Defect Detection Device Based on Machine Vision and Defect Detection Method for Film Products
In the lithium battery separator BSF material detection, a row of cameras and strobe controllers are used to combine multi-layer optical films and different light source configurations, and the problem of difficult to distinguish bright spots, transparent points, semi-transparent and pinholes in the prior art is solved, and efficient and accurate defect detection and distinction are achieved.
Patent Information
- Application Number
- CN202210233219.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-10
- Publication Date
- 2025-06-20
- Estimated Expiration
- 2042-03-10
AI Technical Summary
In the prior art, when detecting defects in lithium battery diaphragm BSF materials, it is difficult to effectively distinguish between bright spots, transparent points, semi-transparent points and pinholes, resulting in increased detection errors and costs.
Using a machine vision defect detection device, by setting a row of cameras and strobe controllers, using multiple layers of optical films and different light source configurations, the alternating lighting and shooting of the first light source and the second light source are realized, and the light propagation distance is adjusted to distinguish different types of defects.
Accurate detection and distinction of different types of defects on lithium battery separator BSF materials is achieved, which reduces detection costs and improves detection efficiency, avoids missed detection and misjudgment.
Smart Images

Figure CN114740011B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a detection device and method for continuous production materials such as films in the field of machine vision technology, and particularly to visual inspection applied in the lithium battery separator BSF industry for detecting various defects on continuous production materials. Background Art
[0002] During the detection of BSF defects on continuous rolls in the field of machine vision, it is necessary to detect certain different types of defect points. For example:
[0003] In the production of lithium battery separators (BSF), defect points often occur on the material surface, and their defect types are: crystal points / black points / bright points / transparent points / missing coating / dark and bright spots / semi-transparent / oil stains / wrinkle stripes and pinholes. Among the BSF materials, the black points / bright points / transparent points / missing coating / dark and bright spots / semi-transparent / oil stains / wrinkle stripes belong to the defects formed on the material surface coating, mostly affecting the performance of the finished battery. However, only the pinholes (perforations in the BSF material) will cause the positive and negative electrodes of the battery to contact and cause a short circuit.
[0004] At present, in the field of machine vision, the above different types of defect detections are carried out using two stations (i.e., two rows of cameras Station1&Station2).
[0005] Station 1:
[0006] The installation angles of the camera, the material, and the light source are perpendicular (the camera is above the material), and the light source configuration used is backlight (the light source is placed below the material), and the light source is 100 mm away from the material. This configuration detects the above-mentioned crystal points / black points / bright points / transparent points / missing coating / dark and bright spots / semi-transparent / oil stains / wrinkle stripes / pinholes through a close-range backlight source. However, in this St1 detection scheme, it is easy to cause the bright points / transparent points / semi-transparent and pinholes to not be distinguishable from each other because using a close-range backlight illumination makes the gray levels and appearance sizes of these defects collected by the camera's photosensitive sensor the same.
[0007] Station 2:
[0008] In order to distinguish between bright points and pinholes, a second station, Station 2, is set up. This station only detects pinhole defects, and its configuration is the same as that of St1. The installation angles of the camera, the material, and the light source are perpendicular, and the light source is 300 mm away from the material below. The detection principle is that because the light source is far from the material and does not provide the brightness of the production material, the gray levels of the bright points / transparent points / semi-transparent defects collected by the camera's photosensitive sensor will be very low. When there are pinhole defects on the material, because the light source directly passes through the hole, the camera captures a high-brightness gray level value. This station completely independently detects pinholes.
[0009] It can be seen that the previous detection technical solutions at least needed to use two stations ST1 and ST2, and used light sources at different distances for lighting in the light source configuration, and the camera view point and the light source point needed to be perpendicular to achieve the detection of different types of defects. Summary of the Invention
[0010] The object of the present invention is to provide a technical solution that can achieve the detection capabilities of two stations with only one station and effectively distinguish the differences between bright points / transparent points / semi-transparent and pinholes. For this purpose, the present invention adopts the following technical solutions
[0011] According to the first aspect of the present invention, the following technical solutions are adopted:
[0012] A defect detection device based on machine vision, including a vision detection system and a backlight system. The backlight system includes a first light source and a second light source. It is characterized in that the detection device is further provided with a stroboscopic controller, and the backlight system further includes an optical structure configured with a reflectivity of 50% and a transmittance of 50%;
[0013] The vision detection system is communicatively connected to the stroboscopic controller, and the stroboscopic controller is connected to the first light source and the second light source, and transmits pulse signals to the first light source and the second light source in sequence;
[0014] Both the first light source and the second light source are backlights. The first light source is located behind the optical structure and provides backlight through transmission. The second light source is located on the side of the optical structure and provides backlight through reflection. By adjusting the installation positions of the two light sources, the light propagation distances from the first light source and the second light source to the camera of the vision detection system are different.
[0015] Further, the optical structure adopts a lens and is provided with multiple layers of optical films.
[0016] Further, the first light source and the second light source are respectively LED light sources, and the first light source and the second light source are respectively arranged along the width direction of the object to be detected; the stroboscopic controller provides the light source trigger timing and controls the first LED light source and the second LED light source to alternately turn on and off under the control of the trigger signal.
[0017] According to the second aspect of the present invention, the following technical solutions are adopted:
[0018] A method for detecting film product defects based on the above-mentioned machine vision defect detection device, characterized in that a row of cameras is set in the vision detection system, the first light source and the second light source are alternately lit at two sets of staggered intervals, and the row of cameras is used to alternately capture the images of the film product to be inspected projected by the first light source and the images of the film product to be inspected projected by the second light source. The light propagation distance from the first light source to the cameras of the vision detection system is longer than the light propagation distance from the second light source to the cameras of the vision detection system. Pinhole defects on the film product are distinguished by the images projected by the first light source; the alternate lighting of the first light source and the second light source by the row of cameras is controlled by the strobe controller transmitting pulse signals to the first light source and the second light source in sequence. When the cameras take pictures, a strobe control signal is generated synchronously, and this signal is used to control the strobe controller to transmit pulse signals to the first light source and the second light source in sequence.
[0019] By adopting multi-layer optical films and the classification function of the vision detection system, the present invention enables the functions originally achieved by two rows of cameras to be satisfied with one row of cameras, achieving the purpose of cost reduction. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 It is a comparison schematic diagram for the present invention and the prior art to distinguish different types of defects.
[0021] Figure 2 It is a schematic diagram of the detection device of the present invention.
[0022] Figure 3 It is a schematic diagram of the light source trigger timing provided by the strobe controller.
[0023] Figure 4 It is a schematic diagram of the working principle of the multi-layer optical film of the present invention.
[0024] Figure 5 It is a control flow chart of the present invention.
[0025] Figure 6-1 and 6-2 It is a schematic diagram that the light source can detect and distinguish holes through the multi-layer optical film.
[0026] Figure 7-1 、 7-2 and 7-3 are respectively the gray scale curve graphs of liquid oil stain 11, hole 2, and liquid oil stain 12.
[0027] Figure 8 It is the actual interface of the detection results for black dots, bright dots, missed coating, and white spots. DETAILED DESCRIPTION OF THE INVENTION
[0028] Refer to the attached drawings. The machine vision defect detection device provided by the present invention includes a vision detection system and a backlight system. The backlight system includes a first light source 101 and a second light source 102. The detection device is also provided with a stroboscopic controller. The backlight system further includes an optical architecture 200, and the optical architecture is configured with a reflectivity of 50% and a transmittance of 50%.
[0029] The vision detection system is communicatively connected to the stroboscopic controller. The stroboscopic controller is connected to the first light source and the second light source, and transmits pulse signals to the first light source and the second light source in sequence.
[0030] Both the first light source 101 and the second light source 102 are backlights. The film product 400 to be detected (BSF in this embodiment) passes between the optical architecture 200 and the camera 300 of the vision detection system. That is, the camera is above the BSF, while the first light source 101, the second light source 102, and the optical architecture are below the BSF.
[0031] The first light source is located behind the optical architecture and provides backlight through transmission. The second light source is located on the side of the optical architecture and provides backlight through reflection. The light propagation distances from the first light source and the second light source to the camera of the vision detection system are different. Among them, the light propagation distance from the first light source to the material to be detected is longer than that from the second light source to the material to be detected. The first light source is used to distinguish crystal points / black dots / bright dots / transparent points / missing coating / dark spot bright spot / half-transparent / oil stain / wrinkle stripe and pinholes existing in the BSF material, and the second light source is used to distinguish black dots / bright dots / transparent points / missing coating / dark spot bright spot / half-transparent / oil stain / wrinkle stripe existing in the BSF material.
[0032] The following further describes each structure with reference to the attached drawings.
[0033] The first light source 101 and the second light source are respectively LED light sources. The first light source 101 and the second light source 102 are respectively arranged along the width direction of the object to be detected, that is, the film product. The stroboscopic controller provides the light source trigger timing and controls the first LED light source and the second LED light source to alternately turn on and off under the control of the trigger signal.
[0034] Refer to Figure 3, the stroboscopic controller can adopt a conventional pulse signal generator (which must be able to accept an external trigger signal of +5V to provide the 40V working voltage required for the normal operation of the light source, and can provide a current adjustment from 0.5A to 5A), provide the trigger timing of the light source, and control the LED light sources of the first light source and the second light source to alternately turn on and off under the control of the trigger signals of Trigger1 and Trigger2. During odd time periods such as T1, T3, T5, T7, T9..., the light source of LED light source A is on, while during even time periods such as T2, T4, T6, T8, T10..., the light source of LED light source B is on.
[0035] The optical architecture uses lenses and is provided with multiple layers of optical films. The main coatings on the surface of the optical films are mostly metals such as silver (Ag), titanium (Ti), iron (Fe), aluminum (Al)... etc. It has three characteristics, reflected as reflection / absorption / penetration. Using the characteristics of reflection and penetration, the multiple layers of optical films are configured to have a reflectivity of 50% and a transmittance of 50%.
[0036] Visual inspection system: A row of cameras is set in the visual inspection system. The cameras output stroboscopic signals to the stroboscopic controller. The stroboscopic controller receives the stroboscopic signals from the cameras, emits pulse signals, and controls the stroboscopy of the light source. The cameras are configured with a dual-scan mode and two groups of instantaneous stroboscopic actions of the light sources, and through the application of the multiple layers of optical films with 50% reflected light and 50% transmitted light, the camera's photosensitive sensor collects two groups of light sources from different projection distances, achieving the detection effect of two cameras with one camera system. Since this system uses an intelligent line-scan camera, under normal non-stroboscopic conditions, the camera scans sequentially according to the scan lines 1, 2, 3... After using stroboscopy, the camera outputs the pulse signals of the odd scan lines 1, 3, 5... as stroboscopic signals to the stroboscopic controller. The stroboscopic controller receives the stroboscopic signals from the camera and controls the first light source 101. The pulse signals of the even scan lines 2, 4, 6... are output as stroboscopic signals to the stroboscopic controller. The stroboscopic controller receives the stroboscopic signals from the camera and controls the second light source 102. In fact, the first light source 101 and the second light source 102 alternate in flashing. It's just that due to the too-fast scanning frequency of the camera, to the human eye, light sources A and B appear to be in a constantly-on state. The visual inspection system can detect surface defects on continuously running rolled, ribbon-shaped, or sheet-shaped materials online or offline, and accurately provide product quality information and defect classification.
[0037] Connection relationship: The visual inspection system and the stroboscopic controller establish communication through a network cable, and the communication protocol is TCP / IP. The stroboscopic controller is connected to the first light source 101 and the second light source 102 by a shielded hard wire, and the transmission is a pulse signal.
[0038] The first light source 101 and the second light source 102 provide light sources to the multi-layer optical film, and provide the reflected or transmitted light sources to the surface of the BSF material. The light source path changes due to the defects in the BSF material, and the change in the light source is received by the visual inspection system.
[0039] As Figure 6-1 , 6-2 shown, when the detection period of the camera 300 is to detect the image provided by the first light source passing through the multi-layer optical film, holes can be detected and distinguished. In Figure 7-1 , 7-2 and 7-3 are the gray-scale analyses of three kinds of defects made with gray-scale curves. It can be seen that the highest gray-scale of the liquid oil stain 11 is 129, the highest gray-scale of the pinhole 2 is 255, and the highest gray-scale of the liquid oil stain 12 is 122; the gray-scale difference between the pinhole and the oil spot and the background is very large, and the pinhole can be detected and distinguished.
[0040] When the detection period of the camera 300 is to detect the image provided by the second light source passing through the multi-layer optical film, other defects such as black dots, bright dots, missed coating, and white spots can be detected. As Figure 8 shown, through the second light source of the present invention, black dots of different sizes and white spots with different transparencies can be distinguished.
[0041] Therefore, the present invention can on-line detect the surface defects of various lithium battery separator materials, accurately provide product quality information, realize accurate identification and classification of defects, truly detect difficult-to-distinguish approximate defects, avoid missed detection and misjudgment, and accurately provide product quality information.
[0042] The above embodiments are only a relatively preferred technical solution of the present invention. Those skilled in the art should understand that without departing from the principle and essence of the present invention, the technical solutions or parameters in the embodiments can be modified or replaced, and all should be covered within the protection scope of the present invention.
Claims
1. A defect detection device based on machine vision, comprising a vision detection system and a backlight system. The backlight system includes a first light source and a second light source, and is characterized in that The detection device is further provided with a stroboscopic controller, and the backlight system further includes an optical architecture configured with a reflectivity of 50% and a transmittance of 50%. The vision detection system is communicatively connected to the stroboscopic controller, and the stroboscopic controller is connected to the first light source and the second light source, and transmits pulse signals to the first light source and the second light source in sequence. Both the first light source and the second light source are backlights. The first light source is located behind the optical architecture and provides backlight through transmission. The second light source is located on the side of the optical architecture and provides backlight through reflection. By adjusting the installation positions of the two light sources, the light propagation distances from the first light source and the second light source to the camera of the vision detection system are different. The vision detection system is provided with a row of cameras, and the first light source and the second light source are alternately lit at two sets of staggered intervals. The row of cameras is used to alternately capture the images of the film product to be inspected projected by the first light source and the images of the film product to be inspected projected by the second light source.
2. The defect detection device based on machine vision according to claim 1, characterized in that The optical architecture uses a lens and is provided with multiple layers of optical films.
3. The defect detection device based on machine vision according to claim 1, characterized in that The first light source and the second light source are LED light sources respectively, and the first light source and the second light source are arranged along the width direction of the object to be detected; the stroboscopic controller provides the light source trigger timing and controls the first LED light source and the second LED light source to alternately turn on and off under the control of the trigger signal.
4. A film product defect detection method based on the defect detection device based on machine vision according to claim 1, characterized in that The vision detection system is provided with a row of cameras, and the first light source and the second light source are alternately lit at two sets of staggered intervals. The row of cameras is used to alternately capture the images of the film product to be inspected projected by the first light source and the images of the film product to be inspected projected by the second light source. The light propagation distance from the first light source to the camera of the vision detection system is longer than the light propagation distance from the second light source to the camera of the vision detection system. The pinhole defects on the film product are distinguished by the images projected by the first light source; the alternate lighting of the first light source and the second light source by the row of cameras is controlled by the stroboscopic controller to transmit pulse signals to the first light source and the second light source in sequence. When the camera takes pictures, a stroboscopic control signal is generated synchronously, and this signal is used to control the stroboscopic controller to transmit pulse signals to the first light source and the second light source in sequence.
Citation Information
Patent Citations
Method and system for detecting penetration state of air film hole and computer storage medium
CN112710637A
Mask plate defect detection method and system based on stroboscopic switching illumination
CN113834818A