vision technology. The receiving component is located directly below the No. 3 positive pressure suction nozzle, and the opening size is 180mm*180mm*150mm. When the mask is judged to be non-defective, it will be rotated downward by 90° by the rotating cylinder connected to the No. 3 positive pressure suction nozzle and horizontally stacked into the receiving box of the non-defective mask collecting component 3. A cylinder pushing device is installed on the side of the receiving box. When the number of masks in the receiving box of the non-defective mask collecting component 3 reaches a preset value, the push rod mechanism is executed once to push all masks to the designated station or belt conveyor line.

[0039] In one embodiment of the present invention, Figure 7As shown, this embodiment specifically discloses the structure of the image acquisition component 5.1, which mainly includes a high-angle ring light source with a diameter of 100mm, a 6-megapixel industrial camera, an industrial lens with a focal length of 16mm and a bracket. The phase axis of the camera and the central axis of the light source are installed in a colinear manner, perpendicular to the side of the nose bridge of the mask, and ensure that the nose bridge is near the center of the field of view. The distance from the lens to the light source is 45mm, and the distance from the lens to the nose bridge is 128mm. The industrial camera bracket and the light source fixing bracket are designed in a way that the position and angle of the three directions of XYZ can be adjusted, and the installation position and posture can be adjusted in the above manner.

[0040] In one embodiment of the present invention, this embodiment provides a detection method for the device: a. Extract the edge contour of the mask cloth surface through frequency domain transformation of the image; c. Preliminarily locate the position of the end of the nose bridge strip according to the cloth surface contour and the positional relationship between the nose bridge strip and the cloth surface; d. Use the nose bridge strip model to find the end of the nose bridge strip in the area where the end of the nose bridge strip may appear; e. Result judgment: if there is a clear distance between the end of the nose bridge strip and the mask cloth surface contour, it can be determined that the mask nose bridge strip is debonded, otherwise it is deemed that there is no such defect.

[0041] In one embodiment of the present invention, Figure 8 As shown, this embodiment specifically discloses and provides image acquisition component No. 2 5.2 and image acquisition component No. 3 5.3. The hardware contained in the two components is consistent, respectively including: 1 industrial camera with 6 million pixels, 1 industrial lens with a focal length of 8mm, 2 strip light sources with a length of 300mm and a width of 60mm and a bracket. The misaligned distance of image acquisition component No. 2 5.2 and image acquisition component No. 3 5.3 is the width of the field of view of an industrial camera, that is, 250mm. The working distance of the industrial camera (the distance from the camera to the mask) is 250mm. The two strip light sources are vertically symmetrically arranged on both sides of the industrial camera with a spacing of 250mm, and the camera installation height is consistent with the midpoint height of the light source. The camera and the strip light source are on the same side of the mask being tested, and the direction of the camera lens is consistent with the direction of the light source. The two light sources are rotated 30° around the long axis in the relative direction. The industrial camera bracket and the light source fixing bracket are designed in a way that the position and angle of the three directions of XYZ can be adjusted in the above manner. The installation position and posture can be adjusted. After the installation of the No. 2 image acquisition component 5.2 and the No. 3 image acquisition component 5.3 is completed, each camera is facing the No. 1 and No. 2 nozzles, that is, the shooting position of the masks. The left and right sides of the two adjacent masks are photographed respectively.

[0042] In one embodiment of the present invention, the images obtained by the image acquisition component No. 1 5.1 are mainly used for defect detection of the side part of the nose bridge of the mask, including but not limited to defects such as debonding and warping of the nose bridge; the images obtained by the image acquisition component No. 2 5.2 and the image acquisition component No. 3 5.3 are mainly used for the detection of defects such as the position, shape, quantity, and contamination of the nose bridge, ultrasonic welds, ultrasonic welding points, ear straps, rubber rings, cloth surfaces and other parts of the mask.

[0043] In one embodiment of the present invention, this embodiment specifically discloses and provides image processing algorithms of the second image acquisition component 5.2 and the third image acquisition component 5.3. The main features of the images acquired by the second image acquisition component 5.2 and the third image acquisition component 5.3 are consistent, and they are images of the left and right sides of the same mask, respectively. Therefore, the defects detected by the two images are consistent, and the algorithms used are consistent. The algorithm flow is:

[0044] a. The image is preprocessed to obtain significant features. The preprocessing includes grayscale stretching, edge detection, color conversion, frequency domain transformation, threshold segmentation, image compression, ROI extraction, etc.

[0045] b. The image is segmented using a semantic segmentation model to locate the various components of the mask, including: fabric, nose bridge, text, sutures, welding points, ear straps, rubber rings, and abnormal components;

[0046] c. Send each component to different AI classifiers to determine whether it is a defective component or a non-defective component. At the same time, according to the positional relationship of each component, determine whether there is a relative position error;

[0047] d. Based on the result of c, summarize the results of each classifier and whether there are position errors to draw conclusions: whether the current mask contains defects, and which part and type of defect the defect belongs to.

[0048] In one embodiment of the present invention, the industrial computer 6 in this embodiment is integrated with detection software, and the functions include: 1. Picture display; 2. Determination result text and position identification display; 3. Production data (total number, yield, etc.); 4. Adjustable detection specification parameters; 5. Historical data storage and upload to server; 6. Alarm information display and log storage, etc. The software is easy to operate. Generally, users only need to open the software, connect the camera and the electronic control system, and it can run automatically. No operation is required during the operation. The software has an encryption system, and the software automatically binds the unique identification code of the industrial computer. After the software is copied to other industrial computers, it will not run normally due to the failure of the identification code verification.

[0049] In one embodiment of the present invention, a medical mask defect detection method based on computer vision is provided, and the operation process of this embodiment is as follows:

[0050] Step S1: A mask A is delivered by the upstream mask machine, and the rotatable angle clamp 12 in the mask moving and transporting component 2 clamps the lower part of the nose bridge strip on the mask cloth. The electrical control unit 7 triggers the image acquisition component No. 1 to collect pictures. This picture mainly represents the side position of the nose bridge strip and whether it fits the mask cloth surface. The picture is acquired and processed by the industrial computer 6 to determine whether the nose bridge strip meets the standard and whether the nose bridge strip fits the mask cloth surface tightly, and the result is saved as a1;

[0051] Step S2: The mask A clamped by the rotatable angle clamp 12 in the mask moving and transporting component 2 is rotated and placed at the first positive pressure suction nozzle 11, and the clamp 12 immediately rotates back to the original position. The electrical control unit 7 triggers the image acquisition component No. 2 to collect pictures and process to obtain the image information of the first surface of the mask A, including the position, shape, quantity, and contamination characteristics of the nose bridge strips on both sides of the mask, ultrasonic welds, ultrasonic welding points, ear straps, rubber rings, cloth surfaces and other components. The picture is acquired and processed by the industrial computer 6, and the judgment result is saved as a2. At the same time, step S1 is performed synchronously to process the mask B delivered by the upstream mask machine to obtain the result b1;

[0052] Step S3: Mask A moves forward once through the reciprocating linear module in the mask moving and transporting component 2, and the distance is the width of a camera field of view. Mask A is moved to the second positive pressure suction nozzle 11, and the electrical control unit 7 triggers the image acquisition component No. 3 to collect pictures and process to obtain the image information of the second side of mask A. The picture features are the same as the first side. The picture is acquired and processed by the industrial computer 6, and the judgment result is saved as a3. At the same time, step 2 is performed synchronously to obtain the result b2 of mask B. The mask C delivered by the upstream mask machine is synchronously performed according to step 1 to obtain the result c1; the reciprocating linear module in the mask moving and transporting component 2 returns to its position backward;

[0053] Step S4: combine the results of a1, a2, and a3, that is, perform an OR operation, and the combined result is recorded as a; if a is judged to be defective, the suction nozzle blows the mask A at the current position to the defective mask discharge component 4; otherwise, the reciprocating linear module in the mobile conveying component 2 moves forward once, and the mask A is moved to the third positive pressure suction nozzle and placed in the receiving box of the non-defective mask collecting component 3; at the same time, step S3 and step S2 are performed synchronously to obtain results b3 and c2, and step S1 simultaneously processes the mask D delivered by the upstream mask machine to obtain result d1; the reciprocating linear module in the mask moving conveying component 2 returns to its position backward;

[0054] Step S5: Repeat the above steps to process the masks delivered from the upstream mask machine in turn. When the number of masks in the receiving box of the defect-free mask collecting component 3 reaches the preset value, the push rod mechanism is executed once to push all the masks to the designated station or belt conveyor line.

[0055] Obviously, the above embodiments are merely examples for clear explanation and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived from these are still within the protection scope of the invention.

Claims

1. A medical mask defect detection method based on computer vision, based on the design of a medical mask defect detection device, characterized in that: The detection device comprises: An appearance component (1), wherein the appearance component (1) is a rectangular parallelepiped structure, and is divided into an upper and lower structure. The lower structure of the appearance component (1) is equipped with an industrial computer (6), and the industrial computer (6) is respectively connected to an electrical control unit (7) and a light source control unit (8) via matching cables. The upper structure of the appearance component (1) is equipped with a mask moving and transporting component (2), a non-defective mask collecting component (3) is provided at the rear of the mask moving and transporting component (2), a defective mask discharging component (4) is provided below the mask moving and transporting component (2), and a matching No. 1 image acquisition component (5.1), No. 2 image acquisition component (5.2), and No. 3 image acquisition component (5.3) are provided at the front of the mask moving and transporting component (2); The detection method comprises the following steps: Step S1: a mask A is delivered by the upstream mask machine, and the rotatable angle clamp (12) in the mask moving and transporting component (2) clamps the lower part of the nose bridge strip of the mask fabric, and the electrical control unit (7) triggers the image acquisition component No. 1 to acquire a picture. This picture mainly represents the side position of the nose bridge strip and whether it fits the mask fabric. The picture is acquired and processed by the industrial computer (6) to determine whether the nose bridge strip meets the standard and whether the nose bridge strip fits the mask fabric tightly, and the result is saved as a1; Step S2: the mask A clamped by the rotatable angle clamp (12) in the mask moving and transporting component (2) is rotated and placed at the first positive pressure suction nozzle (11), after which the clamp (12) immediately rotates back to its original position, and the electrical control unit (7) triggers the image acquisition component No. 2 to acquire an image and process and obtain image information of the first surface of the mask A, including the position, shape, quantity, and contamination characteristics of the nose bridge strips on both sides of the mask, ultrasonic welds, ultrasonic welding points, ear straps, rubber rings, cloth surface and other components. The image is acquired and processed by the industrial computer (6), and the judgment result is saved as a2. At the same time, step S1 is performed synchronously to process the mask B delivered by the upstream mask machine and obtain the result b1; Step S3: Mask A is moved forward once by the reciprocating linear module in the mask moving and transporting component (2), and the distance is the width of the camera field of view. Mask A is moved to the second positive pressure suction nozzle (11), and the electrical control unit (7) triggers the image acquisition component No. 3 to acquire an image and process it to obtain image information of the second side of mask A. The image features are the same as the first side. The image is acquired and processed by the industrial computer (6), and the judgment result is saved as a3. At the same time, step 2 is performed synchronously to obtain the result b2 of mask B. The mask C delivered by the upstream mask machine is performed synchronously according to step 1 to obtain the result c1; the reciprocating linear module in the mask moving and transporting component (2) returns to its original position; Step S4: combine the results of a1, a2, and a3, that is, perform an OR operation, and the combined result is recorded as a; if a is judged to be defective, the suction nozzle blows the mask A at the current position to the defective mask discharge component (4); otherwise, the reciprocating linear module in the movable conveying component (2) moves forward once, and the mask A is moved to the third positive pressure suction nozzle and placed in the receiving box of the non-defective mask collection component (3); at the same time, step S3 and step S2 are performed synchronously to obtain results b3 and c2, and step S1 simultaneously processes the mask D delivered by the upstream mask machine to obtain result d1; the reciprocating linear module in the mask movable conveying component (2) returns to its position; Step S5: Repeat the above steps to process the masks delivered from the upstream mask machine in turn; when the number of masks in the receiving box of the defect-free mask collecting component (3) reaches a preset value, the push rod mechanism is executed once to push all the masks to the designated workstation or belt conveyor line.

2. The method for detecting defects in medical masks according to claim 1, characterized in that: The industrial computer (6) is integrated with detection software, which is used to process the current mask image, determine whether the mask in the image has defects and send the result to the electrical control unit (7), while saving the process data.

3. The method for detecting defects in medical masks according to claim 1, characterized in that: The mask moving and transporting component (2) comprises three positive pressure suction nozzles (11), each of which is mounted on a matching reciprocating linear module, and a matching clamp is mounted on the reciprocating linear module.

4. The method for detecting defects in medical masks according to claim 3, characterized in that: The frame of the mask moving and transporting component (2) is also equipped with a rotatable angle clamp (12).

5. The method for detecting defects in medical masks according to claim 1, wherein: The defect-free mask collecting component (3) is located directly below the third positive pressure suction nozzle (11), and the defect-free mask collecting component (3) comprises a defect-free mask collecting box and a matching push rod mechanism.

6. The method for detecting defects in medical masks according to claim 1, characterized in that: The defective mask discharging component (4) is installed below the mask shooting position corresponding to the third image acquisition component (5.3).

7. The method for detecting defects in medical masks according to claim 1, characterized in that: The first image acquisition component (5.1), the second image acquisition component (5.2), and the third image acquisition component (5.3) are respectively equipped with a high-angle annular light source, a strip light source, an industrial camera, and an industrial fixed-focus lens.

8. The method for detecting defects in medical masks according to claim 7, characterized in that: The cameras of the second image acquisition component (5.2) and the third image acquisition component (5.3) are of the same model, and are installed relative to each other and staggered, with a spacing equal to the width of the camera's field of view.

Images