Machine vision-based cabinet injection housing defect detection method and device

Abstract

The application provides a cabinet machine injection shell defect detection method and device based on machine vision. It relates to data processing technology. The method comprises: obtaining an initial image of the injection shell and a first real-time stress value from a stress sensing bolt arranged at a specified hole position of the injection shell; generating a preset stress simulation strategy, and controlling a mechanical arm to apply a simulated load to a target detection surface of the injection shell based on the preset stress simulation strategy; during the application of the simulated load, obtaining a real-time deformation image of the target detection surface and a second real-time stress value of the stress sensing bolt; determining a deformation detection value of the target detection surface according to the initial image, the real-time deformation image, the first real-time stress value and the second real-time stress value; and determining the detection attribute of the corresponding injection shell based on the matching result of the deformation detection value and the preset defect judgment rule. The application combines visual image analysis and stress deformation monitoring to accurately detect the injection shell and effectively identify potential defects.

Description

Technical Field

[0001] This invention relates to data processing technology, and more particularly to a method and apparatus for detecting defects in the injection-molded outer shell of a cabinet based on machine vision. Background Technology

[0002] Currently, in the manufacturing process of injection-molded housings, the quality inspection of the housings typically relies on manual visual inspection or traditional machine vision appearance inspection. While traditional machine vision inspection improves the efficiency of manual visual inspection, it is mainly limited to the identification of surface defects such as scratches and bubbles, and it is difficult to detect potential structural defects caused by insufficient internal structural strength or uneven material distribution.

[0003] For example, some injection-molded shells may appear intact in their natural state, but when subjected to external forces during actual assembly or use, they may undergo unexpected deformation or even breakage. Such latent defects are difficult to detect effectively in existing technologies. Therefore, there is an urgent need for an efficient and accurate defect detection solution for injection-molded shells. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed to provide a machine vision-based method and apparatus for detecting defects in the injection-molded outer shell of cabinets, which overcomes or at least partially solves the above problems.

[0005] According to one aspect of the present invention, a machine vision-based method for detecting defects in the injection-molded casing of a cabinet air conditioner is provided, comprising the following steps: Acquire an initial image of the injection-molded housing and the first real-time stress value from the stress-sensing bolts located at designated holes in the injection-molded housing; A preset stress simulation strategy is generated, and the robotic arm is controlled to apply simulated loads to the target detection surface of the injection-molded shell based on the preset stress simulation strategy; During the application of the simulated load, a real-time deformation image of the target detection surface and a second real-time stress value of the stress-sensing bolt are acquired. Based on the initial image, the real-time deformation image, the first real-time stress value, and the second real-time stress value, the deformation detection value of the target detection surface is determined; Based on the matching result between the deformation detection value and the preset defect judgment rule, the detection attributes corresponding to the injection molded shell are determined.

[0006] Optionally, in the method according to the invention, before acquiring an initial image of the injection-molded housing and a first real-time stress value from a stress-sensing bolt disposed at a designated hole in the injection-molded housing, the method further includes: Obtain the 3D design model of the injection molded shell, and extract the hole positions of all shell holes located in the injection molded shell based on the 3D design model; Retrieve a preset filter list, and in response to any of the hole positions being in the preset filter list, determine the housing hole position corresponding to that hole position function as the designated hole position; Based on the 3D design model, all remaining shell holes are determined in a linked manner, and based on the determination results, all shell holes with deformation linkage relationships are determined as designated holes.

[0007] Optionally, in the method according to the present invention, the step of determining all remaining housing holes in a coordinated manner based on the three-dimensional design model includes: Determine all structural surfaces of the injection-molded housing based on the 3D design model; When it is determined that any two outer shell holes are located on the same structural surface, the structural rigidity strength of the structural surface is obtained; When it is determined that the structural rigidity is less than a preset first rigidity threshold, the two shell holes are determined to have a deformation linkage relationship. When it is determined that the structural rigidity is greater than or equal to the preset first rigidity threshold and less than the preset second rigidity threshold, the hole distance between the two shell holes is obtained. When the distance between the holes is determined to be less than a preset distance threshold, any two outer shell holes are determined to have a deformation linkage relationship; or... Determine the different structural surfaces where any two outer shell holes are located in adjacent positions, and obtain the surface connection method corresponding to the different structural surfaces; When the surface connection method is determined to be rigid connection, the rigid connector corresponding to the rigid connection is determined according to the three-dimensional design model, and the connection rigidity strength of the rigid connector is obtained. When the connection rigidity is determined to be less than the preset third rigidity threshold, the hole distance between any two corresponding outer shell holes is obtained; When the distance between holes is determined to be less than a preset distance threshold, any two holes in the outer shell are determined to have a deformation linkage relationship.

[0008] Optionally, in the method according to the present invention, generating a preset force simulation strategy includes: Identify the structural features of the injection-molded housing, wherein the structural features include a first structural surface with specified holes, a second structural surface without specified holes, and surface connection relationships between different structural surfaces in adjacent positions; Based on the structural features, the corresponding single-sided force application location and force value are planned for each structural surface to generate a single-sided test strategy. Based on the surface connection relationship between different structural surfaces in adjacent positions, plan the corresponding synchronous force application positions and force values, and generate a fusion test strategy; The single-sided testing strategy is combined with the fusion testing strategy to form the preset force simulation strategy.

[0009] Optionally, in the method according to the present invention, the step of planning the force application location and force value for each structural surface according to the structural features, and generating a single-surface testing strategy, includes: In response to the structural surface being the first structural surface, the force application position corresponding to the single-sided force application is determined as the intersection of the center axis of the hole at the designated hole position located on the first structural surface and the first structural surface. The force value for the corresponding single-sided force application is calculated based on the product of the rated tightening force at the designated hole position and a preset detection coefficient; or... In response to the structural surface being the second structural surface, the applied force value of the corresponding single-surface applied force is calculated based on the product of the rated bearing value of the corresponding second structural surface and the preset detection coefficient, and the surface size of the corresponding second structural surface is compared with the preset size threshold. If the size of the response surface is smaller than a preset size threshold, the second structural surface is divided into an array, and each array point is determined as the force application position for the corresponding single-sided force application. Otherwise, the center point of the corresponding second structural surface is determined as the force application position for the corresponding single-sided force application.

[0010] Optionally, in the method according to the present invention, the step of planning the corresponding synchronous force application positions and force values ​​based on the surface connection relationship between different structural surfaces in adjacent positions, and generating a fusion test strategy, includes: The response is that the surface connection method of different structural surfaces in adjacent positions is a rigid connection, and the specified hole position is determined in the different structural surfaces; When it is determined that the specified holes located on different structural surfaces have a deformation linkage relationship, the different structural surfaces are determined as a strongly correlated set; otherwise, they are determined as a weakly correlated set. Based on the strongly correlated set, the specified hole positions with deformation linkage are determined as the corresponding synchronous force application positions, and the corresponding synchronous force application value is calculated based on the product of the rated tightening force of the corresponding specified hole position and the preset detection coefficient; or, Based on the weak correlation set, the surface connection line corresponding to the rigid connection is determined as the force application position of the corresponding synchronous force application, and the force value of the corresponding synchronous force application is calculated by multiplying the minimum rated bearing value of the corresponding different structural surfaces with the preset detection coefficient.

[0011] Optionally, in the method according to the present invention, the step of controlling the robotic arm to apply a simulated load to the target detection surface of the injection-molded housing based on a preset force simulation strategy includes: According to the preset force simulation strategy, when the robotic arm applies a simulated load to the structural surface identified as the target detection surface, the control fixture unit fixes and clamps all structural surfaces of the injection-molded shell that are not related to the target detection surface.

[0012] Optionally, in the method according to the present invention, determining the deformation detection value of the target detection surface based on the initial image, the real-time deformation image, the first real-time stress value, and the second real-time stress value includes: The initial image and the real-time deformed image are compared for features to extract the image deformation. The stress change is calculated based on the stress difference between the second real-time stress value and the first real-time stress value. The image deformation and the stress change are fused to generate the deformation detection value.

[0013] Optionally, in the method according to the present invention, the step of performing feature comparison between the initial image and the real-time deformed image to extract image deformation includes: When the target detection surface contains a specified hole, the hole contour corresponding to the specified hole in the initial image and the real-time deformation image are identified respectively; Calculate the image deformation of the hole contour in the real-time deformed image relative to the corresponding hole contour in the initial image.

[0014] Optionally, in the method according to the present invention, determining the detection attributes corresponding to the injection-molded shell based on the matching result of the deformation detection value and the preset defect judgment rule includes: The image deformation value in the deformation detection value is compared with a preset image deformation threshold; The stress change in the deformation detection value is compared with a preset stress change threshold. When it is determined that the image deformation is greater than the image deformation threshold, and / or the stress change is greater than the stress change threshold, the resistance change of the injection molded shell is determined according to a preset resistance simulation strategy, and the detection attribute corresponding to the injection molded shell is determined based on the resistance change.

[0015] Optionally, in the method according to the present invention, determining the resistance change of the injection-molded housing according to a preset resistance simulation strategy, and determining the detection attribute corresponding to the injection-molded housing based on the resistance change, includes: A sliding component is assembled inside the injection-molded housing, and a preset track is defined within the housing. The robotic arm is controlled to apply a driving force to the slidable component, causing it to slide along a preset track; If the resistance value of the corresponding drive sliding is obtained, and the resistance value is greater than the preset resistance threshold, then the detection attribute of the corresponding injection molded shell is determined to be a defect attribute.

[0016] Optionally, in the method according to the present invention, the stress-sensing bolt is a passive wireless monitoring bolt; The step of obtaining the first real-time stress value from the stress-sensing bolts located at designated holes in the injection-molded housing includes: The initial pressure data of the passive wireless monitoring bolt in its natural state is obtained through a wireless reading terminal and used as the first real-time stress value.

[0017] According to another aspect of the present invention, a machine vision-based defect detection device for cabinet air conditioner injection molded shells is provided, comprising: The data acquisition module is used to acquire an initial image of the injection-molded housing and the first real-time stress value from the stress-sensing bolts located at designated holes in the injection-molded housing. The load control module is used to control the robotic arm to apply simulated loads to the target detection surface of the injection-molded shell based on a preset force simulation strategy. The real-time monitoring module is used to acquire real-time deformation images of the target detection surface and the second real-time stress value of the stress-sensing bolt during the application of the simulated load. The feature analysis module is used to determine the deformation detection value of the target detection surface based on the initial image, the real-time deformation image, the first real-time stress value, and the second real-time stress value. The defect determination module is used to determine the detection attributes of the corresponding injection molded shell based on the matching result of the deformation detection value and the preset defect determination rule.

[0018] According to the present invention, by simultaneously acquiring the initial image of the injection-molded shell and the first real-time stress value of the stress-sensing bolt at the specified hole position, and then combining it with a preset stress simulation strategy, the robotic arm precisely applies the simulated load, and simultaneously acquires the real-time deformation image of the target detection surface and the second real-time stress value of the stress-sensing bolt, thereby realizing the fusion detection of deformation visual information and stress perception information. This enables the comprehensive and accurate determination of the deformation detection value of the target detection surface. Furthermore, based on the deformation detection value and the preset defect judgment rule, the corresponding detection attributes of the injection-molded shell can be objectively and stably determined, effectively improving the accuracy, reliability, and automation of the structural strength and deformation defect detection of the injection-molded shell, reducing human inspection errors and subjective interference, and ensuring the product quality of the injection-molded shell. Attached Figure Description

[0019] Figure 1 A flowchart of a machine vision-based method for detecting defects in the injection-molded casing of a cabinet air conditioner, according to an embodiment of the present invention, is shown. Figure 2 A schematic diagram showing the force application location according to an embodiment of the present invention is provided; Figure 3A schematic diagram of a surface connecting line according to an embodiment of the present invention is shown; Figure 4 A structural block diagram of a machine vision-based cabinet injection molding shell defect detection device according to another embodiment of the present invention is shown. Detailed Implementation

[0020] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0021] To address the problems existing in the background art described above, the inventors proposed the solution of this invention. One embodiment of this invention provides a machine vision-based method for detecting defects in the injection-molded casing of a cabinet air conditioner, which can be executed in a computing device.

[0022] Figure 1 A flowchart of a machine vision-based defect detection method for cabinet air conditioner injection molded casing according to an embodiment of the present invention is shown, the method being adapted to be executed in a computing device.

[0023] like Figure 1 As shown, the machine vision-based defect detection method for cabinet air conditioner injection molded shells proposed in this embodiment begins with step S102, which includes the following: Acquire an initial image of the injection-molded housing and the first real-time stress value from the stress-sensing bolts located at designated holes in the injection-molded housing.

[0024] For example, in this embodiment, the initial image refers to the image acquired by the industrial camera acquisition unit on the injection molded shell without external force. The designated hole position refers to the hole on the injection molded shell used for assembly connection or other functions, such as the mounting hole and wiring hole on the air conditioner cabinet shell. The stress sensing bolt is pre-laid in these designated holes, which can not only play the role of connection and fastening, but also acquire and feedback the axial pressure or shear force on the bolt shank in real time. When the injection-molded housing is not subjected to external force, the pressure data monitored by the stress-sensing bolt is the first real-time stress value, which can reflect the initial stress state of the injection-molded housing under no external force.

[0025] Furthermore, the aforementioned "acquiring an initial image of the injection-molded housing and a first real-time stress value from the stress-sensing bolts located at designated holes in the injection-molded housing" further includes the following steps: Obtain the 3D design model of the injection molded shell, and extract the hole positions of all shell holes located in the injection molded shell based on the 3D design model; Retrieve a preset filter list, and in response to any of the hole positions being in the preset filter list, determine the housing hole position corresponding to that hole position function as the designated hole position; Based on the 3D design model, all remaining shell holes are determined in a linked manner, and based on the determination results, all shell holes with deformation linkage relationships are determined as designated holes.

[0026] For example, in this embodiment, the server will first obtain the three-dimensional design model of the injection molded shell and automatically identify the hole position function of all shell holes on the injection molded shell, such as determining whether the shell hole is a mounting hole, positioning hole, heat dissipation hole or decorative hole, etc. Next, the server will retrieve the preset filter list, which is established in advance based on the structural mechanical properties of the injection molded shell. The preset filter list usually includes hole types that have a significant impact on structural strength, such as main mounting holes and load-bearing holes. These holes often bear large loads or stresses in the actual use of the injection molded shell. When any hole position function is in the preset filter list, the server will determine the shell hole position corresponding to the hole position function as the designated hole position, so as to improve the targeting of subsequent inspections and avoid wasting inspection resources. In the overall structure of the injection molded housing, some housing holes may not be functionally critical, but they are located on the critical path of stress transmission. That is, when a specified hole is subjected to force, these housing holes on the critical path will also undergo slight deformation. Therefore, for the remaining holes that are not selected in the preset screening list, the server will determine all remaining shell holes based on the 3D design model, and determine all shell holes with deformation linkage relationship as designated holes according to the determination results, so as to ensure the comprehensiveness of subsequent inspection.

[0027] Furthermore, the aforementioned "determining all remaining housing holes based on the 3D design model" also includes the following steps: Determine all structural surfaces of the injection-molded housing based on the 3D design model; When it is determined that any two outer shell holes are located on the same structural surface, the structural rigidity strength of the structural surface is obtained; When it is determined that the structural rigidity is less than a preset first rigidity threshold, the two shell holes are determined to have a deformation linkage relationship. When it is determined that the structural rigidity is greater than or equal to the preset first rigidity threshold and less than the preset second rigidity threshold, the hole distance between the two shell holes is obtained. When the distance between the holes is determined to be less than a preset distance threshold, any two outer shell holes are determined to have a deformation linkage relationship; or... Determine the different structural surfaces where any two outer shell holes are located in adjacent positions, and obtain the surface connection method corresponding to the different structural surfaces; When the surface connection method is determined to be rigid connection, the rigid connector corresponding to the rigid connection is determined according to the three-dimensional design model, and the connection rigidity strength of the rigid connector is obtained. When the connection rigidity is determined to be less than the preset third rigidity threshold, the hole distance between any two corresponding outer shell holes is obtained; When the distance between holes is determined to be less than a preset distance threshold, any two holes in the outer shell are determined to have a deformation linkage relationship.

[0028] For example, in this embodiment, the server first determines all structural surfaces included in the injection-molded shell, such as the side panel and top panel of the air conditioner cabinet, based on the three-dimensional design model. When the server determines that any two shell openings are located on the same structural surface, the server will obtain the structural rigidity strength of that structural surface, which refers to the ability of the structural surface to resist deformation. In the server, a first rigidity threshold and a second rigidity threshold can be preset in advance. These can be empirical values ​​preset based on the general mechanical properties of the injection molding material and the minimum safe deformation of the shell. The first rigidity threshold is less than the second rigidity threshold. For example, the first rigidity threshold can be set to 50 N / mm and the second rigidity threshold can be set to 100 N / mm. When the structural rigidity is determined to be less than the preset first rigidity threshold, it indicates that the structural surface is relatively soft and is very easy to deform as a whole when subjected to force. At this time, there is a significant mechanical transmission effect between any two holes on the structural surface. Therefore, the server will directly determine that these two shell holes have a deformation linkage relationship. When the structural rigidity is determined to be greater than or equal to the preset first rigidity threshold and less than the preset second rigidity threshold, it indicates that the structural mask has a certain rigidity, but local deformation transmission may still occur. It can be understood as medium rigidity. At this time, the server will obtain the hole distance between the two corresponding shell holes. When the distance between the holes is determined to be less than the preset distance threshold, it indicates that the two holes in the outer shell are close to each other and there may be a certain mechanical transmission effect. Therefore, the server will determine that the two holes in the outer shell have a deformation linkage relationship. In addition, there may be cases where the shell openings are distributed on different structural surfaces. When it is determined that any two shell openings are located on different structural surfaces in adjacent positions, the server will first obtain the surface connection method of the two different structural surfaces. Surface connection methods include rigid connection, which refers to a connection method between two structural surfaces that has no relative rotation or displacement and is completely fixed, such as snap-fit ​​connection and bolt fastening. Unlike rotational connection, rigid connection does not move when subjected to force and may have a certain mechanical transmission effect. Rotational connection refers to a connection method between two structural surfaces that can rotate around an axis and have relative movement clearance, such as hinge and pivot, which will rotate relative to each other when subjected to force. When the surface connection method is determined to be rigid connection, the server will determine the corresponding rigid connector based on the 3D design model and obtain the connection rigidity strength of the rigid connector. The connection rigidity strength reflects the ability of the rigid connector to resist relative displacement. For example, if the rigid connector is a metal self-tapping screw, its connection rigidity strength is high; if it is a plastic clip, its connection rigidity strength is relatively low. When the connection rigidity is determined to be less than the preset third rigidity threshold, it indicates that there is a certain elastic deformation space in the connection part, which can transmit stress or cause small displacement. At this time, the server will further obtain the hole distance between the two shell holes. When the hole distance is determined to be less than the preset distance threshold, the server will determine that the two shell holes have a deformation linkage relationship. This embodiment can accurately determine whether two outer shell holes have a deformation linkage relationship, for two cases where the two outer shell holes are located on the same side or different sides.

[0029] Step S104 includes the following: A preset stress simulation strategy is generated, and the robotic arm is controlled to apply simulated loads to the target detection surface of the injection-molded shell based on the preset stress simulation strategy.

[0030] For example, in this embodiment, the server generates a preset stress simulation strategy and controls the robotic arm to apply simulated loads to the target detection surface of the injection molded shell based on the preset stress simulation strategy in order to determine whether there are potential defects inside the injection molded shell. The target inspection surface is a surface area of ​​the injection-molded shell that will be subjected to stress simulation. The simulated load is, for example, the tightening torque or wind pressure load during the installation of an air conditioner shell. Based on the simulated load, potential hidden defects such as internal cracks, bubbles or insufficient material strength in the injection-molded shell can be effectively stimulated. Controlling the robotic arm to apply the simulated load to the target inspection surface of the injection-molded shell can also ensure the controllability and accuracy of the simulation process.

[0031] Furthermore, the aforementioned "generating a preset force simulation strategy" also includes the following steps: Identify the structural features of the injection-molded housing, wherein the structural features include a first structural surface with specified holes, a second structural surface without specified holes, and surface connection relationships between different structural surfaces in adjacent positions; Based on the structural features, the corresponding single-sided force application location and force value are planned for each structural surface to generate a single-sided test strategy. Based on the surface connection relationship between different structural surfaces in adjacent positions, plan the corresponding synchronous force application positions and force values, and generate a fusion test strategy; The single-sided testing strategy is combined with the fusion testing strategy to form the preset force simulation strategy.

[0032] For example, in this embodiment, the server first analyzes the overall structure of the injection-molded shell to identify key structural features. These structural features mainly include a first structural surface with specified holes, a second structural surface without specified holes, and the surface connection relationship between different structural surfaces in adjacent positions. The first structural surface can be understood as the key stress-bearing surface on the injection-molded shell where stress-sensing bolts are arranged, such as the back panel or base of an air conditioner cabinet. These areas are usually assembly connection points, where stress is concentrated and cracks are prone to occur. The second structural surface refers to the area where stress-sensing bolts are not set but where structural strength still needs to be tested, such as large side panels or top panels. These areas may have problems such as uneven material distribution or insufficient wall thickness. The surface connection relationship indicates the connection form between different structural surfaces, such as the snap-fit ​​connection or screw-locking connection between the side plate and the back plate. These connection parts are often the bottleneck of stress transmission and are prone to disengagement or breakage. Next, the server will plan the corresponding force application location and force value for each structural surface based on the structural characteristics, and generate a single-surface test strategy to specifically detect local defects on each structural surface, such as micro-cracks around a mounting hole or uneven wall thickness of a side plate. Next, the server will plan the application positions and values ​​of synchronous force based on the surface connection relationship between different structural surfaces in adjacent positions, and generate a fusion test strategy to detect the cooperative force characteristics between structural surfaces. For example, when a force in a specific direction is applied to two adjacent structural surfaces at the same time, if the buckle strength at the connection is insufficient, significant relative displacement or stress anomalies will occur. The fusion test strategy can effectively detect associated defects that are difficult to find in single-surface testing, such as loose connectors, insufficient welding strength, or cross-surface cracks. Finally, the server combines the single-sided testing strategy with the integrated testing strategy to form the final preset stress simulation strategy, which improves the coverage and accuracy of defect detection and avoids the risk of missed detection.

[0033] Furthermore, the aforementioned "planning the force application location and force value for each structural surface according to the structural features, and generating a single-surface test strategy" also includes the following steps: In response to the structural surface being the first structural surface, the force application position corresponding to the single-sided force application is determined as the intersection of the center axis of the hole at the designated hole position located on the first structural surface and the first structural surface. The force value for the corresponding single-sided force application is calculated based on the product of the rated tightening force at the designated hole position and a preset detection coefficient; or... In response to the structural surface being the second structural surface, the applied force value of the corresponding single-surface applied force is calculated based on the product of the rated bearing value of the corresponding second structural surface and the preset detection coefficient, and the surface size of the corresponding second structural surface is compared with the preset size threshold. If the size of the response surface is smaller than a preset size threshold, the second structural surface is divided into an array, and each array point is determined as the force application position for the corresponding single-sided force application. Otherwise, the center point of the corresponding second structural surface is determined as the force application position for the corresponding single-sided force application.

[0034] For example, in this embodiment, when the structural surface is the first structural surface, that is, the designated holes for stress-sensitive bolts are arranged on this structural surface, during the actual assembly process, the axial tensile force generated when the bolts are tightened will be transmitted to the outer shell surface through the hole center axis. Therefore, the server will determine the force application position corresponding to the single-sided force application as the intersection of the hole center axis corresponding to the designated hole located on the first structural surface and the first structural surface, such as... Figure 2 The solid circle shown allows for a more realistic simulation of the stress state during assembly based on the applied force location, thereby effectively detecting potential hazards such as microcracks around designated holes, stripping risks, or insufficient strength of countersunk holes. After determining the force application location, the server will calculate the corresponding single-sided force value based on the product of the rated tightening force of the corresponding specified hole and the preset detection coefficient. The rated tightening force refers to the axial force corresponding to the standard assembly torque specified in the product manual or design specification for the specified hole. The preset detection coefficient is a parameter used to adjust the detection sensitivity. It usually ranges from 0.8 to 1.2. When the preset detection coefficient is 1.0, it means that the actual assembly stress is completely simulated. When the value is less than 1.0, it is suitable for scenarios where the shell strength is questionable or for non-destructive sampling inspection. When the value is greater than 1.0, it is used for verification of extreme working conditions, which can ensure the reference value of the applied force value and also have a certain degree of flexibility. When the structural surface is the second structural surface, that is, no specified holes are arranged on the structural surface, the server will calculate the force value of the corresponding single-sided force based on the product of the rated load value of the corresponding second structural surface and the preset detection coefficient. The rated load value refers to the maximum static load that the structural surface is expected to bear during normal use, such as the stacking pressure that the side panel of the air conditioner cabinet may bear during transportation. The server also uses the preset detection coefficient to scale the rated load value to determine the safe test load. In determining the location of force application, the server first compares the surface dimensions of the corresponding second structural surface with the preset size threshold. The surface dimensions refer to the outline dimensions of the structural surface, and the preset size threshold is an empirical value set based on the deformation transmission characteristics of the injection molding material. If the surface size is smaller than the preset size threshold, it means that the surface area of ​​the second structural surface is small. If force is applied only at the center point, it may not be able to effectively cover potential defects in the edge area or corner. Therefore, the server will divide the second structural surface into an array and determine each array point as the corresponding force application position for single-sided force application to avoid missing defects at the edge. If the surface size is greater than or equal to the preset size threshold, it means that the surface area of ​​the second structural surface is large. Its central area is usually the location where the deformation is most obvious. Therefore, the server will determine the center point of the surface of the corresponding second structural surface as the force application position of the corresponding single surface to maximize the overall deformation and thus effectively evaluate the overall rigidity of the structural surface. This embodiment improves the coverage and precision of the detection by determining the force application locations on different structural surfaces. It not only simulates real stress conditions but also enhances the applicability and accuracy of the detection scheme.

[0035] Furthermore, the aforementioned "planning the corresponding synchronous force application positions and force values ​​based on the surface connection relationship between different structural surfaces in adjacent positions, and generating a fusion test strategy" also includes the following steps: The response is that the surface connection method of different structural surfaces in adjacent positions is a rigid connection, and the specified hole position is determined in the different structural surfaces; When it is determined that the specified holes located on different structural surfaces have a deformation linkage relationship, the different structural surfaces are determined as a strongly correlated set; otherwise, they are determined as a weakly correlated set. Based on the strongly correlated set, the specified hole positions with deformation linkage are determined as the corresponding synchronous force application positions, and the corresponding synchronous force application value is calculated based on the product of the rated tightening force of the corresponding specified hole position and the preset detection coefficient; or, Based on the weak correlation set, the surface connection line corresponding to the rigid connection is determined as the force application position of the corresponding synchronous force application, and the force value of the corresponding synchronous force application is calculated by multiplying the minimum rated bearing value of the corresponding different structural surfaces with the preset detection coefficient.

[0036] For example, in this embodiment, when the surface connection method of different structural surfaces in adjacent positions is a rigid connection, the server will determine the specified hole positions located on different structural surfaces; When it is determined that the specified holes located on different structural surfaces have a deformation linkage relationship, the server will identify the different structural surfaces as a strongly correlated set. That is, when the specified hole on one structural surface in the strongly correlated set is subjected to force, the stress can be effectively transferred to the other structural surface through the connection part, causing the two structural surfaces to produce synchronous deformation response. Conversely, when it is determined that the specified holes located on different structural surfaces do not have a deformation linkage relationship, the server will determine the different structural surfaces as a weakly associated set. For strongly correlated sets, the specified holes with deformation linkage are usually the installation positions of the connectors. Applying force directly at these positions can most realistically simulate the stress state after assembly. Therefore, the server will determine the specified holes with deformation linkage as the corresponding synchronous force application positions and calculate the corresponding synchronous force application value based on the product of the rated tightening force of the corresponding specified hole and the preset detection coefficient. Synchronous force application refers to the robotic arm simultaneously applying forces in the same or opposite directions to multiple designated holes to simulate complex stress conditions in actual use. The applied force value is the product of the rated tightening force and the coefficient, which can ensure sufficient detection force and avoid overload damage. For weakly correlated sets, due to the weak mechanical transmission between holes, simply applying force to a specified hole may not effectively excite defects at the connection. Therefore, the server will determine the surface connection line corresponding to the rigid connection as the force application location for synchronous force application. The surface connection line refers to the edge or joint where two structural surfaces intersect, such as... Figure 3 As shown, by applying pressure or tension to the surface connecting line, the bonding strength at the joint can be directly tested, for example, to check whether there are problems such as incomplete welding of the weld seam, air bubbles or delamination of the adhesive surface; The server calculates the corresponding synchronous force value based on the product of the minimum rated load-bearing value of different structural surfaces and the preset detection coefficient. The minimum rated load-bearing value of different structural surfaces is selected because the load-bearing capacity of different structural surfaces in a weakly correlated structure may vary. Selecting the minimum rated load-bearing value can ensure that the force application process will not cause irreversible damage to the relatively weak structural surface, thus ensuring the safety of the detection. This embodiment can focus on verifying the strength of connection points for strongly correlated sets, and on verifying the overall stability of connection surfaces for weakly correlated sets. It can effectively solve the problem that defects at connection points are difficult to detect in traditional methods, and significantly improve the comprehensiveness and accuracy of detection.

[0037] Furthermore, the aforementioned "controlling the robotic arm to apply simulated loads to the target detection surface of the injection-molded shell based on a preset force simulation strategy" also includes the following steps: According to the preset force simulation strategy, when the robotic arm applies a simulated load to the structural surface identified as the target detection surface, the control fixture unit fixes and clamps all structural surfaces of the injection-molded shell that are not related to the target detection surface.

[0038] For example, in this embodiment, according to the preset force simulation strategy, when the robotic arm applies a simulated load to the structural surface that is determined to be the target detection surface, the server controls the clamping unit to fix and clamp all structural surfaces of the injection-molded shell that are not related to the target detection surface. In other words, if there is no direct rigid connection between a certain structural surface and the target detection surface, and the conditions for determining the deformation linkage relationship are not met, then the deformation of these structural surfaces will not be the focus of this detection, and the shaking generated during the simulation will become detection interference. Therefore, when the robotic arm applies a simulated load to the structural surface that is identified as the target detection surface, the server controls the clamping unit to fix and clamp all structural surfaces of the injection-molded shell that are not related to the target detection surface. The fixture unit can take various forms, such as pneumatic grippers, hydraulic pressure plates, or vacuum suction cups. The fixture unit can keep all structural surfaces of the injection-molded shell that are not related to the target inspection surface relatively fixed in space, reducing shaking interference and improving the accuracy and specificity of the final inspection.

[0039] Step S106 includes the following: During the application of the simulated load, a real-time deformation image of the target detection surface and a second real-time stress value of the stress-sensing bolt are acquired.

[0040] For example, in this embodiment, during the process of the robotic arm applying the simulated load, the server will simultaneously acquire the real-time deformation image of the target detection surface and the second real-time stress value output by the stress-sensing bolt. Not only can it perform real-time observation from the macroscopic surface, but it can also collect real-time data on the internal force data of the target detection surface based on the second real-time stress value to ensure the accuracy of the simulation results.

[0041] Step S108 includes the following: The deformation detection value of the target detection surface is determined based on the initial image, the real-time deformation image, the first real-time stress value, and the second real-time stress value.

[0042] For example, in this embodiment, the server compares the initial image with the real-time deformation image to determine the image deformation, and determines the stress change based on the second real-time stress value and the first real-time stress value, and finally generates a deformation detection value so that the detection attributes of the injection molded shell can be accurately determined based on the deformation detection value.

[0043] Furthermore, the aforementioned "determining the deformation detection value of the target detection surface based on the initial image, the real-time deformation image, the first real-time stress value, and the second real-time stress value" also includes the following steps: The initial image and the real-time deformed image are compared for features to extract the image deformation. The stress change is calculated based on the stress difference between the second real-time stress value and the first real-time stress value. The image deformation and the stress change are fused to generate the deformation detection value.

[0044] For example, in this embodiment, firstly, the server compares the features of the initial image and the real-time deformation image to extract the image deformation variable. The image deformation variable is an intuitive parameter that characterizes the surface geometric deformation of the injection molded shell. Then, the difference between the second real-time stress value collected under the applied force state and the first real-time stress value under the natural state is calculated to calculate the stress change. Afterwards, the server will fuse the image deformation and stress change to generate deformation detection values. The specific method of data fusion can be a weighted fusion algorithm. Since the image deformation and stress change have different units of measurement, the server will first normalize the image deformation and stress change, then preset the weight coefficients according to the structural characteristics of the injection molded shell, and finally obtain the deformation detection values. By fusing image deformation with stress change, the obtained deformation detection values ​​can more realistically and comprehensively characterize the quality status of the injection molded shell.

[0045] Furthermore, the aforementioned "comparing the features of the initial image and the real-time deformed image to extract the image deformation" also includes the following steps: When the target detection surface contains a specified hole, the hole contour corresponding to the specified hole in the initial image and the real-time deformation image are identified respectively; Calculate the image deformation of the hole contour in the real-time deformed image relative to the corresponding hole contour in the initial image.

[0046] For example, in this embodiment, when the target detection surface contains a specified hole, the server will identify the hole contour corresponding to the specified hole in the initial image and the real-time deformation image respectively. After identifying the hole contour, the server compares the two hole contours to calculate the image deformation of the hole contour in the real-time deformed image relative to the corresponding hole contour in the initial image. This embodiment can determine the specific degree of deformation based on the hole contour, and has a certain degree of accuracy.

[0047] Step S110 includes the following: Based on the matching result between the deformation detection value and the preset defect judgment rule, the detection attributes corresponding to the injection molded shell are determined.

[0048] For example, in this embodiment, the server determines the detection attributes of the corresponding injection-molded shell based on the matching result of the deformation detection value and the preset defect judgment rule, thereby improving the accuracy of the detection results.

[0049] Furthermore, the aforementioned "determining the detection attributes corresponding to the injection-molded shell based on the matching result of the deformation detection value and the preset defect judgment rule" also includes the following steps: The image deformation value in the deformation detection value is compared with a preset image deformation threshold; The stress change in the deformation detection value is compared with a preset stress change threshold. When it is determined that the image deformation is greater than the image deformation threshold, and / or the stress change is greater than the stress change threshold, the resistance change of the injection molded shell is determined according to a preset resistance simulation strategy, and the detection attribute corresponding to the injection molded shell is determined based on the resistance change.

[0050] For example, in this embodiment, the server first compares the image deformation value in the deformation detection value with a preset image deformation threshold. The preset image deformation threshold is a critical value for determining whether the surface deformation of the injection molded shell exceeds the limit, and it can be determined based on statistical analysis of a large amount of historical detection data. If the measured image deformation exceeds the preset image deformation threshold, it indicates that the surface deformation of the injection molded shell has exceeded the normal fluctuation range, and there may be risks of surface cracks, insufficient wall thickness or abnormal structural stiffness. Secondly, the server compares the stress change in the deformation detection value with the preset stress change threshold, which is determined based on the stress response distribution of qualified samples. The stress change reflects the mechanical transmission efficiency of the internal structure of the injection molded shell. If the measured stress change is greater than the preset stress change threshold, it may mean that there are bubbles, shrinkage cavities or uneven material distribution inside the injection molded shell, resulting in stress concentration. If the measured stress change is significantly less than or equal to the preset stress change threshold, it may mean that there is porosity or insufficient material strength, resulting in stress attenuation. Based on the above comparison results, when it is determined that the image deformation is greater than the image deformation threshold and / or the stress change is greater than the stress change threshold, the server will determine the resistance change of the injection molded shell according to the preset resistance simulation strategy, and determine the detection attributes of the corresponding injection molded shell based on the resistance change. The preset resistance simulation strategy is a secondary verification method for functional defects in injection molded shells. After the initial judgment of anomalies, the product is not directly judged as scrap, but resistance simulation is initiated. This is because some non-structural factors may cause abnormal initial test data, such as ambient temperature fluctuations and instantaneous sensor noise. By performing resistance simulations, such as simulating the assembly sliding process of internal components, it is possible to verify whether there is substantial deformation in the injection-molded shell that would cause assembly interference, thus ensuring detection efficiency while significantly reducing the false judgment rate.

[0051] Furthermore, the aforementioned "determining the resistance change of the injection-molded housing according to a preset resistance simulation strategy, and determining the corresponding detection attributes of the injection-molded housing based on the resistance change" also includes the following steps: A sliding component is assembled inside the injection-molded housing, and a preset track is defined within the housing. The robotic arm is controlled to apply a driving force to the slidable component, causing it to slide along a preset track; If the resistance value of the corresponding drive sliding is obtained, and the resistance value is greater than the preset resistance threshold, then the detection attribute of the corresponding injection molded shell is determined to be a defect attribute.

[0052] For example, in this embodiment, in the production and assembly process of the injection-molded shell, in addition to appearance and structural strength issues, the dimensional stability of the internal structure is also crucial. If the shell undergoes internal deformation after injection molding cooling or stress, it often leads to obstruction of the movement of the internal sliding components during subsequent assembly. Therefore, the server will assemble a sliding component in the internal space of the injection molded shell and determine the preset track in the internal space. The sliding component refers to the part of the injection molded shell that needs to move in the actual product function, such as the air guide vane sliding bracket or filter pull bracket inside the air conditioner cabinet. The preset track is the groove or guide rail on the inner wall of the shell used to guide the movement of the component. Subsequently, the server controls the robotic arm to apply driving force to the sliding component, causing it to slide along a preset track. The end of the robotic arm is equipped with a special gripping tool or push rod, which can accurately simulate the pushing and pulling actions during manual assembly or user operation, so as to simulate the movement conditions of the internal components in actual use. During the sliding process, the server will obtain the corresponding resistance value of the sliding drive in real time. The resistance value can be obtained by directly measuring it through a six-axis force sensor installed at the end of the robotic arm, or by calculating it by monitoring the current of the drive motor. The resistance value reflects the friction and interference between the sliding component and the track in real time. For a qualified injection molded housing, the internal track dimensions are stable, the sliding process is smooth, and the resistance value is usually maintained in a low and stable range. When the resistance value is greater than the preset resistance threshold, the server will determine that the detection attribute of the corresponding injection molded shell is a defect attribute. The preset resistance threshold is an upper limit set based on the sliding resistance statistics of a large number of qualified products. When the detected resistance value exceeds the preset resistance threshold, it indicates that the sliding component has encountered abnormal resistance during the movement. This abnormal resistance usually originates from the minute deformation of the internal structure of the injection-molded housing. For example, if the sidewall of the housing shrinks inward during the injection molding cooling process, it will cause the spacing of the preset tracks to narrow, thereby squeezing the sliding components and increasing friction. Or, if the housing undergoes irreversible plastic deformation in the previous simulated load test, causing the track surface to warp or step, it will also cause jamming during sliding. This embodiment can accurately identify products that are outwardly undamaged, have acceptable strength, but have internal dimensional defects, ensuring that the final product not only has a complete structure but also provides a good assembly and user experience.

[0053] Furthermore, the aforementioned "obtaining the first real-time stress value from the stress-sensing bolts located at designated holes in the injection-molded housing" also includes the following steps: The initial pressure data of the passive wireless monitoring bolt in its natural state is obtained through a wireless reading terminal and used as the first real-time stress value.

[0054] For example, in this embodiment, the stress-sensing bolt is a passive wireless monitoring bolt. This passive wireless monitoring bolt is an intelligent fastener that integrates a sensing unit and a wireless communication unit. It can reduce the constraints of battery replacement and cable connection, so that the stress-sensing bolt is not only structurally the same as an ordinary bolt, but can also be easily installed in various designated holes of the injection-molded shell, improving the adaptability of the detection system. The server will obtain the initial pressure data of the passive wireless monitoring bolt in its natural state through the wireless reading terminal as the first real-time stress value. The wireless reading terminal is usually placed in a fixed position at the inspection station or integrated into the end of the robotic arm, and its transmission frequency covers the entire inspection area. During the preparation stage before the start of the testing process, the injection-molded shell is in a natural state without external force. At this time, the wireless reading terminal sends radio frequency query commands to the passive wireless monitoring bolts at each designated hole. The sensing unit inside the stress-sensing bolt responds to the command and converts the current axial preload into a wireless signal to feed back to the wireless reading terminal. The server will record the data collected at this moment as the initial pressure data, i.e. the first real-time stress value. Since the injection molded shell may generate internal residual stress during the injection cooling process, or the shell may be slightly bumped during the flow process, the stress at the hole in its natural state is not zero. Therefore, determining the first real-time stress value can ensure that when applying simulated loads later, the stress difference can be calculated starting from the first real-time stress value, thus ensuring the accuracy of the test results.

[0055] According to the present invention, by simultaneously acquiring the initial image of the injection-molded shell and the first real-time stress value of the stress-sensing bolt at the specified hole position, and then combining it with a preset stress simulation strategy, the robotic arm precisely applies the simulated load, and simultaneously acquires the real-time deformation image of the target detection surface and the second real-time stress value of the stress-sensing bolt, thereby realizing the fusion detection of deformation visual information and stress perception information. This enables the comprehensive and accurate determination of the deformation detection value of the target detection surface. Furthermore, based on the deformation detection value and the preset defect judgment rule, the corresponding detection attributes of the injection-molded shell can be objectively and stably determined, effectively improving the accuracy, reliability, and automation of the structural strength and deformation defect detection of the injection-molded shell, reducing human inspection errors and subjective interference, and ensuring the product quality of the injection-molded shell.

[0056] Another embodiment of the present invention provides a machine vision-based defect detection device for injection-molded casings of cabinet air conditioners. Figure 4 Its corresponding device block diagram includes: The data acquisition module is used to acquire an initial image of the injection-molded housing and the first real-time stress value from the stress-sensing bolts located at designated holes in the injection-molded housing. The load control module is used to control the robotic arm to apply simulated loads to the target detection surface of the injection-molded shell based on a preset force simulation strategy. The real-time monitoring module is used to acquire real-time deformation images of the target detection surface and the second real-time stress value of the stress-sensing bolt during the application of the simulated load. The feature analysis module is used to determine the deformation detection value of the target detection surface based on the initial image, the real-time deformation image, the first real-time stress value, and the second real-time stress value. The defect determination module is used to determine the detection attributes of the corresponding injection molded shell based on the matching result of the deformation detection value and the preset defect determination rule.

[0057] In the specification provided herein, the algorithms and displays are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used with the examples of this invention. The required structure for constructing such systems is apparent from the above description. Furthermore, this invention is not directed to any particular programming language. It should be understood that the contents of the invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing preferred embodiments of the invention.

[0058] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0059] Similarly, it should be understood that, in order to streamline this disclosure and aid in understanding one or more of the various aspects of the invention, in the above description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof.

[0060] Those skilled in the art will understand that modules, units, or components of the devices disclosed in the examples herein can be arranged in the devices described in this embodiment, or alternatively, can be located in one or more devices different from the devices in this example. The modules in the foregoing examples can be combined into a single module or, in addition, can be divided into multiple sub-modules.

[0061] Those skilled in the art will understand that the modules in the device of the embodiment can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiment can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components.

[0062] Furthermore, those skilled in the art will understand that although some embodiments described herein include certain features included in other embodiments but not others, combinations of features from different embodiments are meant to be within the scope of the invention and form different embodiments.

[0063] Furthermore, some of the embodiments described herein are methods or combinations of method elements that can be implemented by a processor of a computer system or by other means of performing the functions. Therefore, a processor having the necessary instructions for implementing the methods or method elements forms means for implementing the methods or method elements. Furthermore, the elements described herein in the apparatus embodiments are examples of means for implementing the functions performed by the objective elements for carrying out the invention.

[0064] As used herein, unless otherwise specified, the use of ordinal numbers such as “first,” “second,” “third,” etc., to describe ordinary objects merely indicates different instances of similar objects and is not intended to imply that the objects being described must have a given order in time, space, ordering, or any other manner.

[0065] Although the invention has been described with reference to a limited number of embodiments, those skilled in the art will understand from the foregoing description that other embodiments are conceivable within the scope of the invention described herein. Furthermore, it should be noted that the language used in this specification has been chosen primarily for readability and edibility purposes, and not for the purpose of explaining or limiting the subject matter of the invention.

Claims

1. A machine vision-based method for detecting defects in the injection-molded casing of a cabinet air conditioner, characterized in that, include: Acquire an initial image of the injection-molded housing and the first real-time stress value from the stress-sensing bolts located at designated holes in the injection-molded housing; A preset stress simulation strategy is generated, and the robotic arm is controlled to apply simulated loads to the target detection surface of the injection-molded shell based on the preset stress simulation strategy; During the application of the simulated load, a real-time deformation image of the target detection surface and a second real-time stress value of the stress-sensing bolt are acquired. Based on the initial image, the real-time deformation image, the first real-time stress value, and the second real-time stress value, the deformation detection value of the target detection surface is determined; Based on the matching result between the deformation detection value and the preset defect judgment rule, the detection attributes corresponding to the injection molded shell are determined.

2. The method according to claim 1, characterized in that, The process of acquiring an initial image of the injection-molded housing and a first real-time stress value from stress-sensing bolts located at designated holes in the injection-molded housing further includes: Obtain the 3D design model of the injection molded shell, and extract the hole positions of all shell holes located in the injection molded shell based on the 3D design model; Retrieve a preset filter list, and in response to any of the hole positions being in the preset filter list, determine the housing hole position corresponding to that hole position function as the designated hole position; Based on the 3D design model, all remaining shell holes are determined in a linked manner, and based on the determination results, all shell holes with deformation linkage relationships are determined as designated holes.

3. The method according to claim 2, characterized in that, The step of determining all remaining outer shell openings in a coordinated manner based on the three-dimensional design model includes: Determine all structural surfaces of the injection-molded housing based on the 3D design model; When it is determined that any two outer shell holes are located on the same structural surface, the structural rigidity strength of the structural surface is obtained; When it is determined that the structural rigidity is less than a preset first rigidity threshold, the two shell holes are determined to have a deformation linkage relationship. When it is determined that the structural rigidity is greater than or equal to the preset first rigidity threshold and less than the preset second rigidity threshold, the hole distance between the two shell holes is obtained. When the distance between the holes is determined to be less than a preset distance threshold, any two outer shell holes are determined to have a deformation linkage relationship; or... Determine the different structural surfaces where any two outer shell holes are located in adjacent positions, and obtain the surface connection method corresponding to the different structural surfaces; When the surface connection method is determined to be rigid connection, the rigid connector corresponding to the rigid connection is determined according to the three-dimensional design model, and the connection rigidity strength of the rigid connector is obtained. When the connection rigidity is determined to be less than the preset third rigidity threshold, the hole distance between any two corresponding outer shell holes is obtained; When the distance between holes is determined to be less than a preset distance threshold, any two holes in the outer shell are determined to have a deformation linkage relationship.

4. The method according to claim 2, characterized in that, The strategy for generating the preset force simulation includes: Identify the structural features of the injection-molded housing, wherein the structural features include a first structural surface with specified holes, a second structural surface without specified holes, and surface connection relationships between different structural surfaces in adjacent positions; Based on the structural features, the corresponding single-sided force application location and force value are planned for each structural surface to generate a single-sided test strategy. Based on the surface connection relationship between different structural surfaces in adjacent positions, plan the corresponding synchronous force application positions and force values, and generate a fusion test strategy; The single-sided testing strategy is combined with the fusion testing strategy to form the preset force simulation strategy.

5. The method according to claim 4, characterized in that, The step of planning the application location and value of force for each structural surface according to the structural features, and generating a single-surface testing strategy, includes: In response to the structural surface being the first structural surface, the force application position corresponding to the single-sided force application is determined as the intersection of the center axis of the hole at the designated hole position located on the first structural surface and the first structural surface. The force value for the corresponding single-sided force application is calculated based on the product of the rated tightening force at the designated hole position and a preset detection coefficient; or... In response to the structural surface being the second structural surface, the applied force value of the corresponding single-surface applied force is calculated based on the product of the rated bearing value of the corresponding second structural surface and the preset detection coefficient, and the surface size of the corresponding second structural surface is compared with the preset size threshold. If the size of the response surface is smaller than a preset size threshold, the second structural surface is divided into an array, and each array point is determined as the force application position for the corresponding single-sided force application. Otherwise, the center point of the corresponding second structural surface is determined as the force application position for the corresponding single-sided force application.

6. The method according to claim 4, characterized in that, The step of planning the corresponding synchronous force application positions and force values ​​based on the surface connection relationship between different structural surfaces in adjacent positions, and generating a fusion test strategy, includes: The response is that the surface connection method of different structural surfaces in adjacent positions is a rigid connection, and the specified hole position is determined in the different structural surfaces; When it is determined that the specified holes located on different structural surfaces have a deformation linkage relationship, the different structural surfaces are determined as a strongly correlated set; otherwise, they are determined as a weakly correlated set. Based on the strongly correlated set, the specified hole positions with deformation linkage are determined as the corresponding synchronous force application positions, and the corresponding synchronous force application value is calculated based on the product of the rated tightening force of the corresponding specified hole position and the preset detection coefficient; or, Based on the weak correlation set, the surface connection line corresponding to the rigid connection is determined as the force application position of the corresponding synchronous force application, and the force value of the corresponding synchronous force application is calculated by multiplying the minimum rated bearing value of the corresponding different structural surfaces with the preset detection coefficient.

7. The method according to claim 4, characterized in that, The method of controlling the robotic arm to apply simulated loads to the target detection surface of the injection-molded shell based on a preset force simulation strategy includes: According to the preset force simulation strategy, when the robotic arm applies a simulated load to the structural surface identified as the target detection surface, the control fixture unit fixes and clamps all structural surfaces of the injection-molded shell that are not related to the target detection surface.

8. The method according to claim 1, characterized in that, Determining the deformation detection value of the target detection surface based on the initial image, the real-time deformation image, the first real-time stress value, and the second real-time stress value includes: The initial image and the real-time deformed image are compared for features to extract the image deformation. The stress change is calculated based on the stress difference between the second real-time stress value and the first real-time stress value. The image deformation and the stress change are fused to generate the deformation detection value.

9. The method according to claim 8, characterized in that, The step of comparing the features of the initial image and the real-time deformed image to extract image deformation includes: When the target detection surface contains a specified hole, the hole contour corresponding to the specified hole in the initial image and the real-time deformation image are identified respectively; Calculate the image deformation of the hole contour in the real-time deformed image relative to the corresponding hole contour in the initial image.

10. The method according to claim 8, characterized in that, The determination of the detection attributes corresponding to the injection-molded shell based on the matching result of the deformation detection value and the preset defect judgment rule includes: The image deformation value in the deformation detection value is compared with a preset image deformation threshold; The stress change in the deformation detection value is compared with a preset stress change threshold. When it is determined that the image deformation is greater than the image deformation threshold, and / or the stress change is greater than the stress change threshold, the resistance change of the injection molded shell is determined according to a preset resistance simulation strategy, and the detection attribute corresponding to the injection molded shell is determined based on the resistance change.

11. The method according to claim 10, characterized in that, The step of determining the resistance change of the injection-molded shell according to a preset resistance simulation strategy, and determining the corresponding detection attributes of the injection-molded shell based on the resistance change, includes: A sliding component is assembled inside the injection-molded housing, and a preset track is defined within the housing. The robotic arm is controlled to apply a driving force to the slidable component, causing it to slide along a preset track; If the resistance value of the corresponding drive sliding is obtained, and the resistance value is greater than the preset resistance threshold, then the detection attribute of the corresponding injection molded shell is determined to be a defect attribute.

12. The method according to claim 1, characterized in that, The stress-sensing bolt is a passive wireless monitoring bolt; The step of obtaining the first real-time stress value from the stress-sensing bolts located at designated holes in the injection-molded housing includes: The initial pressure data of the passive wireless monitoring bolt in its natural state is obtained through a wireless reading terminal and used as the first real-time stress value.

13. A machine vision-based defect detection device for cabinet air conditioner injection molded shells, characterized in that, include: The data acquisition module is used to acquire an initial image of the injection-molded housing and the first real-time stress value from the stress-sensing bolts located at designated holes in the injection-molded housing. The load control module is used to control the robotic arm to apply simulated loads to the target detection surface of the injection-molded shell based on a preset force simulation strategy. The real-time monitoring module is used to acquire real-time deformation images of the target detection surface and the second real-time stress value of the stress-sensing bolt during the application of the simulated load. The feature analysis module is used to determine the deformation detection value of the target detection surface based on the initial image, the real-time deformation image, the first real-time stress value, and the second real-time stress value. The defect determination module is used to determine the detection attributes of the corresponding injection molded shell based on the matching result of the deformation detection value and the preset defect determination rule.

Images