Automatic optical detection equipment control operation system based on human-computer interaction interface

Abstract

The invention relates to the technical field of industrial automatic detection, and discloses an automatic optical detection equipment control operation system based on a human-computer interaction interface, which comprises a detection item intelligent configuration module used for realizing rapid switching and parameter setting of various core detection functions of automatic optical detection equipment through a visual interaction component; the parameter self-adaptive module is used for realizing automatic extraction of detection features and self-adaptive setting of threshold parameters by adopting a deep learning algorithm, the deep learning algorithm has self-adaptive adjustment capability on environment illumination change and detection object material characteristic change, and the detection precision is optimized by dynamically adjusting a detection threshold and a compensation parameter; and the program multiplexing module is used for supporting the import and export of cross-equipment detection parameters and the rapid switching of detection modes so as to adapt to the detection requirements of detection objects with different specifications, greatly reduce the product transformation debugging time and the misjudgment rate, and improve the product yield.

Description

Technical Field

[0001] This invention relates to the field of industrial automation testing technology, and more specifically to a human-machine interface control operating system for automatic optical testing equipment. Background Technology

[0002] In the production process of new energy vehicle power battery modules (CCS), the accuracy and efficiency of automated optical inspection (AOI) equipment in detecting key product features (such as laser engraving codes, connector pins, and laser solder joints) directly affect the production yield. However, existing AOI equipment software systems have the following technical defects:

[0003] 1. Complex configuration of detection parameters

[0004] Traditional systems employ a multi-level nested menu design, requiring debugging personnel to manually configure over 48 parameters (such as laser engraving orientation, connector type, laser solder joint parameters, etc.) across 8 levels of sub-menus, with a single debugging session taking up to 3 hours. Strong coupling exists between parameters (e.g., PIN pin parameters need to be set even when the connector orientation is not enabled), leading to an exponential increase in configuration complexity.

[0005] 2. Poor program reusability

[0006] Testing procedures cannot be directly migrated between different production lines. Even when switching production lines for the same specifications, the photo points still need to be recalibrated. Data from a leading battery manufacturer shows that the equipment idle rate due to program refactoring during production line switchovers reaches 22%. This is because the existing system lacks a standardized data encapsulation protocol, making it impossible to achieve "plug-and-play" testing procedures.

[0007] 3. Low efficiency of human-computer interaction

[0008] Traditional interfaces use a flowchart-based menu structure, requiring users to navigate through more than five layers of the interface to perform key functions (such as switching deep learning models and enabling batch detection items). Third-party evaluations indicate that technicians need to perform over 120 clicks to complete standard debugging tasks, with an error rate as high as 18%.

[0009] 4. Lack of intelligent fault tolerance mechanism

[0010] The existing system lacks fault tolerance mechanisms. For example, it lacks parameter logic verification during the debugging phase. Common errors include: setting pin parameters without enabling connector orientation detection (leading to detection logic conflicts), and failing to synchronously adjust the density of photo capture points in dual-product mode (causing a decrease in detection coverage). Statistics from an OEM manufacturer show that such errors account for 35% of debugging rework. Summary of the Invention

[0011] In view of this, in order to address the shortcomings of traditional automatic optical inspection equipment in the process of changeover and debugging, this invention provides an automatic optical inspection equipment control operating system with a human-machine interface and an automatic optical inspection equipment changeover and debugging method applied to the system.

[0012] In a first aspect, the present invention provides an automatic optical inspection equipment control operating system with a human-computer interaction interface, comprising:

[0013] The core functional layer, hardware control layer, and user interface layer are, among which...

[0014] The core functional layer integrates intelligent configuration, algorithm support, program reuse, and multi-layered fault tolerance mechanisms. It handles detection business logic, ensuring parameter adaptation, function switching, and process stability, driving intelligent detection and efficient model changeover. This includes:

[0015] The intelligent configuration module for detection items is used to enable rapid switching and parameter setting of multiple core detection functions of automatic optical inspection equipment through a visual interactive component.

[0016] The parameter adaptive module is used to automatically extract detection features and adaptively set threshold parameters using a deep learning algorithm. The deep learning algorithm has the ability to adaptively adjust to changes in ambient lighting and changes in the material properties of the detected object, and optimizes detection accuracy by dynamically adjusting the detection threshold and compensation parameters.

[0017] The program reuse module supports the import and export of detection parameters across devices and the rapid switching of detection modes to adapt to the detection needs of different specifications of detection objects.

[0018] The hardware control layer is used to control the CCD camera image acquisition and device movement, and to convert control commands into physical detection actions.

[0019] The user interface layer serves as a human-computer interaction window, presenting status, providing intelligent guidance for operations, and monitoring and debugging processes through a visual interface.

[0020] The automatic optical inspection equipment control operating system based on a human-computer interaction interface provided in this invention has a core functional layer that integrates intelligent configuration, algorithm support, program reuse, and multi-layer fault tolerance mechanisms. Through the intelligent configuration module for inspection items, it enables rapid switching of inspection functions and parameter setting. Relying on the parameter adaptive module, it automatically extracts features using deep learning algorithms, adaptively sets thresholds, and optimizes accuracy to adapt to changes in environment and object. The program reuse module supports cross-device parameter import and mode switching to adapt to diverse needs, ensuring stable inspection business logic processing, parameter adaptation, function switching, and workflow, driving intelligent inspection and efficient model changeover. The hardware control layer precisely controls CCD camera acquisition and equipment movement, converting commands into physical inspection actions. The user interface layer serves as the human-computer interaction window, visually presenting status, intelligently guiding operation, and monitoring the debugging process. These three layers work together to improve inspection accuracy, efficiency, and equipment adaptability while lowering the operational threshold, making the inspection process intelligent, stable, and easy to use, thus facilitating the efficient operation of automatic optical inspection equipment.

[0021] In one optional implementation, the intelligent configuration module for detection items adopts a modular management architecture, dividing each detection function into independent module units. Each module unit can be independently enabled or disabled through a user interface and supports automatic interface refresh.

[0022] Because different product processes require different detection function modules, this invention allows for quick and intuitive activation or deactivation of these modules via a user interface. The interface automatically refreshes to display the activation / deactivation status, significantly reducing equipment debugging time due to model changes and improving production line switchover efficiency. When a function module malfunctions, it can be quickly isolated (disabled) through the interface, ensuring the normal operation of core function modules and preventing excessive impact from faulty modules on testing work, thus improving the overall stability and continuous operation capability of the system. The visually displayed activation / deactivation buttons reduce the cognitive load on operators, enabling them to more clearly understand the status of each module, reducing errors caused by complex operations or unfamiliarity with module functions, and improving human-machine interaction efficiency and operational accuracy.

[0023] In one optional implementation, the parameter adaptation module includes:

[0024] The illumination change adaptive unit is used to automatically adjust the deep learning contrast detection threshold according to the illumination change for low-light and overexposed scenes through a two-level compensation mechanism, and uses logarithmic compensation detection parameters.

[0025] The material batch adaptive unit is used to automatically adjust the deep learning detection threshold based on changes in material reflectivity and surface texture, and employs logarithmic compensation for detection parameters.

[0026] This invention employs a two-stage compensation mechanism to adaptively adjust for low-light and overexposure scenarios. It automatically adjusts the deep learning contrast detection threshold based on changes in illumination and uses logarithmic compensation for detection parameters. This effectively solves the problem of poor adaptability of detection parameters under different lighting conditions, reduces misjudgments caused by illumination fluctuations, and improves detection stability and accuracy in complex lighting environments, thereby increasing product yield. By identifying changes in material reflectivity and surface texture, it automatically adjusts the deep learning detection threshold and uses logarithmic compensation for detection parameters. This adapts to differences in material characteristics between different batches of products, avoiding the tedious process of manually readjusting parameters due to batch material variations. It reduces reliance on operator experience, minimizes parameter configuration time and operational errors, ensures consistency in detection standards across different batches of products, and further improves detection efficiency and reliability.

[0027] In one optional implementation, the deep learning algorithm of the parameter adaptive module integrates a model calling mechanism, supports one-click import and export of models, and adopts a multi-model collaborative detection mode for complex detection items.

[0028] The parameter adaptive module of this invention integrates a deep learning algorithm with a model calling mechanism. By supporting one-click model import and export, it eliminates the complex process of traditional manual configuration and retraining, improving the model reuse efficiency across different production lines or equipment and saving debugging time. For complex detection items such as residual glue morphology and PIN pin misalignment, a multi-model collaborative detection mode is adopted. The detection confidence is improved through multi-model cross-validation, reducing the recognition blind spots and false judgments of a single model, improving the detection accuracy in complex scenarios, and reducing the rate of missed detection of defective products. At the same time, the convenience of model calling and the automated processing of multi-model collaboration reduce the dependence on the operator's deep learning expertise. High-precision detection can be achieved without manually adjusting the model fusion parameters, shortening the personnel training cycle. Combined with the illumination and material adaptive unit, it can flexibly adapt to the detection needs of products with different materials and processes, enhancing the system's adaptability to diverse production scenarios.

[0029] In one optional implementation, the program reuse module supports the import and export of a preset format detection point configuration file, which contains detection coordinate information and associated detection parameters; the quick switching of detection modes includes one-click switching between single-object detection mode and multi-object detection mode, and the motion control logic and image processing resource allocation are adjusted synchronously during the switching.

[0030] The program reuse module of this invention supports the import and export of preset format detection point configuration files (including detection coordinate information and associated detection parameters), realizing the direct migration of cross-device detection programs without recalibrating the photo points (i.e., cross-device point migration). This solves the problem of traditional programs not being reusable between different production line equipment, reducing the equipment switching idle rate from 22% to 5%. It also supports one-click switching between single-object and multi-object detection modes, automatically adjusting motion control logic and image processing resource allocation during switching, replacing the tedious operation of manually modifying more than ten parameters. This reduces the average number of single debugging clicks from 120+ to less than 40, shortens product changeover debugging time from 3 hours to 1 hour, and improves the standardization rate of detection programs to 100%, ensuring the consistency of detection standards across different equipment and production lines.

[0031] In one optional implementation, the system adopts a hierarchical interface architecture. The main interface displays at least the real-time monitoring window of the detection device, device parameter settings, and commonly used function switching components. The sub-interfaces are used to display the enabled / disabled status of the function modules, and the secondary sub-interfaces integrate advanced parameter configuration functions.

[0032] This invention provides a main interface that centrally displays the real-time monitoring window of the testing device, device parameter settings, and commonly used function switching components. Sub-interfaces display the enabled / disabled status of function modules, and secondary sub-interfaces integrate advanced parameter configuration functions. This architecture reduces the frequency of operators switching between multi-level menus, lowers operational complexity and the traditional error rate, while simplifying basic operations and making advanced configurations controllable. It adapts to the needs of all scenarios, from rapid changeover on the production line to complex parameter debugging, shortens the training cycle for technical personnel, and improves human-computer interaction efficiency and system usability.

[0033] In a second aspect, the present invention provides an automatic optical inspection equipment changeover and debugging method based on the system described in the first aspect, the method comprising:

[0034] The system is reset, the testing software is started and the operating system interface is entered, and the testing equipment is controlled to perform a reset operation.

[0035] When switching product modes, the corresponding detection mode is selected based on the mode switching component in the interface, and the system automatically executes the mode switching process.

[0036] Laser code recognition debugging: Switch to the corresponding identifier recognition parameter configuration mode through the interface components;

[0037] Connector testing and debugging: On the connector interface, select the enabled connecting parts and associated testing items based on the testing object configuration. After completing the basic debugging, trigger the automatic threshold setting function. The system automatically calculates the upper and lower limits of the testing parameters. The intelligent testing function can be enabled to realize the detection of residual glue on the connecting parts.

[0038] For equipment photo point debugging, import the pre-saved detection point configuration file. If it is the first detection, manually teach the detection points and save the configuration file for future reuse across devices.

[0039] The testing program is debugged by sequentially calling other testing function modules to complete the debugging of the corresponding testing items.

[0040] The method provided by this invention uses an interface mode switching component to quickly switch product modes. In processes such as laser code recognition and connector detection, the interface component and automatic setting function simplify parameter configuration. Combined with the import of pre-stored point files or the saving of the first teaching, cross-device reuse is achieved. Then, debugging is completed by calling the function modules in sequence. This significantly simplifies the model changeover debugging process, shortens the debugging time, reduces the complexity of operation, and improves the efficiency and flexibility of the equipment in adapting to different specifications of testing objects. At the same time, it ensures the reusability and consistency of the testing parameters.

[0041] In one optional implementation, enabling the intelligent detection function to detect residual adhesive in the connecting parts includes: selecting the deep learning function in the connector interface, importing the trained model data with one click, and automatically detecting residual adhesive in the connecting parts.

[0042] This invention enables rapid and accurate automatic detection of residual adhesive in connector components by selecting the deep learning function on the connector interface and importing pre-trained model data with one click. This eliminates the need for complex manual settings of residual adhesive detection parameters, simplifying the operation process and improving the efficiency and accuracy of residual adhesive detection. The use of mature models also enhances the intelligence and professionalism of the detection, helping equipment to complete changeover debugging and stable testing more efficiently.

[0043] In one optional implementation, during the configuration of detection items, a fault tolerance range corresponding to each detection parameter is preset. When the parameter configuration exceeds the preset range, a prompt is automatically triggered and the parameters exceeding the range are restricted from taking effect. The logical correlation between parameters is verified in real time. When a corresponding detection item is not enabled or a logical conflict in parameter configuration is detected, a prompt is issued. Parameter configuration permissions are restricted for detection items that have not completed the pre-debugging steps to prevent detection items from not being enabled or detection parameters from being configured incorrectly due to skipping steps.

[0044] In the configuration process of the detection items, this invention presets the fault tolerance range of each detection parameter and automatically prompts and restricts the parameters when they exceed the limits. It also verifies the logical correlation between parameters in real time and issues prompts when the corresponding detection item is not enabled or when there is a conflict in the parameter configuration. At the same time, it restricts the parameter configuration permissions of detection items that have not completed the pre-debugging steps. This can effectively prevent configuration errors caused by parameter exceeding limits, logical conflicts, or skipping steps, reduce the risk of detection items not being enabled or improper parameter configuration, improve the accuracy and standardization of detection item configuration, reduce debugging rework and detection deviations caused by configuration problems, and ensure the stability and reliability of the detection process through an intelligent fault tolerance mechanism.

[0045] In one alternative implementation, during the model changeover and debugging process, the operating status of each functional module is monitored in real time. If any functional module malfunctions, the faulty module is automatically disabled within seconds, while the core detection process continues to run.

[0046] This invention provides a solution for quickly isolating faults, preventing the spread of abnormalities from slowing down the overall progress, ensuring that key debugging steps do not stop, reducing the impact of faults on debugging efficiency, maintaining the operation of core testing functions, making the replacement debugging smoother and more stable, reducing the risk of debugging interruption due to module failure, and improving the reliability and continuity of the debugging process. Attached Figure Description

[0047] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0048] Figure 1 This is a schematic diagram of the technical architecture of the automatic optical inspection equipment control operating system based on a human-computer interaction interface according to an embodiment of the present invention;

[0049] Figure 2 This is a flowchart illustrating the changeover and debugging method for automated optical inspection equipment;

[0050] Figure 3 This is a schematic diagram of a certain model of CCS product according to an embodiment of the present invention;

[0051] Figure 4 This is a schematic diagram of the main interface according to an embodiment of the present invention;

[0052] Figure 5 This is a schematic diagram of the connector detection and debugging interface according to an embodiment of the present invention;

[0053] Figure 6This is a schematic diagram of the PIN pin detection interface according to an embodiment of the present invention;

[0054] Figure 7 This is a schematic diagram of the operation and control interface according to an embodiment of the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, 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.

[0056] This invention provides an automated optical inspection equipment control operating system based on a human-computer interaction interface. It is an intelligent software operating system suitable for semi-automatic aerial photography (AOI) equipment on new energy vehicle power battery module (CCS) production lines, covering detection parameter configuration, deep learning algorithm integration, and cross-device program reuse technology. It achieves the following objectives:

[0057] 1. Decoupling detection parameter configuration

[0058] To address the configuration complexity issue caused by the coupling and correlation of multi-dimensional detection items, a modular detection function is established, enabling operators to directly control the activation / deactivation of physical detection items through a visual interface, integrating the correlation parameters that originally required cross-level settings into a single-step operation.

[0059] 2. Refactor the program's interaction logic

[0060] The design incorporates a transfer learning-based program reuse function, unifying the equipment coordinate system, the position of the limit origin, and other parameters. Optical parameters are automatically set, and the parameters and configuration data files are saved together and automatically set in the imported program. This enables the detection program to be used "plug and play" across production line equipment, eliminating production capacity losses caused by program refactoring.

[0061] like Figure 1The diagram shows the system's technical architecture, divided into a core functional layer, a hardware control layer, and a user interface layer. The core functional layer is responsible for business logic and intelligent control, integrating intelligent configuration, algorithm support, program reuse, and multi-layered fault tolerance mechanisms. It handles detection business logic, ensuring parameter adaptation, function switching, and process stability. The hardware control layer interfaces with hardware execution, controlling image acquisition (the source of detection data) via a CCD camera and managing device movement, such as platform movement and camera pose adjustment, to translate software control commands into physical detection actions. The user interface layer presents the status, provides intelligent operation guidance, and monitors the debugging process through a visual interface. A "3D visualization interface" intuitively presents the detection status, intelligently guides user operation (lowering the barrier to entry), and provides real-time feedback on system operation during debugging, allowing users to easily control and monitor the detection process. This complete workflow of "user operation → function processing → hardware execution" enables the equipment to possess intelligent detection capabilities (deep learning, parameter self-adaptation), rapid model changeover through program reuse, and ensures ease of use and system stability, meeting the needs of efficient debugging and accurate detection in complex scenarios for automated optical inspection equipment.

[0062] Furthermore, the three main functional modules of the core functional layer are as follows:

[0063] The intelligent configuration module for inspection items enables rapid switching and parameter setting of various core inspection functions of automated optical inspection equipment through a visual interactive component. Specifically, it allows for rapid switching of 10 core inspection functions, such as laser engraving code orientation recognition (positive / negative / left / right), connector type identification (male pin row count / residual glue / skew detection, one-click setting of pin position deviation limit, female buckle integrity / hole blockage detection), laser solder joint parameter configuration (quantity / type), hot riveting process inspection (missing riveting identification), fuse inspection, female pin code inspection, and auxiliary material inspection. This reduces the cognitive load and operational difficulty of inspection item configuration, reduces the number of debugging clicks (from 120+ to less than 40), shortens product changeover and debugging time, and improves the adaptability and accuracy of inspection functions.

[0064] This invention, through visual interactive components (such as one-click switching buttons and checkboxes), eliminates the need for manual parameter modification in multi-level process flowcharts, reducing product changeover and debugging time by 66.6% (from an average of 3 hours to 1 hour), and significantly minimizing equipment downtime. The complex process, which previously required specialized technicians to configure 48 parameters across 8 sub-menus, is simplified to intuitive operation via a visual interface. The training period for technicians is reduced from 2 weeks to 3 days, decreasing reliance on specialized operators. It supports flexible configuration of multiple types of testing items (such as testing three rows of pins inside connectors, residual adhesive testing, etc.), overcoming the limitations of traditional standard procedures that only support a limited number of testing items, and meeting the testing needs of different product processes.

[0065] The parameter adaptive module is used to automatically extract detection features and adaptively set threshold parameters using deep learning algorithms. The deep learning algorithm has the ability to adaptively adjust to changes in ambient lighting and changes in the material properties of the detected object, and optimizes detection accuracy by dynamically adjusting the detection threshold and compensation parameters.

[0066] This module uses deep learning algorithms to adaptively adjust to changes in ambient lighting (low light, overexposure scenes) and the material characteristics of the object being inspected (reflectivity, surface texture). It dynamically adjusts the detection threshold and employs logarithmic compensation parameters to effectively reduce the false positive rate and improve product yield. It enables automatic extraction of residual adhesive features and one-click setting of PIN pin deviation thresholds, eliminating the need for manual calculation of standard limits and manual adjustment of over twenty inspection tool parameters, reducing human error and improving debugging reliability. For complex inspection items, a multi-model collaborative detection mode is adopted, combined with a one-click model import / export mechanism, enhancing the adaptability to products with different materials and under different lighting conditions, and expanding the applicability of the equipment.

[0067] The program reuse module supports the import and export of detection parameters across devices and the rapid switching of detection modes to adapt to the detection needs of different specifications of objects. Through the operation and control interface, this module supports the import and export of user-preset format (such as CSV) detection point configuration files, containing coordinate information and detection parameters, enabling direct migration of detection programs between different production lines. This reduces the equipment idle rate during switching from 22% to 5%, minimizing production capacity losses due to program reconstruction. It supports one-click switching between single and dual product modes, automatically adjusting motion control logic and image processing resource allocation during switching, eliminating the need for manual modification of more than ten parameters, and reducing the average number of clicks per debugging session from over 120 to less than 40. Through mechanisms such as a unified equipment coordinate system and automatic optical parameter setting, the standardization rate of the detection program reaches 100%, ensuring consistent detection standards for the same product across different devices and production lines, thus improving product quality stability.

[0068] The system provided in this embodiment of the invention, through the synergistic effect of various modules, increases the detection speed by 50% in the flying camera detection method, while achieving a 100% standardization rate of the detection procedure, significantly improving the production efficiency and detection reliability of the production line.

[0069] Specifically, the intelligent configuration module for detection items adopts a modular management architecture, dividing each detection function into independent module units. Each module unit can be independently enabled or disabled through a user interface, and the enabled / disabled status of each functional module is displayed in real time through a sub-interface, supporting automatic interface refresh. The modular management architecture adopted in this embodiment of the invention has the following beneficial effects:

[0070] 1. Rapid production line changeover: Due to differences in product processes, the required detection function modules may vary. The user interface allows for quick and intuitive activation or deactivation of detection modules, and the interface automatically refreshes to display the enabled / disabled status, significantly reducing equipment debugging time caused by product changeover and improving production line changeover efficiency.

[0071] 2. Enhance system resilience: When a functional module fails, it can be quickly isolated (disabled) through the interface, ensuring the normal operation of core functional modules, ensuring that the detection work is not excessively affected by the faulty module, and improving the overall stability and continuous operation capability of the system.

[0072] 3. Accelerate decision-making and reduce operational errors: Visualized enable / disable buttons for functional modules reduce the cognitive load on operators, enabling them to understand the status of each module more clearly, reducing misoperations caused by complex operations or unfamiliarity with module functions, and improving human-computer interaction efficiency and operational accuracy.

[0073] In one embodiment, the parameter adaptation module includes:

[0074] The illumination change adaptive unit is used to automatically adjust the deep learning contrast detection threshold according to illumination changes in low-light and overexposed scenes through a two-level compensation mechanism, and adopts logarithmic compensation for detection parameters. It effectively solves the problem of poor adaptability of detection parameters under different illumination conditions, reduces misjudgments caused by illumination fluctuations, improves detection stability and accuracy in complex illumination environments, and thus improves product detection yield.

[0075] The material batch adaptive unit is used to automatically adjust the deep learning detection threshold based on changes in material reflectivity and surface texture. It also uses logarithmic compensation for detection parameters to adapt to differences in material characteristics between different batches of products. This avoids the tedious process of manually readjusting parameters due to changes in material batches, reduces reliance on operator experience, reduces parameter configuration time and operational errors, ensures consistency of detection standards for different batches of products, and further improves detection efficiency and reliability.

[0076] Through the synergistic effect of the two adaptive units, and through automatic parameter adaptation and compensation, the detection deviation caused by changes in environmental factors (lighting) and product characteristics (material batches) is significantly reduced, manual intervention is reduced, the system's adaptability to complex scenarios is improved, and strong guarantees are provided for detection accuracy and stability.

[0077] The parameter adaptive module of this invention integrates a deep learning algorithm with a model calling mechanism, supporting one-click model import and export. For complex detection items, it employs a multi-model collaborative detection mode, eliminating the complex process of manually configuring model parameters and retraining for adaptation. This allows pre-trained models for residual adhesive detection to be quickly applied to different production lines or equipment, reducing the operational complexity of model migration and saving debugging time. For complex detection items such as residual adhesive morphology and PIN pin misalignment, the multi-model collaborative detection mode improves detection confidence through multi-model cross-validation, reducing potential blind spots or misjudgments in single models, improving detection accuracy in complex scenarios, and reducing the rate of missed defects. Simultaneously, the ease of model calling and automated multi-model collaboration reduce reliance on operators' deep learning model expertise. High-precision detection can be achieved without manual adjustment of model fusion parameters by technicians, shortening personnel training cycles and improving human-machine collaboration efficiency. Combined with an illumination change adaptive unit and a material batch adaptive unit, through rapid model import and export and a multi-model collaborative mode, it can flexibly address the detection needs of products with different materials and processes, enhancing the system's adaptability to diverse production scenarios.

[0078] The program reuse module of this invention supports the import and export of preset format detection point configuration files, which contain detection coordinate information and associated detection parameters. The rapid switching of detection modes includes one-click switching between single-object and multi-object detection modes, with simultaneous adjustment of motion control logic and image processing resource allocation during switching. This enables direct migration of detection programs across devices without recalibrating the image capture points, solving the problem of traditionally non-reusable programs between different production line devices, reducing device idle time from 22% to 5%. It also supports one-click switching between single-object and multi-object detection modes, automatically adjusting motion control logic and image processing resource allocation during switching, replacing the tedious manual modification of more than ten parameters, reducing the average number of single debugging clicks from over 120 to less than 40, shortening product changeover debugging time from 3 hours to 1 hour, and improving the standardization rate of detection programs to 100%, ensuring consistency of detection standards across different devices and production lines.

[0079] The system provided in this embodiment of the invention adopts a hierarchical interface architecture. The main interface displays at least the real-time monitoring window of the detection device, device parameter settings, and commonly used function switching components. Sub-interfaces display the enabled / disabled status of function modules, and secondary sub-interfaces integrate advanced parameter configuration functions. This has the following beneficial effects:

[0080] 1. Improved operational efficiency: The main interface centrally displays the real-time monitoring window of the detection device, commonly used parameter settings, and function switching components, reducing the frequency of operators switching between multi-level menus and compressing the number of single debugging clicks from 120+ to less than 40, thereby reducing operational complexity and improving human-computer interaction efficiency.

[0081] 2. Clear Function Status: The enabled / disabled status of each functional module is displayed in real time through the sub-interface. With the automatic interface refresh mechanism, operators can intuitively grasp the current system configuration, avoid misoperation caused by unclear functional status, and reduce the traditional misoperation rate by 18%.

[0082] 3. Layered optimization of parameter configuration: The sub-interface integrates advanced parameter configuration functions, which not only ensures that the main interface is simple and easy to use (meeting the needs of rapid operation), but also provides a deep configuration entry for professional debugging, achieving the dual goals of "simplified basic operation + controllable advanced configuration", and shortening the training cycle for technical personnel (from 2 weeks to 3 days).

[0083] 4. Adaptable to multiple operating scenarios: The hierarchical architecture takes into account both rapid production line changeover (one-click switching on the main interface) and complex parameter debugging (fine-grained configuration on the secondary interface), adapting to the operational needs of all scenarios from daily production to process optimization, thus improving the system's practicality and flexibility.

[0084] This embodiment also provides a method for changing and debugging an automatic optical inspection device based on the operating system provided in the above embodiments. For example... Figure 2 As shown, the process includes the following steps:

[0085] Step S1: System reset, start the detection software and enter the operating system interface to control the detection device to perform the reset operation.

[0086] Step S2: Product mode switching. Select the corresponding detection mode based on the mode switching component in the interface, and the system will automatically execute the mode switching process.

[0087] Step S3, laser code recognition and debugging: switch to the corresponding identifier recognition parameter configuration mode through the interface components.

[0088] Step S4, Connector Detection and Debugging: In the connector interface, select the enabled connecting parts and associated detection items based on the detection object configuration. After completing the basic debugging, trigger the automatic threshold setting function. The system automatically calculates the upper and lower limits of the detection parameters and enables the intelligent detection function to detect residual adhesive in the connecting parts. Specifically, in the connector interface, select the deep learning function, import the pre-trained model data with one click, and perform automatic detection of residual adhesive in the connecting parts.

[0089] Step S5, Connector testing and debugging: On the connector interface, select the enabled connecting parts and associated testing items based on the testing object configuration. After completing the basic debugging, trigger the automatic threshold setting function. The system automatically calculates the upper and lower limits of the testing parameters and enables the intelligent testing function to detect residual adhesive on the connecting parts.

[0090] Step S6: Debug the detection program by calling other detection function modules in sequence to complete the debugging of the corresponding detection items.

[0091] The method provided by this invention uses an interface mode switching component to quickly switch product modes. In processes such as laser code recognition and connector detection, the interface component and automatic setting function simplify parameter configuration. Combined with the import of pre-stored point files or the saving of the first teaching, cross-device reuse is achieved. Then, debugging is completed by calling the function modules in sequence. This significantly simplifies the model changeover debugging process, shortens the debugging time, reduces the complexity of operation, and improves the efficiency and flexibility of the equipment in adapting to different specifications of testing objects. At the same time, it ensures the reusability and consistency of the testing parameters.

[0092] In one optional implementation, enabling the intelligent detection function to detect residual adhesive in the connecting parts includes: selecting the deep learning function in the connector interface, importing the trained model data with one click, and automatically detecting residual adhesive in the connecting parts.

[0093] This invention enables rapid and accurate automatic detection of residual adhesive in connector components by selecting the deep learning function on the connector interface and importing pre-trained model data with one click. This eliminates the need for complex manual settings of residual adhesive detection parameters, simplifying the operation process and improving the efficiency and accuracy of residual adhesive detection. The use of mature models also enhances the intelligence and professionalism of the detection, helping equipment to complete changeover debugging and stable testing more efficiently.

[0094] In one optional implementation, during the configuration of detection items, a fault tolerance range corresponding to each detection parameter is preset. When the parameter configuration exceeds the preset range, a prompt is automatically triggered and the parameters exceeding the range are restricted from taking effect. The logical correlation between parameters is verified in real time. When a corresponding detection item is not enabled or a logical conflict in parameter configuration is detected, a prompt is issued. Parameter configuration permissions are restricted for detection items that have not completed the pre-debugging steps to prevent detection items from not being enabled or detection parameters from being configured incorrectly due to skipping steps.

[0095] In the configuration process of the detection items, this invention presets the fault tolerance range of each detection parameter and automatically prompts and restricts the parameters when they exceed the limits. It also verifies the logical correlation between parameters in real time and issues prompts when the corresponding detection item is not enabled or when there is a conflict in the parameter configuration. At the same time, it restricts the parameter configuration permissions of detection items that have not completed the pre-debugging steps. This can effectively prevent configuration errors caused by parameter exceeding limits, logical conflicts, or skipping steps, reduce the risk of detection items not being enabled or improper parameter configuration, improve the accuracy and standardization of detection item configuration, reduce debugging rework and detection deviations caused by configuration problems, and ensure the stability and reliability of the detection process through an intelligent fault tolerance mechanism.

[0096] In one alternative implementation, during the model changeover and debugging process, the operating status of each functional module is monitored in real time. If any functional module malfunctions, the faulty module is automatically disabled within seconds, while the core detection process continues to run.

[0097] This invention provides a solution for quickly isolating faults, preventing the spread of abnormalities from slowing down the overall progress, ensuring that key debugging steps do not stop, reducing the impact of faults on debugging efficiency, maintaining the operation of core testing functions, making the replacement debugging smoother and more stable, reducing the risk of debugging interruption due to module failure, and improving the reliability and continuity of the debugging process.

[0098] In one specific embodiment, with Figure 3 Taking the transformation of a certain model of CCS product as an example, the detailed product configuration is as follows: product size 1000*400mm, laser engraving code on the bottom of the product, two male pin connectors on the side, three rows of pins inside each connector (residual adhesive inside the connector needs to be checked), one female header code, nine-point laser solder joint type, a total of 33 laser soldering positions, 23 fuse positions, 18 hot riveting positions, and 7 auxiliary material positions. The model changeover and debugging process using the method provided in this embodiment of the invention is as follows:

[0099] ① Open the vision software, enter the software program interface, click the device return button, and the device will automatically return to its original position.

[0100] ② Product Mode Switching: Click the mode switching button on the main interface, and the program will execute the switching process internally, allowing you to switch to single-product mode with one click, such as... Figure 4 As shown. Compared to the original function which required entering the program's internal workflow bar and manually modifying more than ten parameters, the existing function interface allows for one-click switching. This simplifies the multi-step operation that previously required manually modifying more than ten parameters into a single-button operation, reducing the mode switching time from 15 minutes to within 3 seconds. The number of operation steps has been reduced by 95%, and the risk of configuration errors caused by manual parameter modification has been completely eliminated, significantly improving production changeover efficiency and operational reliability.

[0101] ③ Laser Code Recognition and Debugging: The main interface can be switched to the bottom scanning mode with one click, and the program will automatically switch to the bottom scanning parameter configuration. Compared with the original function that required manual camera connection and debugging, the existing function interface button can switch with one click, which greatly reduces the laser code recognition and debugging time. At the same time, by automatically loading the preset parameter configuration, the detection accuracy is improved by 40%, which greatly reduces the technical requirements of the operator.

[0102] ④ Connector Testing and Debugging: Click to enter the connector module interface and configure according to the product settings described above, such as... Figure 5 As shown, check Connector 1 / Connector 2 to enable. The connector interface, image window, and configuration parameters will be automatically displayed. Then check PIN1 / PIN2 / PIN3 to enable the three-row PIN detection process. Click the PIN processing button to enter the connector processing interface, as shown. Figure 6 As shown, after completing the debugging by sorting the program tools, click "One-click Standard Setting," and the program will automatically calculate the upper and lower limits. For connector residual adhesive detection: In the connector interface, select the deep learning function and import pre-trained model data with one click. Compared to the original function, which required entering the program's internal workflow and manually modifying over forty parameters with incomplete detection capabilities, the current function features a separate interface window with comprehensive and rich functionality, visual buttons, and switch-controlled detection items. Debugging is simple and convenient, eliminating the need to navigate through complex workflow parameters and significantly reducing debugging time and error rates.

[0103] ⑤ Equipment photo point debugging: Enter the operation control interface and import the CSV file of the photo points for this product with one click. (If this product is being produced for the first time, manual teaching is required. After completion, save the CSV file for cross-device import and export.) Figure 7 As shown. Compared to the original function, which required manual debugging of points in sequence, lacked the ability to insert or delete intermediate single points, and prevented the sharing of points between devices with low reuse rate, the current function allows for the individual insertion or deletion of points, import and export of program points, and reuse across devices and production lines, significantly reducing debugging time.

[0104] ⑥ Testing Program Debugging: Proceed sequentially to the testing interfaces for motherboard barcodes, laser welding, fuses, hot riveting, and auxiliary materials for debugging. Compared to the original function, which required finding and debugging over forty testing tools and parameters within a complex logical testing process, and involving over 120 debugging clicks and switching between multiple process levels, the current function features a single-item testing interface. Function integration combines multiple tools into a single tool, automatically linking parameters and simplifying the debugging process. The program is debugged according to the interface tool sequence, eliminating the need to navigate multiple process levels and search for corresponding tools within complex logical testing procedures. This significantly reduces debugging time, the skill level required of technical personnel, and the error rate.

[0105] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A control operating system for an automated optical inspection device based on a human-computer interaction interface, characterized in that, include: The core functional layer, hardware control layer, and user interface layer are as follows: The core functional layer integrates intelligent configuration, algorithm support, program reuse, and multi-layered fault tolerance mechanisms. It handles detection business logic, ensuring parameter adaptation, function switching, and process stability, driving intelligent detection and efficient model changeover. This includes: The intelligent configuration module for detection items is used to enable rapid switching and parameter setting of multiple core detection functions of automatic optical inspection equipment through a visual interactive component. The parameter adaptive module is used to automatically extract detection features and adaptively set threshold parameters using a deep learning algorithm. The deep learning algorithm has the ability to adaptively adjust to changes in ambient lighting and changes in the material properties of the detected object, and optimizes detection accuracy by dynamically adjusting the detection threshold and compensation parameters. The program reuse module supports the import and export of detection parameters across devices and the rapid switching of detection modes to adapt to the detection needs of different specifications of detection objects. The hardware control layer is used to control the CCD camera image acquisition and device movement, and to convert control commands into physical detection actions. The user interface layer serves as a human-computer interaction window, presenting status, providing intelligent guidance for operations, and monitoring and debugging processes through a visual interface.

2. The system according to claim 1, characterized in that, The intelligent configuration module for the detection items adopts a modular management architecture, which divides each detection function into an independent module unit. Each module unit can be independently enabled or disabled through the user interface and supports automatic interface refresh.

3. The system according to claim 1, characterized in that, The parameter adaptive module includes: The illumination change adaptive unit is used to automatically adjust the deep learning contrast detection threshold according to the illumination change for low-light and overexposed scenes through a two-level compensation mechanism, and uses logarithmic compensation for detection parameters. The material batch adaptive unit is used to automatically adjust the deep learning detection threshold based on changes in material reflectivity and surface texture, and employs logarithmic compensation for detection parameters.

4. The system according to claim 1 or 3, characterized in that, The parameter adaptive module integrates a deep learning algorithm with a model calling mechanism, supports one-click import and export of models, and adopts a multi-model collaborative detection mode for complex detection items.

5. The system according to claim 1, characterized in that, The program reuse module supports the import and export of detection point configuration files in a preset format. The configuration file contains detection coordinate information and associated detection parameters. The quick switching of detection modes includes one-click switching between single-object detection mode and multi-object detection mode. During the switching, motion control logic and image processing resource allocation are adjusted simultaneously.

6. The system according to claim 1, characterized in that, The system adopts a hierarchical interface architecture. The main interface displays at least the real-time monitoring window of the detection device, device parameter settings, and commonly used function switching components. The sub-interfaces are used to display the enabled / disabled status of function modules, and the secondary sub-interfaces integrate advanced parameter configuration functions.

7. A method for changing and debugging an automatic optical inspection device based on the system described in any one of claims 1-6, characterized in that, Includes the following steps: The system is reset, the testing software is started and the operating system interface is entered, and the testing equipment is controlled to perform a reset operation. When switching product modes, the corresponding detection mode is selected based on the mode switching component in the interface, and the system automatically executes the mode switching process. Laser code recognition debugging: Switch to the corresponding identifier recognition parameter configuration mode through the interface components; Connector testing and debugging: On the connector interface, select the enabled connecting parts and associated testing items based on the testing object configuration. After completing the basic debugging, trigger the automatic threshold setting function. The system automatically calculates the upper and lower limits of the testing parameters and enables the intelligent testing function to detect residual adhesive in the connecting parts. For equipment photo point debugging, import the pre-saved detection point configuration file. If it is the first detection, manually teach the detection points and save the configuration file for future reuse across devices. The testing program is debugged by sequentially calling other testing function modules to complete the debugging of the corresponding testing items.

8. The method according to claim 7, characterized in that, The method of enabling intelligent detection to detect residual adhesive in connecting parts includes: selecting the deep learning function in the connector interface, importing pre-trained model data with one click, and automatically detecting residual adhesive in connecting parts.

9. The method according to claim 7, characterized in that, During the configuration of detection items, a fault tolerance range is preset for each detection parameter. When the parameter configuration exceeds the preset range, a prompt is automatically triggered and the parameter exceeding the range is restricted from taking effect. The logical correlation between parameters is verified in real time. When a corresponding detection item is not enabled or a logical conflict in parameter configuration is detected, a prompt is issued. Parameter configuration permissions are restricted for detection items that have not completed the pre-debugging steps to prevent detection items from not being enabled or detection parameters from being configured incorrectly due to skipping steps.

10. The method according to claim 7, characterized in that, Also includes: During the model changeover and debugging process, the operating status of each functional module is monitored in real time. If any functional module malfunctions, the faulty module is automatically disabled within seconds, while the core testing process continues to run.

Images