An automatic trimming method and system based on visual detection
Patent Information
- Application Number
- CN202611313088.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-27
- Publication Date
- 2026-10-23
AI Technical Summary
然而，这种方式存在诸多缺陷：首先，修边效果高度依赖于操作人员的技术熟练度和精神状态，不同的工人或同一工人在不同时间的操作，会导致修边位置、切口角度和表面光滑度产生显著波动，产品一致性和尺寸精度难以保证，尤其无法满足现代汽车工业对严格公差的追求；其次，手动修边速度缓慢，远远无法匹配自动化生产线的节拍需求，而且需要大量人工参与，人力成本居高不下；再者，操作人员长时间重复同一动作，极易引发腱鞘炎等职业病，同时使用锋利刀具还存在割伤的安全风险；最后，对于形状复杂、曲率变化大的产品边缘，手动修边难度大、耗时长，且更容易造成瑕疵甚至报废
1.本发明采用基于CNC控制原理的工业机器人，通过视觉识别自动判断产品类型，并自适应切换抓取方式（吸盘/夹爪）和加工工位（固定工位/多工位转盘）。对于大平面产品采用吸盘抓取+固定工位高速打磨，保证节拍；对于复杂曲面小尺寸产品采用夹爪抓取+转盘并行作业，保证可达性。这种设计在同一产线上同时实现了CNC级的重复定位精度（±0.1mm以内）和机器人级的作业柔性，避免了手动修边的质量波动和刀模修边的高额换型成本，显著提升了产品一致性和生产效率；
Smart Images
Figure CN122956623A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of automotive interior parts manufacturing technology, specifically to an automatic trimming method and system based on visual inspection. Background Technology
[0002] In the production of automotive polyurethane foam products (such as dashboards, armrests, headrests, and other interior parts), due to the gap between the mold parting surfaces or the injection pressure, a thin layer of excess material (also known as flash or burrs) inevitably forms on the edges of the molded product. This excess material must be precisely removed to ensure the appearance quality of the product and the accuracy of subsequent assembly.
[0003] Currently, a widely adopted trimming method in the industry is manual trimming, where operators use utility knives, scissors, or specialized tools to manually remove excess material from product edges based on experience and visual judgment. However, this method has several drawbacks: First, the trimming effect is highly dependent on the operator's skill level and mental state. Different workers or the same worker operating at different times can lead to significant fluctuations in trimming position, cut angle, and surface smoothness, making it difficult to guarantee product consistency and dimensional accuracy, especially failing to meet the stringent tolerance requirements of the modern automotive industry. Second, manual trimming is slow, far from matching the pace demands of automated production lines, and requires a large amount of manual labor, resulting in high labor costs. Third, operators who repeat the same actions for extended periods are highly susceptible to occupational diseases such as tenosynovitis, and the use of sharp tools also poses a risk of cuts. Finally, for product edges with complex shapes and large curvature variations, manual trimming is difficult, time-consuming, and more likely to cause defects or even scrap.
[0004] Another common trimming method is die-cutting trimming (or die-cutting trimming), which uses stamping equipment and customized steel die-cutting molds to cut away excess material. While this method improves trimming efficiency to some extent, its limitations are also obvious: On the one hand, a set of high-precision die-cutting molds needs to be customized for each type or category of product shape, making the molds very expensive. Furthermore, if the product design changes, the die-cutting molds must be remade, resulting in additional costs and time investment. On the other hand, the die-cutting molds wear down and become dull during use, requiring regular maintenance, grinding, or replacement, increasing equipment downtime and maintenance costs. In addition, die-cutting trimming may compress or stretch soft foam materials during punching, resulting in uneven or slightly deformed trimmed edges, and the perpendicularity of the cut surface is difficult to guarantee. More importantly, die-cutting molds are mainly suitable for trimming flat or near-flat surfaces. For three-dimensional curved surfaces or complex irregular structures, die-cutting trimming is difficult to achieve, or requires extremely complex, multi-step punching designs, severely limiting its applicability.
[0005] To address the shortcomings of traditional trimming methods, some companies have begun to explore automated trimming in recent years. However, most existing automated trimming solutions are designed for single products and processes, lacking the ability to recognize and adapt to different product characteristics. For example, when production lines need to process large flat products and complex curved small-sized products in mixed flow, robots often have to use fixed gripping methods (such as using only suction cups or only grippers) and fixed processing stations (such as only configuring fixed stations or only configuring turntables), resulting in either unstable gripping of complex curved products or low efficiency for large flat products. In addition, the accuracy of trimming also needs to be improved. Post-trimming quality inspection (such as checking for metal foreign objects, weight, and dimensions), gluing, labeling, and sorting are usually separate processes, requiring manual transfer or independent equipment connections, failing to form a continuous, closed-loop intelligent manufacturing process. Summary of the Invention
[0006] In view of the above-mentioned shortcomings of the prior art, the present invention provides an automatic trimming method and system based on visual detection to solve the problems mentioned in the background art.
[0007] To achieve the above objectives, the present invention provides the following technical solution: an automatic edge trimming method based on visual detection, comprising the following steps: Step 1: Place the product to be trimmed into the conveyor unit; Step 2: The visual recognition unit acquires product images, and the control system identifies the product type; the control system determines whether the current product is a large flat product or a small-sized product with a complex curved surface based on the recognition results; Step 3: Based on the judgment result, the feeding unit places the product in the trimming station corresponding to the product type. The feeding unit includes a suction cup gripping mechanism and a gripper gripping mechanism. The suction cup gripping mechanism can place large flat products in the large flat trimming unit, and the gripper gripping mechanism can place small-sized products with complex curved surfaces in the small-sized curved surface trimming unit. Step 4: The large flat surface trimming unit or the curved surface small size trimming unit trims the product according to the pre-stored trajectory program; The trimming robots in the large planar trimming unit and the curved small-size trimming unit adopt the CNC control principle. The trimming trajectory is set through offline programming, and the control system performs interpolation calculations and linkage control on each joint of the robot.
[0008] Furthermore, after step 4, the following is also included: Step 5: The product enters the dust removal and weighing unit for surface dust removal and weight detection; Step 6: The product enters the automatic glue application unit for glue application; Step 7: The product enters the automatic labeling unit and an identification label is affixed; Step 8: The product enters the CCD vision inspection unit to inspect the quality of the trimmed edges, the position of the adhesive, and the label information; Step 9: The automatic sorting unit sorts products to the qualified product channel or different categories of defective product channels based on the combined results of weight detection and visual inspection.
[0009] Furthermore, the large flat surface trimming unit adopts a fixed trimming station; the curved surface small size trimming unit adopts a multi-station turntable trimming station.
[0010] Furthermore, the multi-station turntable trimming station includes three independent trimming stations for parallel processing of multiple small-sized curved surface products.
[0011] Furthermore, for the small-sized trimming unit on curved surfaces, a manual assembly unit is provided between the dust removal and weighing unit and the automatic glue application unit for installing small plastic clips.
[0012] An automatic edge trimming system based on visual detection, used to implement an automatic edge trimming method based on visual detection, includes: A conveyor unit is used to transport products to be trimmed. A visual recognition unit, located above the conveyor unit, is used to capture product images and identify product types; The feeding unit includes a suction cup gripping mechanism and a gripper gripping mechanism, which are used to feed different types of products; The large flat surface trimming unit uses a suction cup gripping mechanism for feeding materials and is used for trimming large flat surface products. The curved surface small-size trimming unit is fed by a gripper-type grasping mechanism and is used for trimming operations of curved surface small-size products. The control system is electrically connected to the conveying unit, vision recognition unit, feeding unit, large flat surface trimming unit, and curved surface small size trimming unit, and is used to execute the work of each unit.
[0013] Furthermore, it also includes a dust removal and weighing unit, an automatic glue application unit, an automatic labeling unit, a CCD vision inspection unit, and an automatic sorting and unloading unit arranged sequentially along the product conveying direction.
[0014] Furthermore, it also includes a dust collection system, which is connected to the large flat trimming unit and the small curved trimming unit to collect the dust generated during the trimming process.
[0015] Furthermore, the visual recognition unit includes a first visual detection component and a second visual detection component. The first visual detection component is suitable for accurate identification of detailed features of complex curved surface products, while the second visual detection component is suitable for rapid extraction of the overall outline of large flat surface products.
[0016] This invention provides an automatic edge trimming method and system based on visual detection, which has the following beneficial effects: 1. This invention employs an industrial robot based on CNC control principles. It automatically determines the product type through visual recognition and adaptively switches between gripping methods (suction cup / gripper) and processing stations (fixed station / multi-station turntable). For large flat products, suction cup gripping combined with high-speed grinding at a fixed station ensures cycle time; for complex curved surface products with small dimensions, gripper gripping combined with a turntable in parallel operation ensures accessibility. This design simultaneously achieves CNC-level repeatability (within ±0.1mm) and robot-level operational flexibility on the same production line, avoiding quality fluctuations from manual trimming and the high changeover costs of die-cutting trimming, significantly improving product consistency and production efficiency. 2. This invention integrates functional modules such as metal detection, weighing, CCD vision inspection, automatic gluing, automatic labeling, and automatic sorting and unloading, forming a closed-loop intelligent manufacturing process. Weighing monitors the amount of material removed during trimming in real time; CCD detection detects the quality of the trimmed edges, the gluing position, and label information; all data is bound to the product ID for quality traceability; and finally, the sorting mechanism automatically distinguishes qualified products from various types of defective products. The entire process requires no manual intervention, freeing up operators, eliminating the risk of work-related injuries and occupational diseases, and reducing labor costs and inter-process transfer costs. Furthermore, when switching products, only different CNC programs need to be called, and the robot automatically changes the end effector, eliminating the need for customized expensive die-cutting molds. Changeover time is reduced to a few seconds, and the unit product trimming cost is significantly lower than traditional methods over long-term operation. 3. This invention features a high-efficiency dust collection system that centrally collects and treats the dust generated during trimming, improving workshop air quality and meeting occupational health and safety standards. Simultaneously, each functional module (adhesive application, labeling, testing, sorting, etc.) adopts a modular design, allowing for flexible addition, deletion, or upgrades based on product process requirements. It is easily expandable to other similar processing scenarios involving trimming, grinding, and deburring of soft or semi-rigid materials, offering excellent investment protection. Attached Figure Description
[0017] Figure 1 This is a flowchart of an automatic trimming system based on visual inspection according to the present invention; Figure 2 For the present invention Figure 1 A magnified view of a portion of the image.
[0018] In the diagram: 1. Conveying unit; 2. Vision recognition unit; 21. First vision inspection component; 22. Second vision inspection component; 3. Feeding unit; 31. Gripper gripping mechanism; 32. Suction cup gripping mechanism; 4. Curved small-size trimming unit; 5. Large flat trimming unit; 6. Dust removal and weighing unit; 7. Automatic glue application unit; 8. Automatic labeling unit; 9. CCD vision inspection unit; 10. Automatic sorting and unloading unit; 11. Automatic sorting and unloading unit. Detailed Implementation
[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0020] The present invention will now be described in detail through specific embodiments, as follows: According to the instruction manual Figures 1-2 As can be seen, the automatic trimming system based on visual detection of the present invention includes: Conveying Unit 1: Uses a feeding conveyor belt to transport products to be processed.
[0021] Visual recognition unit 2: Located above the feeding conveyor belt, it includes a high-precision industrial camera and an image processing module, used to acquire product images and identify product type (such as shape outline and size characteristics), and send the recognition results to the control system.
[0022] Of course, the visual recognition unit 2 can also include two independent visual detection components, namely the first visual detection component 21 and the second visual detection component 22. The first visual detection component 21 is configured for small-sized products with complex curved surfaces, equipped with a high-resolution macro lens and a ring light source, suitable for accurate identification of detailed features (such as surface curvature and edge shape) of complex curved surface products. The second visual detection component 22 is configured for large-flat products, equipped with a wide-angle lens and a structured light source, suitable for rapid extraction of the overall contour of large-flat products. During detection, the product first passes through the first visual detection component 21 to determine if it is a small-sized curved product. If so, it is directly gripped by the gripper-type grasping mechanism 31 to the small-sized curved surface trimming unit 4; otherwise, it is determined to be a large-flat product. However, the second visual detection component 22 will also inspect products that have passed through the first visual detection component 21 and have been determined to be large-flat products to ensure the accuracy of the detection.
[0023] Loading Unit 3: This refers to the loading / unloading robot, which is a six-axis articulated industrial robot. Loading Unit 3 includes two types of robots: one with a suction cup gripping mechanism 32 and the other with a gripper gripping mechanism 31. Loading Unit 3 is responsible for picking up products from the loading conveyor belt and placing them at the corresponding trimming station, as well as removing the products after trimming and placing them on the downstream conveyor belt.
[0024] Large plane trimming unit 5: Large plane trimming unit 5 adopts a fixed trimming station and is equipped with an independent six-axis articulated industrial robot. The robotic arm carries a grinding head and follows a trajectory to process large plane products.
[0025] Curved Surface Small-Dimensional Trimming Unit 4: This unit employs a multi-station rotary table trimming station, comprising three independent trimming stations, each equipped with an independent six-axis articulated industrial robot. A high-speed grinding spindle is fixedly mounted at the end of each trimming robot, with a milling cutter or grinding head clamped at its front end for precise grinding and trimming of excess material from the product edges. The three trimming robots can simultaneously process three different curved surface small-dimensional products, achieving parallel operation.
[0026] In addition, the system includes a dust removal and weighing unit 6, an automatic gluing unit 7, an automatic labeling unit 8, a CCD vision inspection unit 9, and an automatic sorting and unloading unit 10, arranged sequentially along the product conveying direction. The dust removal and weighing unit 6 performs surface dust removal and weight detection on the products; the automatic gluing unit 7 applies glue to the products; the automatic labeling unit 8 affixes identification labels to the products; the CCD vision inspection unit 9 detects the quality of the trimmed edges, the glue application position, and the label information; and the automatic sorting and unloading unit 10 sorts the products to the qualified product channel or different categories of defective product channels based on the combined results of weight detection and vision inspection.
[0027] Dust collection system: installed near fixed trimming stations and multi-station rotary trimming stations to collect dust generated during trimming.
[0028] Control system: PLC or industrial computer, electrically connected to all the above units, used to coordinate and control the actions of each unit and execute the trimming method steps.
[0029] To ensure the long-term stable, safe, and environmentally friendly operation of the automated trimming process of this invention, the system also integrates the following auxiliary systems, including: Positioning and clamping system: Each trimming station is equipped with vacuum adsorption or pneumatic clamping to fix the product; Dust removal and cooling system: Dust collection hood + central dust collector removes dust; cooling airflow prevents materials from overheating and sticking to the blade; Visual inspection system (optional): Scans the location, quantity, and thickness of overflow material before trimming and provides feedback on the fine-tuning trajectory; scans the trimming quality (burrs, gaps, etc.) after trimming. Safety protection system: safety fences, light curtains, emergency stops, to ensure personnel safety.
[0030] Example 1 (for small-sized products with complex curved surfaces); This embodiment uses a car armrest (complex curved surface, small size) as an example for illustration.
[0031] (I) Methods and Steps Step 1: Place the product to be trimmed into conveyor unit 1.
[0032] Specifically, operators or automated feeding systems place small, complex curved products (car handrails) onto the loading conveyor belt. The conveyor belt then transports the products to subsequent workstations at a set speed.
[0033] Step S2: The visual recognition unit 2 acquires product images and controls the system to identify the product type.
[0034] Specifically, when the product is conveyed to the visual recognition unit 2, the industrial camera captures an image of the product, and the image processing module extracts features such as the product's outline and curvature. The control system compares the extracted features with a pre-stored product database, and the recognition result determines whether the current product is a large flat product or a small-sized product with a complex curved surface. In this embodiment, the current product is determined to be a "small-sized product with a complex curved surface".
[0035] Step S3: Based on the judgment result of step S2, the control system issues an instruction to the loading unit 3, and the loading unit 3 places the product in the trimming station corresponding to the product type.
[0036] Specifically, in this embodiment, a gripper-type grasping mechanism 31 is used for grasping to ensure grasping stability and positioning accuracy. The robot with the gripper-type grasping mechanism 31 transports the complex curved surface small-sized product to the curved surface small-sized trimming unit 4.
[0037] Step S4: Trim the edges of the product.
[0038] Specifically, the curved small-size trimming unit 4 employs a multi-station rotary trimming station, including three independent trimming stations. Each trimming station is equipped with a product positioning fixture to firmly fix the product and ensure that it does not shift during the trimming process. The three trimming robots operate independently, each with a high-speed grinding spindle fixedly mounted at its end. When a product is placed at a trimming station, the trimming robot at that station starts and precisely grinds and trims the excess material at the product's edges according to the trajectory program pre-stored in the control system. During the trimming process, a dust collection system is activated to remove the generated dust in real time.
[0039] Three trimming stations can process three different products simultaneously, achieving true parallel operation. The control system dynamically monitors the processing status of each station, allocating newly arrived products to idle stations to maximize equipment utilization and output efficiency. The motion control of the trimming robot adopts the CNC (Computer Numerical Control) principle: technicians use specialized software to generate trimming tool trajectories on a computer based on the product's 3D CAD model, which are then converted into code executable by the robot. The controller performs real-time interpolation calculations and linkage control on each joint, achieving a trimming accuracy of ±0.1mm. During the trimming process, force sensors can be integrated to monitor cutting forces and adjust trimming speed and path in real time to adapt to subtle material differences and improve surface quality.
[0040] Step S5: Remove dust and weigh the trimmed product.
[0041] Specifically, after trimming, the material is unloaded onto a conveyor belt via a gripper-type grasping mechanism 31 and enters the dust removal and weighing unit 6. First, high-pressure airflow blows away the dust adhering to the product surface, and then a high-precision weighing sensor detects the product weight. The control system compares the measured weight with the standard weight range: if the weight exceeds the standard (too much or too little trimming), it is recorded as unqualified. The weighing data provides real-time feedback on the stability of the trimming process, providing a basis for adjusting process parameters. Furthermore, a metal detector can be added to the dust removal and weighing unit 6 to detect metal foreign objects.
[0042] Step S6: Automatically apply glue to the dust-removed and weighed product.
[0043] Specifically, the product continues to be conveyed to the automatic glue application unit 7. This unit includes a glue supply system and a glue application valve. The loading / unloading robot (or a dedicated glue application robot) picks up the product and applies adhesive to the designated location on the product (such as the snap-fit mounting surface) according to a pre-stored trajectory.
[0044] Step S7: Automatically label the product after applying the adhesive.
[0045] Specifically, the product enters the automatic labeling unit 8. The labeling machine prints and affixes identification labels (including product model, production date, serial number, etc.) to the product surface based on the product information issued by the control system.
[0046] Step S8: Perform CCD visual inspection on the labeled product.
[0047] Specifically, the product enters CCD vision inspection unit 9. This unit is equipped with multiple industrial cameras to capture images of the trimmed edges, adhesive application areas, and labels from different angles. The image processing software automatically detects the following: Check for burrs, gaps, or unevenness at the trimmed edges; Is the adhesive application path continuous and is the position accurate? Are the labels complete? Is the information readable?
[0048] The test results are linked to the product ID and stored in the quality traceability database.
[0049] Step S9: Automatically sort and remove products after CCD vision inspection.
[0050] Specifically, the products eventually arrive at the automatic sorting and unloading unit 10. The control system integrates the weight data from step S5 and the visual inspection results from step S8 to make a sorting decision: If all test items are qualified, the sorting organization will push the product into the qualified product channel; If the weight is not up to standard but the appearance is up to standard, push it into the weight defective products channel; If the appearance is not up to standard (edge trimming defects, glue application defects, etc.), push it into the appearance defective product channel; If the label or adhesive is severely missing, push it into the repair channel.
[0051] This concludes the complete process of trimming, inspection, gluing, labeling, and sorting for a small product with a complex curved surface.
[0052] In addition, if necessary, a manual assembly unit 11 can be set between the dust removal and weighing unit 6 and the automatic glue application unit 7 for installing small plastic clips.
[0053] Example 2 (for large flat surface products) This embodiment uses an automotive dashboard (large flat surface, relatively regular shape) as an example for illustration. The main differences from Embodiment 1 are: different product type, different gripping method, and different processing station.
[0054] Step S1: Place the product to be trimmed in the conveying unit 1 (same as step 1 in Example 1).
[0055] Step S2: The visual recognition unit 2 acquires product images and controls the system to identify the product type.
[0056] After the visual recognition unit 2 acquires the image, the control system determines that the current product is a "large flat product".
[0057] Step S3: Based on the judgment result of step S2, the control system issues an instruction to the loading unit 3, and the loading unit 3 places the product in the trimming station corresponding to the product type.
[0058] Specifically, in this embodiment, a suction cup gripping mechanism 32 is used for gripping. The surface of the large flat product is flat, and the suction cup can form a stable adhesion, resulting in fast gripping speed without damaging the product surface. The robot using the suction cup gripping mechanism 32 transports the large flat product to the large flat trimming unit 5.
[0059] Step S4: Trim the edges of the product.
[0060] A suction cup gripping mechanism 32 precisely places the product onto a fixed trimming station. This station is equipped with a vacuum adsorption or mechanical clamping device to firmly secure the product. The control system calls upon a trimming trajectory program designed for large flat products; this program is typically planned as a relatively simple planar or high-curvature path. The motion control of the trimming robot also employs CNC principles. The high-speed grinding spindle, fixed at the end, trims the product fixed at the station according to a pre-stored trajectory, allowing for high feed rates and efficient grinding.
[0061] Steps S5 to S9, which involve dust removal and weighing, automatic glue application, automatic labeling, CCD vision inspection, and automatic sorting and unloading of the trimmed products, are essentially the same as steps S5 to S9 in Example 1. Furthermore, large flat products typically do not require manual assembly of clips, enabling fully automated, unmanned production.
[0062] Those skilled in the art should understand that the specific configurations described in the above embodiments (such as the number of six-axis robots, the number of trimming stations, and the gluing method) are merely illustrative and not intended to limit the invention. In practical applications, reasonable modifications can be made according to product characteristics and production capacity requirements. The number of trimming robots can be configured to be two, four or more, not limited to three, depending on production capacity requirements; Five-axis or seven-axis robots can be used instead of six-axis robots; Automatic glue application can be achieved through spraying, dispensing, or rolling. The CCD detection unit can be equipped with three-dimensional laser contour measurement as needed; For products that do not require adhesive or labeling, the corresponding unit can be skipped.
[0063] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. An automatic edge trimming method based on visual detection, characterized in that, Includes the following steps: Step 1: Place the product to be trimmed into the conveyor unit (1); Step 2: The visual recognition unit (2) acquires product images and controls the system to identify the product type; the control system determines whether the current product is a large flat product or a small-sized product with a complex curved surface based on the recognition results; Step 3: Based on the judgment result, the loading unit (3) places the product in the trimming station corresponding to the product type. The loading unit (3) includes a suction cup gripping mechanism (32) and a gripper gripping mechanism (31). The suction cup gripping mechanism (32) can place large flat products in the large flat trimming unit (5), and the gripper gripping mechanism (31) can place complex curved small-sized products in the curved small-sized trimming unit (4). Step 4: The large flat surface trimming unit (5) or the curved surface small size trimming unit (4) trims the product according to the pre-stored trajectory program; The trimming robots in the large planar trimming unit (5) and the curved small-size trimming unit (4) adopt the CNC control principle. The trimming trajectory is set by offline programming, and the control system performs interpolation calculations and linkage control on each joint of the robot.
2. The automatic edge trimming method based on visual detection according to claim 1, characterized in that, Step 4 is followed by: Step 5: The product enters the dust removal and weighing unit (6) for surface dust removal and weight detection; Step 6: The product enters the automatic glue application unit (7) for glue application; Step 7: The product enters the automatic labeling unit (8) and the identification label is affixed; Step 8: The product enters the CCD vision inspection unit (9) to inspect the quality of the trimmed edges, the glue application position, and the label information; Step 9: Automatic sorting and unloading unit (10) sorts products to qualified product channels or different categories of defective product channels based on the combined results of weight detection and visual inspection.
3. The automatic edge trimming method based on visual detection according to claim 1, characterized in that, The large flat trimming unit (5) adopts a fixed trimming station; the curved small-size trimming unit (4) adopts a multi-station turntable trimming station.
4. The automatic edge trimming method based on visual detection according to claim 3, characterized in that, The multi-station turntable trimming station includes three independent trimming stations for parallel processing of multiple small-sized curved surface products.
5. The automatic edge trimming method based on visual detection according to claim 2, characterized in that, For the small-sized trimming unit (4) on curved surfaces, a manual assembly unit (11) is set between the dust removal and weighing unit (6) and the automatic glue application unit (7) for installing small plastic clips.
6. An automatic edge trimming system based on visual detection, used to implement the automatic edge trimming method based on visual detection as described in any one of claims 1 to 5, characterized in that, include: Conveying unit (1) is used to convey products to be trimmed; A visual recognition unit (2) is located above the conveying unit (1) and is used to collect product images and identify product types. The feeding unit (3) includes a suction cup gripping mechanism (32) and a gripper gripping mechanism (31) for feeding different types of products; The large flat surface trimming unit (5) is loaded with material through a suction cup gripping mechanism (32) and is used for trimming large flat surface products. The curved small-size trimming unit (4) is fed by a gripper-type grasping mechanism (31) and is used for trimming of curved small-size products. The control system is electrically connected to the conveying unit (1), the vision recognition unit (2), the feeding unit (3), the large flat trimming unit (5), and the curved small-size trimming unit (4) to perform the work of each unit.
7. The automatic trimming system based on visual inspection according to claim 6, characterized in that, It also includes a dust removal and weighing unit (6), an automatic glue application unit (7), an automatic labeling unit (8), a CCD vision inspection unit (9), and an automatic sorting and unloading unit (10) arranged sequentially along the product conveying direction.
8. The automatic trimming system based on visual inspection according to claim 6, characterized in that, It also includes a dust collection system, which is connected to the large flat trimming unit (5) and the curved small-size trimming unit (4) to collect the dust generated during the trimming process.
9. The automatic trimming system based on visual inspection according to claim 6, characterized in that, The visual recognition unit (2) includes a first visual detection component (21) and a second visual detection component (22). The first visual detection component (21) is suitable for accurate identification of detailed features of complex curved surface products, and the second visual detection component (22) is suitable for rapid extraction of the overall outline of large flat surface products.