High-speed intelligent defect detection equipment for wafer biscuits based on deep learning
By combining multi-channel feeding with spiral propulsion and a deep learning vision inspection system, the problems of low efficiency, low accuracy, and poor adaptability of wafer barrel appearance defect detection equipment have been solved, achieving high-speed, online, automated, and high-precision inspection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GRACE (TIANJIN) VISION TECH CO LTD
- Filing Date
- 2026-04-23
- Publication Date
- 2026-07-21
AI Technical Summary
Existing wafer barrel appearance defect detection equipment is inefficient, inaccurate, and has poor adaptability. It also relies on manual labor or traditional machine vision has insufficient generalization ability and is difficult to adapt to changes in the shape of different products and lighting conditions.
The structure design combines multi-channel feeding with spiral propulsion for material transfer, and is equipped with a deep learning-based vision inspection system to achieve high-speed, online, and automated defect detection of wafer drums.
It improves detection accuracy and adaptability, reduces false detection rate and labor costs, and enhances production efficiency and product quality consistency.
Smart Images
Figure CN122425001A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of food packaging inspection technology, and in particular relates to a high-speed intelligent defect detection device for wafer buckets based on deep learning. Background Technology
[0002] Wafer buckets are commonly used packaging for ice cream, candy, and other food products, and their appearance quality directly affects product quality and brand image. During the production process, wafer buckets may develop various defects, such as surface cracks, structural deformation, poor printing, stains, and edge damage. Therefore, efficient and accurate appearance defect inspection before shipment is a crucial step in ensuring product quality.
[0003] Currently, the industry mainly relies on the following methods to inspect the appearance quality of wafer drums:
[0004] This method relies on workers visually inspecting and manually sorting samples at the production line. It is not only inefficient and difficult to integrate with modern high-speed production lines, but also prone to causing visual fatigue from prolonged work, resulting in high rates of missed and false detections, poor consistency in testing, and continuously rising labor costs.
[0005] Some production lines employ automated equipment equipped with photoelectric sensors or simple image processing algorithms. These systems typically rely on pre-set fixed thresholds, edge detection, or rigid template matching. While their detection speed is faster than manual methods, significant limitations remain: First, the algorithms lack flexibility and generalization ability, making it difficult to effectively identify and segment defects with diverse shapes and irregular features (such as fine, gradual cracks or irregular stains); second, the system is extremely sensitive to changes in lighting conditions, material position, and orientation, requiring improved stability; third, existing equipment designs are often tailored to specific shapes (such as single columnar or cap-shaped wafer drums), necessitating shutdowns and readjustment of equipment parameters, or even replacement of mechanical parts, when product specifications or shapes change, resulting in poor equipment adaptability and production line flexibility.
[0006] In addition, existing testing equipment has a low degree of integration in terms of material conveying, precise positioning and visual inspection, which often leads to materials stacking, rolling or being misaligned at the testing station, affecting the quality of image acquisition and thus limiting the final recognition accuracy.
[0007] In recent years, deep learning technology, especially convolutional neural networks (CNNs), has demonstrated powerful performance in image recognition and classification tasks, providing new solutions for the intelligent detection of complex industrial appearance defects. However, how to deeply integrate high-performance deep learning algorithms with high-speed, stable, and flexible material conveying and sorting actuators to build an integrated device that can adapt to different product shapes and achieve high-precision real-time online detection remains a pressing technical problem to be solved in this field.
[0008] Therefore, it is necessary to develop a new type of intelligent defect detection equipment to overcome the shortcomings of existing technologies in terms of detection efficiency, accuracy, adaptability, and intelligence. Summary of the Invention
[0009] The purpose of this invention is to provide a high-speed intelligent defect detection device for wafer barrels based on deep learning, so as to solve the technical problems of slow efficiency and low accuracy in traditional detection.
[0010] To solve the above-mentioned technical problems, the specific technical solution of the present invention is as follows: In some embodiments of this application, a high-speed intelligent defect detection device for wafer barrels based on deep learning is provided, including: A feeding component, wherein the feeding component is provided with multiple feeding channels; A material transfer component is located at the discharge end of the feeding component. It is used to receive the material conveyed by the feeding component and transfer it to the discharge component. A detection component is located above the material transfer component and detects the material on the material transfer component. A collection component, located below the discharge component, is used to collect defective materials.
[0011] In some embodiments of this application, the feeding component is a modular structure, including: A first support frame, on which a first conveying component is provided; The unloading component is located at one end of the first conveying component and is arranged at an inclination relative to the first conveying component. The first conveying component and the unloading component are provided with several feeding channels, and the end of the first conveying component is connected to the transfer component.
[0012] In some embodiments of this application, the material transfer component is a modular structure, including: The second bracket is located on one side of the first bracket and is fixedly connected to the first bracket. The transfer channel is arranged in a matrix on the second support, and is fixedly connected to the second support, and corresponds to the position of the feeding channel; A material conveying component is located on a transfer channel and has a rotating end thereon, which conveys materials by rotating a spiral. The unloading component is located at the discharge end of the transfer channel and is rotatably connected to the second support. It is provided with several unloading channels, which correspond to the positions of the transfer channel and are connected to the discharge component. The detection components are respectively located above the transfer conveyor and above the unloading component.
[0013] In some embodiments of this application, the discharge component is a modular structure, including: The third support is located on one side of the second support and is fixedly connected to the second support. The second conveying component is mounted on the third support and has several discharge channels on it; A swinging component is mounted on a third support and located at the feed end of the second conveying component. It is rotatably connected to the third support, wherein the swinging ends of the swinging component correspond to the discharge channels. The bottom of the swinging component is equipped with a collecting component.
[0014] In some embodiments of this application, the feeding component is a modular structure, including: A support frame, wherein the support frame is provided with an installation area; The feeding channels are arranged in a matrix on the support frame, wherein the feeding channels are arc-shaped. A horizontal channel is located at the end of the feed channel and is rotatably connected to the feed channel. A stop component is provided on a support frame and is rotatably connected to the support frame. Several baffles are provided on the stop component, wherein the baffles correspond to the positions of the horizontal track.
[0015] In some embodiments of this application, the material transfer component is a modular structure, including: The first vertical frame is located on one side of the horizontal path and is arranged symmetrically. The first rotating component is mounted on the first vertical frame and is rotatably connected to the first vertical frame. It is provided with two first receiving ends. The second vertical frame is located between the first vertical frame and the discharge component, and they are arranged symmetrically. The second rotating component is mounted on the second vertical frame and is rotatably connected to the second vertical frame. It is provided with multiple second receiving ends. The third vertical frame is located on both sides of the discharge component; The detection components are respectively disposed between the first vertical frame, the second vertical frame and the third vertical frame.
[0016] In some embodiments of this application, the discharge component is a modular structure, including: A conical truncated platform, wherein a trapezoidal frame is provided on one side of the conical truncated platform; The trapezoidal frame is provided with several discharge channels, and the discharge channels are provided with movable discharge valves; The bottom of the trapezoidal frame is equipped with a collecting component, which corresponds to the position of the discharge valve.
[0017] In some embodiments of this application, the height of the second rotating component is lower than the height of the first rotating component.
[0018] Compared with the prior art, the beneficial effects of the present invention are that by adopting a structural design that combines multi-channel feeding with spiral propulsion material transfer, and in conjunction with a deep learning-based vision inspection system, high-speed, online, and automated defect detection of wafer barrels is achieved. It has the beneficial effects of high detection accuracy, strong adaptability, accurate and reliable sorting, and high degree of automation, which significantly improves production efficiency and product quality consistency, reduces labor costs and false detection rate, and solves the problems of reliance on manual labor, poor generalization ability of traditional machine vision, and insufficient equipment flexibility in the prior art. Attached Figure Description
[0019] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a schematic diagram of the overall structure provided in Embodiment 2 of the present invention; Figure 2 This is a schematic diagram of the overall structure from different perspectives provided in Embodiment 3 of the present invention; Figure 3 This is a schematic diagram of the overall structure from different perspectives provided in Embodiment 3 of the present invention; Figure 4 This is a schematic diagram of the overall structure from different perspectives provided in Embodiment 3 of the present invention. Detailed Implementation
[0020] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.
[0021] To better understand the purpose, structure, and function of this invention, the invention will be described in further detail below with reference to the accompanying drawings.
[0022] Example 1 See appendix Figure 1 -Appendix Figure 3 As shown, according to the embodiments of this application, it includes: Feeding component 1, which is provided with multiple feeding channels 105 for feeding materials; Material transfer component 2 is located at the discharge end of the feeding component 1. It is used to receive the material conveyed by the feeding component 1 and transfer it to the discharge component 4. The detection component 3 is an industrial camera, image processing device, etc. The detection component 3 is located above the material transfer component 2. It detects the material on the material transfer component 2, identifies the defective material according to the image information, marks the channel, and feeds back to the discharge component 4 and the collection component 5. The collecting component 5 is located below the discharging component 4 and is used to collect defective materials. When the collecting component 5 receives a signal from the detection component 3, it opens the channel for the defective materials, allowing the materials to fall automatically into the collecting frame.
[0023] The technical effects achieved by the above technical solution in the embodiments of this application are as follows: By using the detection component 3 to perform image recognition on the material on the transfer component 2, the accuracy of identifying defective materials is improved, thereby enabling the defective materials to fall accurately into the collection component 5 for collection, thus providing a foundation for improving the accuracy of identification.
[0024] Example 2 See appendix Figure 1 As shown in the embodiment of this application, the feeding component 1 is a combined structure, including: First support 101, and first conveying component 102 (conveyor belt) is provided on the first support 101. The unloading component 103 has a plate-shaped structure. The unloading component 103 is located at one end of the first conveying component 102 and is inclined to the first conveying component 102. The inclination angle can be selected according to actual needs and is not limited here. The first conveying component 102 and the unloading component 103 are provided with a plurality of feeding channels 105, and the end of the first conveying component 102 is connected to the transfer component 2.
[0025] After the material first enters the feeding channel 105 of the feeding component 103, it is accelerated to enter the feeding channel 105 of the first conveying component 102 due to the inclined arrangement of the feeding component 103, and is then transferred to the transfer component 2 by the first conveying component 102.
[0026] It should be noted that the material transfer component 2 is a modular structure, including: The second bracket 201 is disposed on one side of the first bracket 101 and is fixedly connected to the first bracket 101. The transfer channel 202 is arranged in a matrix on the second support 201, and is fixedly connected to the second support 201, and corresponds to the position of the feed channel 105. Material conveying component 203 is provided on the transfer channel 202 and has a rotating end thereon, which conveys materials by spiral rotation. It should be further explained that the material conveying component 203 is a spiral propulsion device, which is arranged in a matrix on the transfer channel 202. There is a gap between two adjacent material conveying components 203. After the material enters the transfer channel 202, it is spirally propelled by the material conveying components 203 in sequence. During the propulsion process, the material rotates, which is conducive to the detection component 3 to identify the material from all angles and improve the detection accuracy.
[0027] The unloading component 204 is located at the discharge end of the transfer channel 202 and is rotatably connected to the second support 201. It has several unloading channels, which correspond to the positions of the transfer channel 202 and are connected to the discharge component 4. In other words, the unloading component 204 is a plate-shaped structure and is hinged to the second support 201. The unloading angle can be controlled by pre-setting the angle between the unloading component 204 and the second support 201.
[0028] The detection component 3 is respectively located above the transfer channel 202 and above the unloading component 204.
[0029] It should be further noted that the discharge component 4 is a modular structure, including: The third bracket 401 is located on one side of the second bracket 201 and is fixedly connected to the second bracket 201. The second conveying component 402 is mounted on the third support 401 and has several discharge channels on it. The swing component 403 is mounted on the third support 401 and is located at the feed end of the second conveying component 402. It is rotatably connected to the third support 401. The swing ends of the swing component 403 correspond to the discharge channels. The bottom of the swing component 403 is provided with a collecting component 5.
[0030] In other words, the swing component 403 consists of a swing plate and a telescopic cylinder. The swing plate is hinged to the third support 401. The telescopic cylinder supports the swing plate. When the signal from the detection component 3 is received, it performs individual operations based on the material information on the current discharge channel. For example, when there is a defect in the material on the first discharge channel, the swing cylinder retracts, thereby causing the swing plate to swing downward, so that the defective material enters the collection component 5 for collection, and so on.
[0031] The technical effects achieved by the above technical solution in the embodiments of this application are as follows: The feeding component 1 uses a conveyor belt with an inclined feeding plate, suitable for the smooth conveying and alignment of cylindrical wafer barrels; the spiral propulsion device rotates the cylindrical material axially during the transfer process, facilitating omnidirectional image acquisition of the cylindrical outer surface by an industrial camera, with no blind spots in detection; the swing component 403 corresponds to each discharge channel and can be independently controlled according to the detection signal to achieve precise sorting in a single channel, especially suitable for the rapid separation of cylindrical materials in high-speed conveying, avoiding material accumulation or collision; the components are fixedly connected, with strong overall rigidity, suitable for continuous high-intensity production environments; the modular design facilitates disassembly and replacement, reducing maintenance costs.
[0032] Example 3 See appendix Figure 2 -Appendix Figure 4 As shown, the difference between this embodiment and Embodiment 2 is that the material in Embodiment 2 has a columnar structure, while the material in this embodiment has a cap-shaped structure. The feeding component 1 is a combined structure, including: Support frame 104, wherein the support frame 104 is provided with an installation area; Feeding channel 105, the feeding channel 105 is arranged in a matrix on the support frame 104, wherein the feeding channel 105 is an arc-shaped structure; A horizontal channel 106 is located at the end of the feeding channel 105 and is rotatably connected to the feeding channel 105, i.e., it is driven by a motor to control the rotation angle of the horizontal channel 106. Stopping component 107, etc. Figure 2 As shown, the stop component 107 is mounted on the support frame 104 and is rotatably connected to the support frame 104. It is provided with several baffles, wherein the baffles are positioned corresponding to the horizontal channel 106. In other words, it is composed of a motor and a stop frame. The motor drives the stop frame to rotate, thereby connecting with the horizontal channel 106 to block the material.
[0033] The material slides down from the feed channel 105 under the action of gravity and enters the horizontal channel 106. The motor drives the stop frame to align the material and drives the horizontal channel 106 to rotate, causing the horizontal channel 106 to tilt, so that the material enters the transfer component 2.
[0034] It should be noted that the material transfer component 2 is a modular structure, including: The first vertical frame 205 is located on one side of the horizontal channel 106 and is arranged symmetrically. The first rotating component 206 is mounted on the first vertical frame 205 and is rotatably connected to the first vertical frame 205. It is provided with two first receiving ends. In other words, the first rotating component 206 is a motor-driven rotating roller, and a first receiving end is provided on the rotating roller in a symmetrical arrangement. The shape of the first receiving end is selected according to actual needs and is not limited here. The second vertical frame 207 is located between the first vertical frame 205 and the discharge component 4, and they are arranged symmetrically. The second rotating component 208 is mounted on the second vertical frame 207 and is rotatably connected to the second vertical frame 207. It is provided with a plurality of second receiving ends. In other words, the second rotating component 208 is a motor-driven rotating roller with second receiving ends arranged in a ring array on the rotating roller. The shape and number of the second receiving ends are selected according to actual needs and are not limited here. The third vertical frame 209 is located on both sides of the discharge component 4; The height of the second rotating component 208 is lower than the height of the first rotating component 206.
[0035] The detection component 3 is respectively disposed between the first vertical frame 205, the second vertical frame 207 and the third vertical frame 209.
[0036] It should be noted that the discharge component 4 is a modular structure, including: A conical platform 404, wherein a trapezoidal frame 405 is provided on one side of the conical platform 404; The trapezoidal frame 405 is provided with several discharge channels, and a movable discharge valve 406 is provided on the discharge channel; The bottom of the trapezoidal frame 405 is provided with a collecting component 5, which corresponds to the position of the feeding valve 406.
[0037] The discharge valve 406 is controlled by a telescopic cylinder. When it receives a signal from the detection component 3, it sends a signal to the telescopic cylinder on the discharge channel of the defective material, thereby opening the discharge valve 406 on the discharge channel, so that the defective material enters the collection component 5.
[0038] The technical effects achieved by the above technical solution in the embodiments of this application are as follows: The arc-shaped feeding channel 105 and the adjustable horizontal channel 106 adapt to the special shape of the cap-shaped wafer bucket, preventing the material from tipping over during transportation. The stepped rotating roller design keeps the material in a stable posture with the opening facing up or down during transfer, facilitating image acquisition of multiple parts such as the cap edge and the inner side. Multiple cameras are arranged between the first, second, and third vertical frames 209, which can simultaneously perform top, side, and internal visual inspection of cap-shaped materials at different transfer stages, realizing parallel image processing and adapting to the comprehensive quality inspection needs of complex-shaped materials. The trapezoidal frame 405, in conjunction with the pneumatic discharge valve 406, can control the opening of specific discharge channels according to the detection results, realizing the directional dropping and classified recycling of defective materials, avoiding the mixing of qualified and defective products, and facilitating quality traceability and statistical analysis.
[0039] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.
[0040] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.
[0041] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.
[0042] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.
[0043] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A high-speed intelligent defect detection device for wafer barrels based on deep learning, characterized in that, include: A feeding component, wherein the feeding component is provided with multiple feeding channels; A material transfer component is located at the discharge end of the feeding component. It is used to receive the material conveyed by the feeding component and transfer it to the discharge component. A detection component is located above the material transfer component and detects the material on the material transfer component. A collection component, located below the discharge component, is used to collect defective materials.
2. The high-speed intelligent defect detection device for wafer barrels based on deep learning according to claim 1, characterized in that, The feeding component is a modular structure, including: A first support frame, on which a first conveying component is provided; The unloading component is located at one end of the first conveying component and is arranged at an inclination relative to the first conveying component. The first conveying component and the unloading component are provided with several feeding channels, and the end of the first conveying component is connected to the transfer component.
3. The high-speed intelligent defect detection device for wafer barrels based on deep learning according to claim 2, characterized in that, The material transfer component is a modular structure, including: The second bracket is located on one side of the first bracket and is fixedly connected to the first bracket. The transfer channel is arranged in a matrix on the second support, and is fixedly connected to the second support, and corresponds to the position of the feeding channel; A material conveying component is located on a transfer channel and has a rotating end thereon, which conveys materials by rotating a spiral. The unloading component is located at the discharge end of the transfer channel and is rotatably connected to the second support. It is provided with several unloading channels, which correspond to the positions of the transfer channel and are connected to the discharge component. The detection components are respectively located above the transfer conveyor and above the unloading component.
4. The high-speed intelligent defect detection device for wafer barrels based on deep learning according to claim 3, characterized in that, The discharge component is a modular structure, including: The third support is located on one side of the second support and is fixedly connected to the second support. The second conveying component is mounted on the third support and has several discharge channels on it; A swinging component is mounted on a third support and located at the feed end of the second conveying component. It is rotatably connected to the third support, wherein the swinging ends of the swinging component correspond to the discharge channels. The bottom of the swinging component is equipped with a collecting component.
5. The high-speed intelligent defect detection device for wafer barrels based on deep learning according to claim 1, characterized in that, The feeding component is a modular structure, including: A support frame, wherein the support frame is provided with an installation area; The feeding channels are arranged in a matrix on the support frame, wherein the feeding channels are arc-shaped. A horizontal channel is located at the end of the feed channel and is rotatably connected to the feed channel. A stop component is provided on a support frame and is rotatably connected to the support frame. Several baffles are provided on the stop component, wherein the baffles correspond to the positions of the horizontal track.
6. The high-speed intelligent defect detection device for wafer barrels based on deep learning according to claim 5, characterized in that, The material transfer component is a modular structure, including: The first vertical frame is located on one side of the horizontal path and is arranged symmetrically. The first rotating component is mounted on the first vertical frame and is rotatably connected to the first vertical frame. It is provided with two first receiving ends. The second vertical frame is located between the first vertical frame and the discharge component, and they are arranged symmetrically. The second rotating component is mounted on the second vertical frame and is rotatably connected to the second vertical frame. It is provided with multiple second receiving ends. The third vertical frame is located on both sides of the discharge component; The detection components are respectively disposed between the first vertical frame, the second vertical frame and the third vertical frame.
7. The high-speed intelligent defect detection device for wafer barrels based on deep learning according to claim 6, characterized in that, The discharge component is a modular structure, including: A conical truncated platform, wherein a trapezoidal frame is provided on one side of the conical truncated platform; The trapezoidal frame is provided with several discharge channels, and the discharge channels are provided with movable discharge valves; The bottom of the trapezoidal frame is equipped with a collecting component, which corresponds to the position of the discharge valve.
8. The high-speed intelligent defect detection device for wafer barrels based on deep learning according to claim 6, characterized in that, The height of the second rotating component is lower than the height of the first rotating component.