Cutting machine module alignment detection method and system

Abstract

The invention relates to the technical field of semiconductor manufacturing equipment, in particular to a cutting machine module alignment detection method and system. The core of the method is as follows: in a teaching stage of module alignment detection, in a coarse adjustment pattern inspection link, a plurality of detection rounds including normal, bright and dark light source conditions are set; in each turn, a pattern comparison process including rotation comparison and spiral inspection is executed; finally, only when the pattern comparison under all the light source detection rounds meets the preset tolerance requirement, the teaching is judged to be qualified. The system comprises a light source control module for realizing the multi-turn light source control, a comparison module for executing pattern comparison and a judgment module for final qualification judgment. The method has the advantages that light source fluctuation can be actively simulated, the risk of unreliable recognition caused by light source decline or interference is recognized and eliminated in advance in the teaching stage, and therefore module alignment preciseness and the yield and stability of the production process are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of semiconductor manufacturing equipment technology, and specifically to a method and system for aligning and detecting cutting machine modules. Background Technology

[0002] In the back-end processes of semiconductor manufacturing, wafer dicing is a crucial step in separating a wafer containing numerous chips into individual chips. During this process, the dicing machine must precisely cut along pre-designed dicing tracks on the wafer; any deviation can directly lead to damage to chip functionality or decreased reliability, resulting in significant economic losses. Therefore, module alignment before dicing—that is, the accurate identification and positioning of dicing track marks (feature points) on the wafer by the vision system—is the primary prerequisite for ensuring dicing accuracy and product yield.

[0003] Currently, the most common approach in this field is module alignment based on fixed light source parameters. Specifically, during equipment debugging (training) phases, operators set a fixed set of axial and ring light intensity parameters based on experience. In subsequent mass production, the cutting machine's vision system consistently uses these fixed light source conditions to illuminate the wafer surface and acquires feature point images via image sensors, comparing them with a preset template to complete alignment. While simple and direct, this method suffers from inherent and insurmountable defects. First, it cannot effectively monitor and warn of performance degradation in the light source system. Light source components, especially light-emitting devices, inevitably experience luminous efficiency degradation over time. When the actual light source intensity falls below the set value, the contrast of the acquired feature point images decreases, details become blurred, and the vision system is prone to misidentification or failure. This potential risk cannot be detected before production begins under the traditional fixed-parameter mode, potentially leading to batch-wide quality incidents. Second, this method exhibits insufficient alignment rigor when faced with similar interfering feature points on the wafer surface. Because the light source conditions are singular and fixed, the vision system may misidentify similar patterns in non-target areas as genuine feature points, leading to incorrect definition of the cutting path. Current technology lacks a mechanism to actively simulate and detect the impact of light source fluctuations on recognition stability during the teaching phase. This leaves the entire production process built on a potentially unstable foundation, posing a continuous threat to the quality and reliability of high-end chip manufacturing.

[0004] Therefore, existing technologies still need further development. Summary of the Invention

[0005] The purpose of this invention is to overcome the above-mentioned technical deficiencies and provide a method and system for aligning and detecting cutting machine modules, so as to solve the problems existing in the prior art.

[0006] To achieve the above-mentioned technical objectives, according to a first aspect of the present invention, the present invention provides a method for alignment detection of a cutting machine module, comprising: S100, The teaching stage for module alignment detection, the teaching stage includes coarse pattern inspection; S200. In the coarse pattern inspection, multiple light source detection rounds are set, including normal light source conditions, brightened light source conditions, and darkened light source conditions. S300. In each round of light source detection, pattern comparison is performed; S400. Based on the results of the pattern comparison, determine whether the module alignment is qualified.

[0007] Specifically, the brightening and darkening light source conditions are obtained by adjusting the light source intensity of the normal light source conditions by a preset percentage.

[0008] Specifically, the preset percentage adjustment applies to both axial light and ring light.

[0009] Specifically, the pattern comparison includes rotational comparison and spiral inspection.

[0010] Specifically, the rotation comparison is a coordinate error comparison performed within a preset angle range of the material sheet rotation.

[0011] Specifically, the spiral check involves comparing coordinate errors along a preset spiral path.

[0012] Specifically, determining whether the module alignment is qualified includes: When the pattern comparison in all light source detection rounds meets the preset error threshold, the coarse adjustment pattern inspection is deemed qualified.

[0013] Specifically, the preset error thresholds include the coordinate error threshold for rotation comparison and the coordinate error threshold for spiral inspection.

[0014] Specifically, the number of light source detection rounds is three, corresponding to the normal light source condition, the brightened light source condition, and the darkened light source condition, respectively.

[0015] According to a second aspect of the present invention, a system for aligning and detecting a cutting machine module is provided, comprising: The light source control module is used to set up multiple light source detection rounds during the teaching phase of module alignment detection. The multiple light source detection rounds include normal light source conditions, brightened light source conditions, and darkened light source conditions. The comparison module is used to perform pattern comparison in each round of light source detection; The determination module is used to determine whether the module alignment is qualified based on the pattern comparison result.

[0016] Beneficial effects: The novel cutting module alignment detection method and system provided by this invention brings significant and multifaceted benefits compared to the prior art, fundamentally improving the reliability of the alignment process and the stability of production quality.

[0017] The most significant benefit of this invention lies in its paradigm shift from "passive response" to "proactive prevention." By introducing a multi-round light source floating detection mechanism into the coarse-tuning pattern inspection during the teaching phase, this invention can proactively simulate potential enhancement and attenuation boundary conditions of the light source before mass production begins, conducting rigorous "stress testing" or "boundary capability verification" of the visual recognition system. This allows any potential decrease in recognition reliability caused by changes in light source performance to be detected and intercepted in its early stages, thus completely preventing the spread of hidden quality risks caused by slow light source degradation to the mass production stage, greatly ensuring product yield.

[0018] Secondly, this invention significantly enhances the rigor and robustness of the alignment process. By incorporating precise alignment algorithms such as rotational comparison and spiral checking into each light source detection round, the scheme not only verifies the system's ability to "find" feature points under different lighting conditions, but also tests the stability of its ability to accurately locate the precise coordinates of feature points. This dual verification mechanism ensures the strong adaptability of the alignment results to changes in lighting conditions, effectively avoiding alignment coordinate shifts introduced by slight changes in the feature point image morphology, resulting in extremely accurate and reliable final cut path positions.

[0019] Third, this invention significantly improves the automation level and quality control capabilities of the production process. The method transforms the previously vague judgment of light source reliability, which relied on operator experience, into an automated and standardized process based on clear thresholds and logical decisions. This reduces reliance on manual labor, ensures a high degree of consistency in execution standards across different equipment and operators, and enhances the overall consistency and controllability of the production line.

[0020] Finally, while achieving the aforementioned superior reliability, this invention also considers production efficiency. Through optimized design (such as setting the number of inspection rounds to three), this solution minimizes the time required for the teaching phase while ensuring sufficient verification results, avoiding a significant impact on equipment uptime and demonstrating good engineering practicality and economy. In summary, this invention provides a highly reliable, high-precision, and automated alignment inspection solution for semiconductor wafer dicing, possessing extremely high industrial application value. Attached Figure Description

[0021] Figure 1This is a flowchart illustrating the method for aligning and detecting a cutting machine module provided in a specific embodiment of the present invention; Figure 2 This is a schematic diagram of the system composition of the cutting machine module alignment detection system provided in a specific embodiment of the present invention. Detailed Implementation

[0022] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Based on the embodiments in this application, other similar embodiments obtained by those skilled in the art without creative effort should all fall within the scope of protection of this application. Furthermore, directional terms mentioned in the following embodiments, such as "up," "down," "left," and "right," are only for reference to the directions in the accompanying drawings; therefore, the directional terms used are for illustrative purposes and not for limiting the invention.

[0023] The present invention will be further described below with reference to the accompanying drawings and preferred embodiments.

[0024] Please see Figure 1 This invention provides a method for alignment detection of a cutting machine module, comprising: S100, The teaching phase for module alignment detection, which includes coarse pattern inspection.

[0025] It should be further explained that the core of the method claimed in this invention lies in the fundamental improvement made to the existing dicing machine module alignment process, especially in the coarse pattern inspection step during the teaching phase. Specifically, the "teaching phase" refers to the process of setting and debugging the parameters of the visual recognition system of the dicing equipment after changing to a new wafer batch or product model and before formally putting it into mass production. The "coarse pattern inspection" is a key step in this phase, aiming to initially locate and identify the feature patterns (i.e., feature points) on the wafer dicing track.

[0026] S200. In the coarse pattern inspection, multiple light source detection rounds are set, including normal light source conditions, brightened light source conditions, and darkened light source conditions.

[0027] It should be further explained that the innovation of this invention lies in expanding the detection under traditional single-light source conditions to detection rounds under multiple different light source conditions. Specifically, in addition to using preset normal light source conditions (i.e., standard light source intensity set according to workpiece data), detection conditions simulating potential changes in the light source must also be added, namely "brighter light source conditions" and "darkener light source conditions." This is equivalent to setting a more rigorous "qualification test" for the device's visual recognition system.

[0028] S300. In each round of light source detection, a pattern comparison is performed.

[0029] It should be further explained that in each round, the system needs to execute a pattern comparison algorithm, such as acquiring images through an image sensor (CCD) and performing correlation matching or edge feature extraction comparison with preset feature templates.

[0030] S400. Based on the results of the pattern comparison, determine whether the module alignment is qualified.

[0031] It should be further explained that, ultimately, the training is only deemed successful and the equipment is allowed to proceed to the subsequent fine-tuning or direct production process if the comparison results of all rounds meet the requirements. If any round fails, the equipment is deemed unqualified, and the system will issue an alarm and interrupt the cutting process.

[0032] Understandably, this kind of pre-production multi-condition testing can detect and intercept the risk of decreased recognition reliability caused by aging, decay or instability of light source components before production begins, thereby avoiding batch quality accidents caused by alignment errors from the source and significantly improving production yield and equipment operation robustness.

[0033] Furthermore, a key improvement of this invention lies in extending the traditional single-stage light source condition detection into multiple consecutive rounds of light source detection. In specific implementation, after the operator initiates the training program on the device's human-machine interface (HMI), the system will prompt for a coarse pattern check. At this time, the system will automatically execute three detection rounds in a preset sequence: ① First round (normal lighting conditions): Use the standard lighting parameters preset in the workpiece data (Recipe). For example, set the coaxial light intensity to 2000 units and the ring light intensity to 1500 units. This round serves as a baseline test; ② Second round (light source enhancement conditions): The system automatically adjusts the intensity upwards based on the light source intensity in the first round; ③ Third round (dim light source conditions): The system automatically adjusts the light source intensity downwards based on the light source intensity in the first round.

[0034] Furthermore, each round involves independent pattern comparison. This "pattern comparison" is an automated image processing procedure, the core of which is a template matching algorithm. The system calls up pre-stored feature point template images and compares them with real-time images acquired by the current CCD camera under specific lighting conditions, calculating similarity and locating the coordinates of the feature points.

[0035] Understandably, the final pass / fail determination is based on the comparison results of all rounds. Only if all three rounds of comparison are successful will the system determine that the coarse pattern inspection training is qualified, allowing the user to proceed to the next operation step (such as fine adjustment or direct production). If any round fails, the system will determine that the device is unqualified, lock the equipment, and issue an audible and visual alarm, prompting the operator to check the light source hardware or adjust the recognition parameters.

[0036] Specifically, the brightening and darkening light source conditions are obtained by adjusting the light source intensity of the normal light source conditions by a preset percentage.

[0037] It should be further explained that the present invention specifies the specific method of light source adjustment, which is an efficient and easy-to-implement preferred solution. In specific implementation, the "preset percentage" is a key parameter. The system stores the light source intensity values ​​of the axial light and the ring light under normal light source conditions (e.g., expressed as digital values ​​or current values). When generating a brightening condition, the system multiplies this value by a coefficient of (1 + preset percentage); when generating a darkening condition, it multiplies it by a coefficient of (1 - preset percentage). As a preferred embodiment, the preset percentage can be selected as 15%. The reason for selecting 15% is that this value is determined based on statistical analysis of typical decay curves of light source devices (such as LEDs) and field fault data. It can effectively cover the significant decay range that occurs in the middle and late stages of the light source's lifespan (e.g., a brightness decay of 10%-20% is a common range that leads to identification errors), while avoiding misjudgment due to excessive adjustment exceeding the normal process window, thus achieving a good balance between detection sensitivity and practicality.

[0038] Furthermore, let the intensity of the axial light be under normal light source conditions. The intensity of the ring light is The intensity under brighter light source conditions is calculated as follows: The intensity calculation under dimmed light source conditions is as follows: in: and These represent the axial light intensity and the ring light intensity under the brightening condition, respectively; and These represent the axial light intensity and the ring light intensity under dimming conditions, respectively; This represents the preset percentage of light source fluctuation (preferably 0.15).

[0039] Understandably, the advantage of this percentage-based linear adjustment method is that it makes the simulated changes in light source conditions quantifiable and repeatable, which is very convenient for engineers to unify standards, perform parameter porting and consistency management between devices, and also simplifies the implementation logic of software algorithms.

[0040] Specifically, the preset percentage adjustment applies to both axial light and ring light.

[0041] It should be further clarified that this invention further defines the object of adjustment, namely, the simultaneous and proportional brightness adjustment of the axial beam and the ring beam. In practical visual illumination systems, the axial beam (typically providing perpendicular incident illumination) and the ring beam (typically providing oblique incident illumination) work together to highlight the three-dimensional morphology and contrast of the wafer surface. In specific implementation, the control system sends commands to the light source driver to synchronously increase or decrease the axial beam drive current and the ring beam drive current according to the aforementioned preset percentages.

[0042] Furthermore, in practical implementation, when the system executes the second and third rounds, the light source control module will simultaneously send adjustment commands to both the axial optical driver and the ring optical driver to ensure that their brightness is at the same percentage. Make changes. For example, when When the normal light source is 2000 for the axial beam and 1500 for the ring beam, the brightening condition is 2300 for the axial beam and 1725 for the ring beam; the darkening condition is 1700 for the axial beam and 1275 for the ring beam. This synchronous change is completed through the same software instruction loop of the device, with almost no hardware delay.

[0043] Understandably, the above scheme simulates the most stringent and most likely real-world scenario of light source degradation, namely, a decline in the overall performance of the entire lighting module, rather than a change in a single light source. This ensures that the detection can comprehensively cover the changes in feature point imaging under various lighting angles, avoiding potential risks that might be missed by adjusting only a single light source. This makes the "stress test" more comprehensive and reliable, further enhancing the robustness of the detection scheme.

[0044] Specifically, the pattern comparison includes rotational comparison and spiral inspection.

[0045] It should be further explained that this invention clarifies the specific algorithms or steps involved in pattern matching, namely "rotational matching" and "spiral checking." These are two continuous and complementary precision alignment processes. In specific implementation, rotational matching refers to controlling the wafer stage (Chuck) to move the wafer within a certain angle range (e.g., rotating a certain angle clockwise and counterclockwise around the identified initial position) after roughly locating the feature points. By calculating the point with the highest matching degree of the feature point image at different angles, the mechanical angle deviation of the wafer is precisely corrected. Spiral checking, after angle correction, involves the vision system performing a fine search within a preset planar area centered on the current identification point, following a spiral or grid-like path, to find the most accurate sub-pixel coordinates of the feature points. These two are combined into each light source detection round of this invention.

[0046] Furthermore, this invention clarifies two key sub-steps involved in pattern comparison. The specific implementation process is as follows: 1. Rotation Match: After the system initially locates the feature points through template matching, this step is performed first to correct any possible angular mounting deviations of the wafer. The system controls the motor carrying the wafer (θ-axis motor) to rotate the wafer within a small angular range around the current identification point for forward and reverse scanning. 2. Spiral Search: This step is performed immediately after rotational comparison is completed and angular deviations are corrected to achieve sub-pixel-level precise localization of feature points. The vision system instructs the stage to make micro-movements in the X and Y directions, and the search path is a spiral line expanding outward from the center.

[0047] Understandably, this invention provides a dual guarantee of alignment accuracy. It not only verifies whether the system can "find" feature points under different lighting conditions, but also verifies whether it can "accurately locate" them, i.e., whether the final positioning coordinates are precise and stable. This effectively avoids alignment coordinate shifts caused by slight changes in the feature point morphology due to variations in lighting, ensuring the absolute accuracy of subsequent cutting path definitions.

[0048] Specifically, the rotation comparison is a coordinate error comparison performed within a preset angle range of the material sheet rotation.

[0049] It should be further explained that the rotation comparison in this invention has been refined. In specific implementation, the "preset angle range" is an important error-tolerance parameter. As a preferred embodiment, this range is set to ±15 degrees. The reason for choosing ±15 degrees is based on the engineering experience value of the maximum mechanical angle deviation that may exist during wafer loading. This range is sufficient to cover all normal mounting deviations, ensuring that feature points are not lost due to excessive initial angle deviation. During the comparison process, the system records the coordinate values ​​(X, Y) of the center of the feature point before and after rotation, and calculates its Euclidean distance as the coordinate error.

[0050] Furthermore, during rotational alignment, the system acquires images and performs template matching at three positions: -15°, 0°, and +15° (or in smaller increments). Upon successful matching, the system records the pixel coordinates of the feature points at each angle. Coordinate error typically refers to the offset between the coordinates of a feature point calculated from the 0° position and the ±15° position. This offset should be within an allowable tolerance range.

[0051] It is understandable that by setting a reasonable and sufficient rotation search range and supplementing it with the quantitative judgment of coordinate error, the present invention provides an objective and measurable qualification standard for rotation comparison, so that this step is no longer a vague judgment that relies on operator experience, but becomes an automated and standardized detection link.

[0052] Specifically, the spiral check involves comparing coordinate errors along a preset spiral path.

[0053] It should be further explained that this invention refines the spiral inspection process. In specific implementation, the "preset spiral path" defines the step size and range of the fine-grained search. Starting from the current coordinates, the system acquires images and performs matching calculations along a spiral path with gradually increasing step size from the inside out to locate the optimal matching point. The coordinate error of the spiral inspection refers to the deviation between the coordinates of the optimal point found by the spiral search and the coordinates of the center point determined after rotation and comparison.

[0054] Furthermore, the spiral inspection algorithm works as follows: starting from the center point, following a spiral path, the similarity between the small region image and the feature template is calculated point by point in increments of 1 pixel. A commonly used similarity metric is the Sum of Absolute Differences (SAD). Its formula is as follows: in: Indicates the location in the real-time image. The matching difference value at each location; the smaller the value, the more similar the match. Represents a grayscale image captured in real time; Represents a pre-stored feature template image; and These represent the width and height (in pixels) of the template image, respectively. and It is the pixel index within the template image; Furthermore, the system will find the point with the smallest SAD value, which is the optimal matching point. The coordinate error of the spiral check refers to the Euclidean distance between the coordinates of this optimal matching point and the coordinates of the center point after rotation and comparison.

[0055] Understandably, this search strategy balances search efficiency and positioning accuracy, enabling it to quickly and accurately converge to the true center of the feature points, further eliminating alignment errors.

[0056] Specifically, determining whether the module alignment is qualified includes: when the pattern comparison in all light source detection rounds meets the preset error threshold, the coarse adjustment pattern inspection is deemed qualified.

[0057] It should be further explained that this invention clearly defines the final pass / fail judgment logic, which is the core of the quality control of this invention. In specific implementation, this is an "AND" logical relationship: that is, the tests for normal light source, brightened light source, and darkened light source must all pass; none can be omitted. If the error of rotation comparison or spiral inspection in any round exceeds the threshold, the overall judgment will be deemed unqualified.

[0058] Furthermore, in practice, the decision module internally performs a strict "AND" logical judgment. Its pseudocode is as follows: “if((Round_1_Result==PASS) and (Round_2_Result==PASS) and (Round_3_Result==PASS)) then Teaching status = qualified Unlock the production process Else Teaching status = unsatisfactory Trigger alarm and lock device end if".

[0059] Furthermore, the PASS result for each round means that the coordinate errors of the rotation comparison and spiral check in that round are both less than their respective preset thresholds.

[0060] Understandably, this invention constructs a highly reliable "safety gate," where only identification parameters that can withstand the test of simulated environmental fluctuations are adopted for production, thus minimizing wishful thinking and potential risks and ensuring the consistency of product quality.

[0061] Specifically, the preset error thresholds include the coordinate error threshold for rotation comparison and the coordinate error threshold for spiral inspection.

[0062] It should be further explained that the present invention specifies the error thresholds. In specific implementation, thresholds need to be set separately for the two steps with different precision requirements: rotational alignment and spiral inspection. As a preferred embodiment, the coordinate error threshold for rotational alignment can be set to 15 pixels, and the coordinate error threshold for spiral inspection can be set to 5 pixels. The reason for choosing 15 pixels as the rotational alignment threshold is that this value fully considers the backlash of the mechanical rotation mechanism itself and the deformation of the image under large-angle rotation, and is a lenient and reasonable tolerance value. The main purpose is to confirm whether the approximate angle correction is correct. The reason for choosing 5 pixels as the spiral inspection threshold is that spiral inspection is the final fine positioning, and its error will be directly transmitted to the cutting head, so higher precision is required; 5 pixels is a safety value derived from the cutting track width and the positioning accuracy of the cutting head, ensuring that the alignment error will not cause the cutting head to damage the chip's functional area.

[0063] Understandably, the beneficial effect of this hierarchical threshold management is that it ensures both the pass rate of detection (avoiding false alarms caused by overly stringent rotation comparison) and the accuracy of final alignment, achieving an optimal balance between efficiency and accuracy.

[0064] Specifically, the number of light source detection rounds is three, corresponding to the normal light source condition, the brightened light source condition, and the darkened light source condition, respectively.

[0065] It should be further noted that this invention limits the optimal number of detection rounds to three. In specific implementation, the choice of three rounds is based on a comprehensive consideration of effectiveness and efficiency. The three rounds represent the "baseline state" (normal light source), the "performance enhancement boundary" (brighter light source), and the "performance degradation boundary" (darkened light source). These three points are sufficient to delineate the two extremes and the center of the light source's operating range, thereby effectively assessing the stability of the recognition system within the expected fluctuation range. If the number of rounds is too small (e.g., two), it will not be able to fully cover the fluctuation scenarios; if it is too large (e.g., four or more), it will significantly increase the time of the teaching phase, reduce the equipment uptime, and the marginal effect of the resulting reliability improvement will diminish. Therefore, three rounds is the optimal and most economical number of rounds to achieve the purpose of this invention.

[0066] Understandably, choosing three rounds is based on the engineering principle of "boundary-benchmark" testing, which is the optimal and most economical number of tests to achieve the purpose of this invention. Three points (benchmark point, upper boundary point, and lower boundary point) are sufficient to define the critical range of a working curve. Fewer than three points (such as two) cannot fully assess stability; more than three points (such as four or five) will significantly increase training time, leading to a decrease in equipment uptime, while the marginal effect of the resulting reliability improvement diminishes.

[0067] It is understandable that this invention achieves a full verification of the robustness of the recognition system without significantly increasing time costs.

[0068] Please see Figure 2 The present invention provides another embodiment, which provides a system for aligning and detecting a cutting machine module. The system for aligning and detecting a cutting machine module includes: The light source control module 100 is used to set up multiple light source detection rounds during the teaching phase of module alignment detection. The multiple light source detection rounds include normal light source conditions, brightened light source conditions, and darkened light source conditions.

[0069] It should be further explained that the light source control module 100 runs in software within the IPC and communicates with the PLC or directly with the light source controller (such as an LED controller). It incorporates the aforementioned percentage adjustment algorithm. When a brightening or dimming cycle needs to be executed, this module sends a new current setting command to the light source driver via a communication bus (such as EtherCAT or Ethernet / IP) to precisely control the brightness of the axial light and the ring light.

[0070] The comparison module 200 is used to perform pattern comparison in each round of light source detection.

[0071] It should be further explained that the core of the comparison module 200 is an image processing software library (such as Halcon, OpenCV, etc.) running on an IPC or dedicated vision processing card. It receives digital images from the CCD camera transmitted via the image acquisition card, and calls its built-in template matching, rotation scanning, and spiral search algorithms to perform operations such as SAD calculation, and finally outputs the precise coordinates of the feature points and the comparison error value.

[0072] The determination module 300 is used to determine whether the module alignment is qualified based on the result of the pattern comparison.

[0073] It should be further noted that the judgment module 300 is also a software logic module, integrated into the main control program. It continuously receives error data (rotational comparison error and spiral inspection error) from the comparison module and compares it with preset thresholds (such as 15 pixels and 5 pixels as described above) stored in the Recipe. Based on the control logic described in this invention, a final pass / fail judgment is made, and a control signal (such as activating the alarm or enabling the cutting start button) is output.

[0074] It should be further noted that this invention claims protection for a system that implements the above-described method. The various modules of this system can be upgraded from the existing programmable logic controller (PLC) and industrial computer (IPC) of the cutting machine, along with corresponding vision processing software. Specifically, the light source control module 100 can be a software functional block within the IPC. It sends instructions to the drivers of the axial and ring lights via a digital I / O card or a dedicated light source controller interface, precisely controlling their output current to generate the three light source conditions. The comparison module 200 can be integrated into the vision processing card or run using the CPU resources of the IPC. It receives image data from a CCD camera and executes image processing algorithms such as rotation comparison and spiral inspection. The judgment module 300 is also a software logic unit. It receives the coordinate error data output by the comparison module and compares it with a preset error threshold, making a final pass / fail judgment based on the logic of this invention.

[0075] It is understood that the present invention materializes the innovative method and process into hardware and software modules, enabling the technical solution of the present invention to be directly embedded into existing semiconductor cutting equipment. The modular design facilitates functional upgrades and maintenance, providing a solid hardware foundation and execution guarantee for achieving highly reliable module alignment detection.

[0076] In a preferred embodiment, this application also provides an electronic device, the electronic device comprising: The computer device includes a memory and a processor, wherein the memory stores computer-readable instructions that, when executed by the processor, implement the method for aligning and detecting the cutting machine module. The computer device can be broadly categorized as a server, terminal, or any other electronic device with the necessary computing and / or processing capabilities. In one embodiment, the computer device may include a processor, memory, network interface, communication interface, etc., connected via a system bus. The processor of the computer device can be used to provide the necessary computing, processing, and / or control capabilities. The memory of the computer device may include a non-volatile storage medium and internal memory. The non-volatile storage medium may store an operating system, computer programs, etc. The internal memory can provide an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface and communication interface of the computer device can be used to connect and communicate with external devices via a network. When the computer program is executed by the processor, it performs the steps of the method of the present invention.

[0077] This invention can be implemented as a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, causes the steps of the methods of embodiments of the invention to be performed. In one embodiment, the computer program is distributed across multiple network-coupled computer devices or processors, such that the computer program is stored, accessed, and executed in a distributed manner by one or more computer devices or processors. A single method step / operation, or two or more method steps / operations, may be executed by a single computer device or processor or by two or more computer devices or processors. One or more method steps / operations may be executed by one or more computer devices or processors, and one or more other method steps / operations may be executed by one or more other computer devices or processors. One or more computer devices or processors may execute a single method step / operation, or execute two or more method steps / operations.

[0078] Those skilled in the art will understand that the method steps of this invention can be performed by a computer program instructing related hardware, such as a computer device or processor, to perform the steps of this invention when executed. Depending on the context, any references herein to memory, storage, databases, or other media may include non-volatile and / or volatile memory. Examples of non-volatile memory include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid-state drive, etc. Examples of volatile memory include random access memory (RAM), external cache memory, etc.

[0079] The technical features described above can be combined arbitrarily. Although not all possible combinations of these technical features are described, any combination of these technical features should be considered to be covered by this specification, provided that such combination does not contain contradictions.

[0080] The specific embodiments of the present invention described above do not constitute a limitation on the scope of protection of the present invention. Any other corresponding changes and modifications made in accordance with the technical concept of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for aligning and detecting a cutting machine module, characterized in that, Includes the following steps: S100, The teaching stage for module alignment detection, the teaching stage includes coarse pattern inspection; S200. In the coarse pattern inspection, multiple light source detection rounds are set, including normal light source conditions, brightened light source conditions, and darkened light source conditions. S300. In each round of light source detection, pattern comparison is performed; S400. Based on the results of the pattern comparison, determine whether the module alignment is qualified.

2. The method according to claim 1, characterized in that, The brightening and darkening light source conditions are obtained by adjusting the light source intensity of the normal light source conditions by a preset percentage.

3. The method according to claim 2, characterized in that, The preset percentage adjustment applies to both axial light and ring light.

4. The method according to claim 3, characterized in that, The pattern comparison includes rotational comparison and spiral inspection.

5. The method according to claim 4, characterized in that, The rotation comparison is a coordinate error comparison performed within a preset angle range of the material sheet rotation.

6. The method according to claim 4, characterized in that, The spiral check involves comparing coordinate errors along a preset spiral path.

7. The method according to claim 6, characterized in that, Determining whether the module alignment is qualified includes: When the pattern comparison in all light source detection rounds meets the preset error threshold, the coarse adjustment pattern inspection is deemed qualified.

8. The method according to claim 7, characterized in that, The preset error thresholds include the coordinate error threshold for rotation comparison and the coordinate error threshold for spiral inspection.

9. The method according to claim 1, characterized in that, The number of light source detection rounds is three, corresponding to the normal light source condition, the brightened light source condition, and the darkened light source condition, respectively.

10. A system for aligning and detecting a cutting machine module, characterized in that, include: The light source control module is used to set up multiple light source detection rounds during the teaching phase of module alignment detection. The multiple light source detection rounds include normal light source conditions, brightened light source conditions, and darkened light source conditions. The comparison module is used to perform pattern comparison in each round of light source detection; The determination module is used to determine whether the module alignment is qualified based on the pattern comparison result.

Images