Aircraft component coating spraying quality detection method

Abstract

The invention relates to the field of spraying quality detection, in particular to an aircraft component coating spraying quality detection method, which comprises the following steps of: determining the category of concern points according to neighborhood vector difference degree and maximum reference difference degree, and determining whether to adopt an adaptive division mode or not based on the proportion of the second category of concern points and the proportion of the area of an analysis region; in the adaptive division mode, determining seed points according to the curvature processing values, and obtaining a primary selection sequence of the seed points based on the curvature processing values; according to the curvature processing value corresponding to the comparison condition, determining whether sequence adjustment is carried out or not so as to obtain a final sequence; determining associated points in the target area corresponding to the growth of the seed points based on the curvature processing difference value and the symbol similarity index; determining whether to perform merging processing according to the area of the region; the combination processing mode is determined according to the curvature similarity parameter difference value and the spraying distance difference index. According to the invention, the efficiency of spraying quality detection is improved.

Description

Technical Field

[0001] This invention relates to the field of coating quality inspection, and more particularly to a method for inspecting the coating quality of aircraft components. Background Technology

[0002] With the gradual maturation of automated aircraft painting technology, more and more aircraft parts are being painted automatically by robots. However, the quality of the coating is still mostly inspected manually by visual inspection. To improve inspection efficiency and accuracy, laser light sources are gradually being used to inspect the coating quality. However, the inspection of complex aircraft parts is often difficult to complete in one go. Therefore, it is necessary to set up areas according to the structure of the actual object being inspected. How to improve the setting effect of the inspection area to improve inspection efficiency is a problem of great concern to those skilled in the art.

[0003] Chinese Patent Publication No. CN108212719A discloses a method for automated spraying area division and robot positioning planning for aircraft. The method includes: firstly, dividing the aircraft into three parts—nose, fuselage, and wings (including tail)—based on the surface morphology, and abstracting these areas into basic shapes such as frustums, cylinders, and planes; then, establishing single-pass spraying area units (referred to as "area units") with regular geometric shapes based on the maximum single intersection of the robot's available workspace and the aforementioned geometric surfaces; finally, using the area units to complete the area arrangement of the aircraft's outer surface, thus realizing robot positioning planning. It is evident that the above technical solution, by employing maximum inscribed area units, ensures fewer robot movements and improves coating efficiency. However, it suffers from the following problems: relying solely on the simple shapes of aircraft components for area acquisition fails to meet the requirements for efficient and flexible detection of the structural characteristics of aircraft components. Summary of the Invention

[0004] Therefore, the present invention provides a method for inspecting the coating quality of aircraft components, which overcomes the problem that the existing technology only obtains regions based on the simple shape of the aircraft components, and the region acquisition method cannot meet the inspection efficiency and inspection flexibility of the structural characteristics of the aircraft components.

[0005] To achieve the above objectives, the present invention provides a method for inspecting the coating quality of aircraft components, comprising: The categories of concerns are determined based on the neighborhood vector difference degree and the maximum reference difference degree, and whether to adopt an adaptive partitioning method is determined based on the proportion of the second category of concerns and the proportion of the analysis area. Under the adaptive partitioning method, first-class and second-class interest points are selected according to the curvature processing value to obtain the first seed point and the second seed point. The seed points are arranged in descending order of curvature processing value to obtain the initial selection sequence. Determine whether to adjust the order of the initial sequence based on the curvature processing value corresponding to the comparison conditions in order to obtain the final sequence; Region growing is performed sequentially on various sub-points in the final sequence to obtain several target regions including seed points and corresponding associated points. The associated points are selected based on curvature processing difference and sign similarity index. For target areas whose area is smaller than the preset area, it is determined to merge them. In the merging process, the state of adjacent target areas is determined based on the difference value of curvature similarity parameters and the difference index of spraying distance, and the merging process method is determined to be curvature similarity merging or spraying similarity merging based on the state of adjacent target areas.

[0006] Furthermore, regions are constructed for the two types of concerns of the target component model to obtain the model region with the smallest area that can include all two types of concerns, which is denoted as the analysis region; If the proportion of Category II concerns is greater than the preset proportion of Category II concerns, or the proportion of the analysis area is less than or equal to the preset proportion of the analysis area, then the adaptive division method shall be adopted. If the proportion of Category II concerns is less than or equal to the preset proportion of Category II concerns and the proportion of the analysis area is greater than the preset proportion of the analysis area, then the benchmark division method shall be adopted.

[0007] Furthermore, points of interest whose neighborhood vector difference is greater than the preset neighborhood vector difference or whose maximum reference difference is greater than the preset maximum reference difference are denoted as a type of point of interest. Points of interest whose neighborhood vector difference is less than or equal to the preset neighborhood vector difference and whose maximum reference difference is less than or equal to the preset maximum reference difference are denoted as second-class points of interest.

[0008] Furthermore, the adaptation methods include: The curvature reference values ​​for each point of interest are filtered and normalized to obtain the curvature processed values. Points of interest whose curvature processing value is greater than the preset maximum curvature processing value are denoted as the first seed point; Points of interest with curvature processing values ​​less than or equal to the preset minimum curvature processing value are designated as second seed points.

[0009] Furthermore, for comparison results where the curvature processing value comparison degree is greater than the preset curvature processing value comparison degree, it is determined to adjust the order of the initial selection sequence based on the neighborhood similar reference values.

[0010] Furthermore, the order adjustment of the initial selection sequence includes: Neighborhood analysis is performed on various sub-points in descending order of curvature processing values; Neighborhood analysis for a single seed point includes detecting the difference between its neighboring reference values ​​and those of the same type that are adjacent seed points in the sequence order. If the difference between the neighboring reference values ​​is greater than or equal to 0, then no order change is required. If the difference between the reference values ​​of the same type in the neighborhood is less than 0, then the two seed points are swapped sequentially.

[0011] Furthermore, when performing region growth for a single seed point, the focus point whose curvature processing difference is less than the preset curvature processing difference and whose symbol similarity index is greater than the preset symbol similarity index is recorded as the associated point, and the region of the target component model including the seed point and the corresponding associated point with the minimum area is recorded as the target region. If the order is adjusted, the initial selection sequence after the order adjustment is completed is recorded as the final sequence; If no order adjustment is performed, the initial sequence is recorded as the final sequence.

[0012] Furthermore, for adjacent target areas where the curvature similarity parameter difference value is less than the preset curvature similarity parameter difference value and the spraying distance difference index is less than the preset spraying distance difference index, the merging processing method is determined to be curvature similarity merging, and the adjacent area with the smallest curvature similarity parameter is selected for merging.

[0013] Furthermore, for adjacent target areas where the curvature similarity parameter difference value is greater than or equal to the preset curvature similarity parameter difference value or the spraying distance difference index is greater than or equal to the preset spraying distance difference index, the merging processing method is determined to be spraying similarity merging. The trajectory reference direction of the adjacent areas is detected. If the number of adjacent areas with the same trajectory reference direction as the target area is 1, the target area is merged into the adjacent area with the same trajectory reference direction.

[0014] Furthermore, if the number of adjacent regions with the same trajectory reference direction as the target region is not 1, the target region is merged into the adjacent region with the highest similarity in spraying output parameters.

[0015] Compared with the prior art, the beneficial effect of the present invention is that the present invention reflects the degree of difference between the point of interest and each of its neighboring points of interest in terms of normal vector through the neighborhood vector difference degree, and reflects the maximum degree of difference between the normal vector of the point of interest and the points of interest in the neighborhood. Based on the neighborhood vector difference degree and the maximum reference difference degree, the category of the point of interest is determined, which realizes the accurate division of the point of interest and provides accurate data support for the subsequent determination of the division method.

[0016] Furthermore, this invention determines whether to adopt an adaptive partitioning method by using the proportion of the two types of concerns and the proportion of the analysis area. The proportion of the two types of concerns reflects the numerical distribution of the two types of concerns within the entire model area of ​​the target component, while the proportion of the analysis area reflects the proportion of the smallest model area, including all two types of concerns, to the target component's model area. This allows for the use of an adaptive partitioning method to analyze the target component when the proportion of the two types of concerns is greater than a preset proportion or the proportion of the analysis area is less than or equal to a preset proportion of the analysis area. This avoids misjudgments caused by the poor applicability of a single analysis method and further improves the flexibility of the processing method.

[0017] Furthermore, this invention uses filtering and normalization to obtain curvature processing values ​​that reflect the relative differences in curvature between the first and second types of interest points within the neighborhood of the corresponding target component. First-type interest points with curvature processing values ​​greater than a preset maximum curvature processing value are designated as first seed points, and second-type interest points with curvature processing values ​​less than or equal to a preset minimum curvature processing value are designated as second seed points. This achieves precise selection of seed points, providing an accurate foundation for the subsequent construction of the target region and further improving the efficiency of target region construction.

[0018] Furthermore, this invention uses neighborhood similarity reference values ​​to reflect the number of interest points of the same category as the seed point in the neighborhood. By adjusting the seed point numbering based on the neighborhood similarity reference values ​​of seed points with adjacent numbering, the order of seed point growth is determined, providing accurate data support for the subsequent construction of the target region and further improving the flexibility of seed point order analysis.

[0019] Furthermore, this invention uses curvature processing difference to reflect the degree of difference in curvature processing values ​​between the seed point and other points of interest, and sign similarity index to reflect whether the curvature signs of the seed point and other points of interest are the same. By using curvature processing difference and sign similarity index, associated points are determined, achieving accurate determination of associated points. Based on the seed point and the corresponding associated points, the corresponding target region is determined, improving the effect of target region construction.

[0020] Furthermore, this invention determines the state of adjacent target areas by using the curvature similarity parameter difference value and the spraying distance difference index, and determines the merging processing method accordingly. The curvature similarity parameter difference value reflects the degree of similarity in curvature between the area to be merged and the adjacent areas, and the spraying distance difference index reflects the degree of difference in the corresponding spraying distance of each target area when spraying the target component. This determines the merging processing method for target areas with an area smaller than the preset area, avoiding the situation where the area is too small or the number is too large, which would increase the difficulty of analysis and further improve the efficiency of spraying detection. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the aircraft component coating quality inspection method of the present invention; Figure 2 This is a flowchart illustrating how the present invention determines the category of interest based on the neighborhood vector difference and the maximum reference difference. Figure 3 This is a flowchart illustrating the process of determining whether to adopt an adaptive partitioning method based on the proportion of two types of concerns and the proportion of the analysis area in this invention. Figure 4 This is a flowchart illustrating the process of determining whether to perform a merging process on a target region based on the region area, according to the present invention. Detailed Implementation

[0022] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0023] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0024] Please see Figures 1 to 4 As shown, the present invention provides a method for inspecting the coating quality of aircraft components, comprising: The categories of concerns are determined based on the neighborhood vector difference degree and the maximum reference difference degree, and whether to adopt an adaptive partitioning method is determined based on the proportion of the second category of concerns and the proportion of the analysis area. Under the adaptive partitioning method, first-class and second-class interest points are selected according to the curvature processing value to obtain the first seed point and the second seed point. The seed points are arranged in descending order of curvature processing value to obtain the initial selection sequence. Determine whether to adjust the order of the initial sequence based on the curvature processing value corresponding to the comparison conditions in order to obtain the final sequence; Region growing is performed sequentially on various sub-points in the final sequence to obtain several target regions including seed points and corresponding associated points. The associated points are selected based on curvature processing difference and sign similarity index. For target areas whose area is smaller than the preset area, it is determined to merge them. In the merging process, the state of adjacent target areas is determined based on the difference value of curvature similarity parameters and the difference index of spraying distance, and the merging process method is determined to be curvature similarity merging or spraying similarity merging based on the state of adjacent target areas.

[0025] In this invention, the target component is an aircraft component, including but not limited to the fuselage, wings, tail, and cabin doors. When performing quality inspection on the target component, a three-dimensional model of the target component is first constructed, denoted as the target component model. Then, the target component model is triangulated to obtain a triangular patch mesh. The maximum side length of the triangular patch mesh is set to not exceed 1 / 3 of the expected scan point distance. The point of interest is the point that constitutes the triangular patch mesh, and the normal vector of the point of interest is a unit vector. The expected scan point distance is the distance between two adjacent sampling points when the painting robot plans its path. In this embodiment of the invention, the expected scan point distance is 5mm.

[0026] Specifically, points of interest whose neighborhood vector difference is greater than the preset neighborhood vector difference or whose maximum reference difference is greater than the preset maximum reference difference are categorized as a type of point of interest. Points of interest whose neighborhood vector difference is less than or equal to the preset neighborhood vector difference and whose maximum reference difference is less than or equal to the preset maximum reference difference are denoted as second-class points of interest.

[0027] For a single point of interest, the corresponding neighborhood is identified by obtaining other points of interest within the neighborhood distance of that point of interest, and recording the circular area centered on that point of interest and with the neighborhood distance as the radius as the neighborhood of that point of interest. The neighborhood distance is the average side length of each triangular facet mesh of the target component.

[0028] For a single point of interest, the neighborhood vector difference is the variance of the difference values ​​of each neighborhood vector. The neighborhood vector difference is the minimum angle between the point of interest and the normal vector of any point of interest in the neighborhood. The maximum reference difference is the maximum value of the neighborhood vector difference. How to obtain the minimum angle between two unit vectors is a topic that is already known to those skilled in the art and will not be elaborated here.

[0029] For a single point of interest, its normal vector is calculated as follows:

[0030] in, This focus is on the first The area of ​​a triangular mesh. For the first The normal vector of a triangular mesh formed by this point of interest. The number of triangular meshes formed by this point of interest. The summation index is used to traverse the triangular patch mesh formed by the point of interest.

[0031] Regarding the first The normal vector of the triangular mesh formed by the points of interest. The calculation method is as follows:

[0032] in, , , for and The edge vectors formed for and The edge vectors formed The three-dimensional coordinates of this point of interest. and To form the third with this concern The three-dimensional coordinates of the other two points of interest in the triangular mesh.

[0033] In this embodiment of the invention, the preset neighborhood vector difference degree is set to 0.5°. It can be understood that the neighborhood vector difference degree reflects the degree of difference in angle between the normal vector of the point of interest and the normal vectors of other points of interest in the neighborhood. The greater the difference in angle between the normal vector of the point of interest and the normal vectors of other points of interest in the corresponding neighborhood, the worse the flatness of the corresponding target part area. The more sensitive the user is to the flatness of the position corresponding to the point of interest, the smaller the value of the preset neighborhood vector difference degree.

[0034] In this embodiment of the invention, the preset maximum reference difference is set to 1.5°. It can be understood that the maximum reference difference reflects the maximum degree of change in the angle between the normal vector of the point of interest and other points of interest in the neighborhood. The larger the maximum reference difference, the larger the angle between the normal vector of the point of interest in the neighborhood and the normal vector of the point of interest, the sharper the actual area of ​​the target component, and the greater the difficulty of spraying analysis of the point of interest. The more sensitive the user is to the difficulty of spraying analysis of the point of interest, the larger the preset maximum reference difference value.

[0035] Specifically, regions are constructed for the two types of concerns of the target component model to obtain the model region with the smallest area that can include all two types of concerns, which is denoted as the analysis region; If the proportion of Category II concerns is greater than the preset proportion of Category II concerns, or the proportion of the analysis area is less than or equal to the preset proportion of the analysis area, then the adaptive division method shall be adopted. If the proportion of Category II concerns is less than or equal to the preset proportion of Category II concerns and the proportion of the analysis area is greater than the preset proportion of the analysis area, then the benchmark division method shall be adopted.

[0036] The percentage of Category II concerns = number of Category II concerns / total number of concerns; the percentage of the analysis area = analysis area / surface area of ​​the target component model.

[0037] In this embodiment of the invention, the preset value of the proportion of the second type of concern is 0.7. It can be understood that the proportion of the second type of concern reflects the degree of the number of the second type of concern in the entire target component model area. The larger the proportion of the second type of concern, the more the number of the second type of concern in the target component model. The more flat the actual area of ​​the target component corresponding to the second type of concern, the larger the proportion of the flat actual area of ​​the target component, the lower the analysis difficulty of the target component, and the higher the user's tolerance for the analysis difficulty of the target component. Therefore, the preset value of the proportion of the second type of concern is higher.

[0038] In this embodiment of the invention, the preset analysis area ratio is 0.8. It can be understood that the analysis area ratio reflects the coverage of the analysis area in the target component model area. The larger the analysis area ratio, the larger the corresponding analysis area, the larger the actual area of ​​the relatively flat target component, the lower the analysis difficulty of the target component model, and the more sensitive the user is to the analysis difficulty of the target component model, the higher the preset analysis area ratio.

[0039] Specifically, the adaptation methods include: The curvature reference values ​​for each point of interest are filtered and normalized to obtain the curvature processed values. Points of interest whose curvature processing value is greater than the preset maximum curvature processing value are denoted as the first seed point; Points of interest with curvature processing values ​​less than or equal to the preset minimum curvature processing value are designated as second seed points.

[0040] For a single point of interest, the curvature processing value is calculated as follows:

[0041]

[0042] in, This is the normal vector corresponding to the point of interest. For the neighborhood corresponding to this point of interest, excluding the point of interest itself, the [number]th [item] The normal vector corresponding to each point of interest. The summation index is used to iterate through other points of interest within the neighborhood of the current point of interest. These are the three-dimensional coordinates of the location corresponding to this point of interest. For the neighborhood corresponding to this point of interest, excluding the point of interest itself, the [number]th [item] The three-dimensional coordinates of the location corresponding to each point of interest. The number of neighborhood attention points. It is an inverse cosine function. This is a clamping function. Focus on the entire target component model The 5th percentile, Focus on the entire target component model The 95th percentile, This is a curvature reference value. This is the curvature processing value.

[0043] The curvature reference value of the point of interest is filtered to eliminate noise or abnormal fluctuations. The filtering methods used can be, but are not limited to, mean filtering, median filtering, and Gaussian filtering.

[0044] When arranging seed points in descending order of curvature processing values ​​to obtain the initial selection sequence, if there are seed points with the same curvature processing value, the seed points with the same curvature processing value are randomly sorted.

[0045] In this embodiment of the invention, the preset maximum curvature processing value is 0.5, and the preset minimum curvature processing value is 0.1. It can be understood that the curvature processing value reflects the degree of geometric transformation of a type I and a type II point of interest in their corresponding neighborhood. The smaller the curvature processing value, the slower the geometric transformation of the neighborhood corresponding to the point of interest. The more sensitive the user is to the geometric transformation of the neighborhood corresponding to the point of interest, the higher the preset maximum curvature processing value and the lower the preset minimum curvature processing value.

[0046] Specifically, for comparison results where the curvature processing value comparison degree is greater than the preset curvature processing value comparison degree, the order adjustment of the initial selection sequence is determined based on the neighborhood similar reference values.

[0047] The curvature processing value comparison degree is the variance of the curvature processing values ​​of various sub-points. In this embodiment of the invention, the preset curvature processing value comparison degree is 0.35. It can be understood that the curvature processing value comparison degree reflects the degree of difference in the curvature processing values ​​of various sub-points. The larger the curvature processing value comparison degree, the greater the difference between the curvature processing values ​​of various sub-points, the greater the difference in the difficulty of constructing regions for various sub-points, and the more sensitive the user is to the difference in the difficulty of constructing seed point regions. Therefore, the preset curvature processing value comparison degree is lower.

[0048] Specifically, the order adjustment of the initial selection sequence includes: Neighborhood analysis is performed on various sub-points in descending order of curvature processing values; Neighborhood analysis for a single seed point includes detecting the difference between its neighboring reference values ​​and those of the same type that are adjacent seed points in the sequence order. If the difference between the neighboring reference values ​​is greater than or equal to 0, then no order change is required. If the difference between the reference values ​​of the same type in the neighborhood is less than 0, then the two seed points are swapped sequentially.

[0049] For the initial selection sequence, the sequence order is represented in the form of numbers. Various sub-points are numbered, and the numbers correspond to the order of the seed points in the initial selection sequence. For example, the seed point numbered 1 corresponds to the seed point with the largest curvature processing value in the initial selection sequence, and the seed point numbered 2 is located one position after the seed point numbered 1.

[0050] For a single seed point, its neighborhood reference value is the number of interest points in the corresponding neighborhood that have the same interest category as the seed point. For any seed point in the initial selection sequence and the next seed point with the corresponding number, the difference between their corresponding neighborhood reference values ​​is equal to the difference between the neighboring reference values ​​of the seed point with the smaller number and the neighboring reference values ​​of the seed point with the larger number. If the difference between the neighboring reference values ​​is less than 0, the order of the seed points in the two adjacent initial selection sequences is swapped, and their corresponding numbers are also swapped. If the order of any seed point has been changed, neighborhood analysis is not required.

[0051] Specifically, when performing region growth for a single seed point, the points of interest whose curvature processing difference is less than the preset curvature processing difference and whose symbol similarity index is greater than the preset symbol similarity index are recorded as associated points, and the region of the target component model including the seed point and the corresponding associated points with the minimum area is recorded as the target region. If the order is adjusted, the initial selection sequence after the order adjustment is completed is recorded as the final sequence; If no order adjustment is performed, the initial sequence is recorded as the final sequence.

[0052] Region growing is performed on the seed points of the final sequence to construct the target region corresponding to each seed point. Seed points are grown sequentially from smallest to largest according to their corresponding numerical numbers. When growing a single seed point, the nearest unassigned point of interest is detected. The curvature processing difference is calculated as: curvature processing value of the seed point minus curvature processing value of the point of interest. The symbolic similarity index is determined by checking if the curvature signs of the seed point and the point of interest are the same. If they are the same, the symbolic similarity index is 1; otherwise, it is 0. In this embodiment, the symbolic similarity index is preset to 0. The curvature sign is calculated for any point of interest as follows:

[0053] in, Let this be the normal vector of the point of interest. The displacement vector is obtained by subtracting the average of the three-dimensional coordinates of all neighboring points of interest from the three-dimensional coordinates of the point of interest. Functions with positive and negative signs This is the curvature symbol.

[0054] In this embodiment of the invention, the preset curvature processing difference is set to 0.05. It can be understood that the curvature processing difference reflects the degree of difference in curvature processing values ​​between the seed point for region growing and the nearest point of interest that is not recorded as an associated point. The smaller the curvature processing difference, the closer the curvature processing values ​​of the seed point and the point of interest are, and the smaller the degree of difference between the seed point and the point of interest, the greater the possibility that the point of interest will be recorded as the corresponding associated point when performing region growing on the seed point. The more sensitive the managers are to the degree of difference between the seed point and the associated point in the target area, the lower the value of the preset curvature processing difference.

[0055] If a single seed point is constructed as an associated point within the target region corresponding to other seed points, then there is no need to perform region growing on that seed point.

[0056] For target regions whose area is smaller than a preset area, a merging process is determined for the target region, and the target region is recorded as a region to be merged. Merging is carried out in order of increasing area according to the region to be merged. For a single region to be merged, the region to be merged is divided into the adjacent target regions with the smallest curvature similarity parameter. For target regions whose area is greater than or equal to the preset area, no merging process is required. In this embodiment of the invention, the preset area is the average area of ​​each region.

[0057] Specifically, for adjacent target areas where the curvature similarity parameter difference value is less than the preset curvature similarity parameter difference value and the spraying distance difference index is less than the preset spraying distance difference index, the merging processing method is determined to be curvature similarity merging, and the adjacent area with the smallest curvature similarity parameter is selected for merging.

[0058] The curvature similarity parameter difference value is the range of each curvature similarity parameter. For the region to be merged and any adjacent target region, the curvature similarity parameter is the absolute value of the difference between the average curvature processing value of each point of interest in the region to be merged and the average curvature processing value of each point of interest in the adjacent target region. The spraying distance difference index is the variance of the spraying distance between the target region and each neighboring region. For a single target region, the spraying distance is the distance between the spray gun nozzle and the actual area of ​​the target component corresponding to the target region when the spray gun performs spraying operation on the actual area of ​​the target component corresponding to the target region in the historical process. The preset spraying distance difference index is the average spraying distance of each target region.

[0059] In this embodiment of the invention, the preset curvature similarity parameter difference value is 0.18. It can be understood that the curvature similarity parameter difference value reflects the degree of difference in curvature between the region to be merged and its adjacent regions. The larger the curvature similarity parameter difference value, the more obvious the curvature difference between the region to be merged and its adjacent regions, and the higher the difficulty of merging. The higher the user's requirements for the difficulty of merging the target region, the lower the preset curvature similarity parameter difference value.

[0060] Specifically, for adjacent target areas where the curvature similarity parameter difference value is greater than or equal to the preset curvature similarity parameter difference value or the spraying distance difference index is greater than or equal to the preset spraying distance difference index, the merging processing method is determined to be spraying similarity merging. The trajectory reference direction of the adjacent areas is detected. If the number of adjacent areas with the same trajectory reference direction as the target area is 1, the target area is merged into the adjacent area with the same trajectory reference direction.

[0061] Specifically, if the number of adjacent regions with the same trajectory reference direction as the target region is not 1, the target region will be merged into the adjacent region with the highest similarity in spray output parameters.

[0062] During spraying, the spraying shape of the spray gun nozzle is elliptical. For a single target area and a single adjacent area, if the difference in trajectory angle is less than the preset trajectory angle difference, then the trajectory reference direction of the target area and the adjacent area is the same. If the difference in trajectory angle is greater than or equal to the preset trajectory angle difference, then the trajectory reference direction of the target area and the adjacent area is different. The trajectory angle difference = the average minimum angle between the spray gun nozzle movement direction and the minor axis direction when spraying the target area - the average minimum angle between the spray gun nozzle movement direction and the minor axis direction when spraying the adjacent area. The average of the minimum angles between the spray gun nozzle and the minor axis direction, which are uniformly sampled 5 times during spraying, is recorded as the minimum angle average.

[0063] In this embodiment of the invention, the preset trajectory angle difference is set to 5°. It can be understood that the trajectory angle difference reflects the degree of difference between the angle between the moving direction of the spray gun nozzle and the minor axis direction when spraying different target areas. The smaller the trajectory angle difference, the closer the spray trajectory reference direction is. The higher the user's sensitivity to the similarity of the trajectory reference direction, the lower the preset trajectory angle difference is.

[0064] For the area to be merged and any adjacent area, the similarity of the spray output parameters is the reciprocal of the absolute value of the difference between the spray output parameters of the area to be merged and the spray output parameters of the adjacent area. For any target area, the spray output parameters are the actual moving speed of the spray gun through the target area corresponding to the target component.

[0065] When performing baseline partitioning, interconnected points of interest of the same type are divided into one target region, and interconnected points of interest of the same type are divided into another target region.

[0066] After the regions are merged, users can control automated robots to automatically plan detection paths based on the geometric features of each target region to perform quality analysis on each target region. The quality analysis method involves using a pulsed laser to excite ultrasonic waves and using a laser interferometer to receive the ultrasonic signals, thereby detecting defects such as delamination, delamination, and porosity under the coating. Example

[0067] For an aircraft wing component, after 3D scanning and triangulation, there are 18,500 points of interest. The maximum side length of the triangular mesh is set to 1.6 mm. The preset neighborhood vector difference is 0.5°, and the preset maximum reference difference is 1.5°.

[0068] After calculating the neighborhood vector difference and maximum reference difference for each point of interest, 4,200 points of interest in the first category and 14,300 points of interest in the second category were obtained. The proportion of points of interest in the second category is 0.773, which is greater than the preset proportion of points of interest in the second category of 0.7, so an adaptive partitioning method is adopted.

[0069] Curvature reference values ​​are calculated for each point of interest, and then processed by median filtering and normalization to obtain the processed curvature values. The preset maximum curvature processing value is 0.5, and the preset minimum curvature processing value is 0.1.

[0070] From the first category of concerns, 186 points with curvature values ​​greater than 0.5 were selected as the first seed points; from the second category of concerns, 204 points with curvature values ​​less than or equal to 0.1 were selected as the second seed points, for a total of 390 seed points. All seed points were then sorted in descending order of curvature values ​​to form an initial selection sequence.

[0071] The variance of the curvature processing values ​​for various sub-points is calculated to be 0.44, which is greater than the preset curvature processing value comparison degree of 0.35. Therefore, sequential adjustment is required.

[0072] Adjacent seed points in the initial sequence are compared with their neighboring reference values. After multiple rounds of swapping, the final sequence is obtained. The preset curvature processing difference is 0.05, and the preset sign similarity index is 0. Region growing is then performed on the seed points in the final sequence. Points of interest with a curvature processing difference less than 0.05 and the same curvature sign are recorded as associated points, forming target regions. Finally, 47 target regions are obtained, with an average area of ​​0.092 m².

[0073] Nine target areas with an area less than 0.092 m² were selected as areas to be merged. Taking the smallest area, A, as an example, its area is 0.031 m², and its adjacent areas are B, C, and D. The difference in curvature similarity parameters and the difference in spraying distance between each adjacent area were calculated. The calculated difference in curvature similarity parameters between area A and areas B, C, and D was 0.14, which is less than the preset value of 0.18; the difference in spraying distance was 0.08, which is less than the preset value of 0.10. Therefore, curvature similarity merging was adopted, and the adjacent area B with the smallest curvature similarity parameter was selected for merging. The remaining areas to be merged were processed sequentially, resulting in 41 target areas remaining after merging.

[0074] After the region is divided, the robot automatically plans the detection path based on the geometric features of each target region. Compared with the traditional fixed grid division method, this method reduces the number of detection regions by about 35%, shortens the detection path length by 28%, and improves the overall detection efficiency by about 40%.

[0075] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A method for inspecting the coating quality of aircraft components, characterized in that, include: The categories of concerns are determined based on the neighborhood vector difference degree and the maximum reference difference degree, and whether to adopt an adaptive partitioning method is determined based on the proportion of the second category of concerns and the proportion of the analysis area. Under the adaptive partitioning method, first-class and second-class interest points are selected according to the curvature processing value to obtain the first seed point and the second seed point. The seed points are arranged in descending order of curvature processing value to obtain the initial selection sequence. Determine whether to adjust the order of the initial sequence based on the curvature processing value corresponding to the comparison conditions in order to obtain the final sequence; Region growing is performed sequentially on various sub-points in the final sequence to obtain several target regions including seed points and corresponding associated points. The associated points are selected based on curvature processing difference and sign similarity index. For target areas whose area is smaller than the preset area, it is determined to merge them. In the merging process, the state of adjacent target areas is determined based on the difference value of curvature similarity parameters and the difference index of spraying distance, and the merging process method is determined to be curvature similarity merging or spraying similarity merging based on the state of adjacent target areas.

2. The method for inspecting the coating quality of aircraft components according to claim 1, characterized in that, For the two types of concerns of the target component model, a region is constructed to obtain the model region with the smallest area that can include all two types of concerns, which is denoted as the analysis region; If the proportion of Category II concerns is greater than the preset proportion of Category II concerns, or the proportion of the analysis area is less than or equal to the preset proportion of the analysis area, then the adaptive division method shall be adopted. If the proportion of Category II concerns is less than or equal to the preset proportion of Category II concerns and the proportion of the analysis area is greater than the preset proportion of the analysis area, then the benchmark division method shall be adopted.

3. The method for inspecting the coating quality of aircraft components according to claim 2, characterized in that, Points of interest whose neighborhood vector difference is greater than the preset neighborhood vector difference or whose maximum reference difference is greater than the preset maximum reference difference are categorized as a type of point of interest. Points of interest whose neighborhood vector difference is less than or equal to the preset neighborhood vector difference and whose maximum reference difference is less than or equal to the preset maximum reference difference are denoted as second-class points of interest.

4. The method for inspecting the coating quality of aircraft components according to claim 3, characterized in that, Adaptive partitioning methods include: The curvature reference values ​​for each point of interest are filtered and normalized to obtain the curvature processed values. Points of interest whose curvature processing value is greater than the preset maximum curvature processing value are denoted as the first seed point; Points of interest with curvature processing values ​​less than or equal to the preset minimum curvature processing value are designated as second seed points.

5. The method for inspecting the coating quality of aircraft components according to claim 4, characterized in that, For comparison results where the curvature processing value comparison degree is greater than the preset curvature processing value comparison degree, the order adjustment of the initial selection sequence is determined based on the neighborhood similar reference values.

6. The method for inspecting the coating quality of aircraft components according to claim 5, characterized in that, Adjusting the order of the initial selection sequence includes: Neighborhood analysis is performed on various sub-points in descending order of curvature processing values; Neighborhood analysis for a single seed point includes detecting the difference between its neighboring reference values ​​and those of the same type that are adjacent seed points in the sequence order. If the difference between the neighboring reference values ​​is greater than or equal to 0, then no order change is required. If the difference between the reference values ​​of the same type in the neighborhood is less than 0, then the two seed points are swapped sequentially.

7. The method for inspecting the coating quality of aircraft components according to claim 6, characterized in that, When performing region growth for a single seed point, the points of interest whose curvature processing difference is less than the preset curvature processing difference and whose symbol similarity index is greater than the preset symbol similarity index are recorded as associated points. The region of the target component model that includes the minimum area of ​​the seed point and the corresponding associated points is recorded as the target region. If the order is adjusted, the initial selection sequence after the order adjustment is completed is recorded as the final sequence; If no order adjustment is performed, the initial sequence is recorded as the final sequence.

8. The method for inspecting the coating quality of aircraft components according to claim 7, characterized in that, For adjacent target areas where the curvature similarity parameter difference value is less than the preset curvature similarity parameter difference value and the spraying distance difference index is less than the preset spraying distance difference index, the merging processing method is determined to be curvature similarity merging, and the adjacent area with the smallest curvature similarity parameter is selected for merging.

9. The method for inspecting the coating quality of aircraft components according to claim 8, characterized in that, For adjacent target areas where the curvature similarity parameter difference value is greater than or equal to the preset curvature similarity parameter difference value or the spraying distance difference index is greater than or equal to the preset spraying distance difference index, the merging processing method is determined to be spraying similarity merging. The trajectory reference direction of the adjacent areas is detected. If the number of adjacent areas with the same trajectory reference direction as the target area is 1, the target area is merged into the adjacent area with the same trajectory reference direction.

10. The method for inspecting the coating quality of aircraft components according to claim 9, characterized in that, If the number of adjacent regions with the same trajectory reference direction as the target region is not 1, then the target region will be merged into the adjacent region with the highest similarity in spray output parameters.

Images