Automobile seat ironing robot control method and system

Abstract

The present application relates to the technical field of non-electric variable control system, and particularly relates to a control method and system of a car seat ironing robot, which comprises the following steps: determining an initial wrinkle comprehensive score of each ironing area of a current car seat according to the gray scale and depth performance in the same ironing area of the current car seat and a plurality of historically ironed car seats of the same model; obtaining a fabric-curvature correction coefficient of each ironing area; correcting the initial wrinkle comprehensive score to obtain a corrected wrinkle comprehensive score; determining a final wrinkle score in combination with the area of each ironing area; determining an updated ironing frequency of each ironing area of the current car seat in combination with a basic ironing frequency; and controlling the car seat ironing robot to perform ironing operation on the current car seat by using the updated ironing frequency. The present application can improve the ironing effect of car seats.

Description

Technical Field

[0001] This invention relates to the field of control system technology for non-electrical variables, and specifically to a control method and system for an automotive seat ironing robot. Background Technology

[0002] As a core comfort component in a vehicle, the smoothness of car seats directly impacts consumers' visual and user experience. In the seat manufacturing process, ironing is a crucial step in eliminating fabric wrinkles and improving the overall appearance. With the continuous expansion of mass production in the automotive industry, automakers are placing higher demands on the efficiency and consistency of seat production.

[0003] Existing problems: Current ironing technology still has many shortcomings and is difficult to match the needs of modern production. Traditional mechanical ironing often uses a fixed number of ironing cycles or sets parameters uniformly according to the seat model, ignoring the individual differences in wrinkles caused by the tension of the leather and fluctuations in the production process within the same model of seat. For example, in the same model of genuine leather seat, some seats have more severe wrinkles on the side wings due to insufficient leather tension, and a fixed number of ironing cycles cannot completely eliminate them; while other seats have only slight wrinkles in the flat area of ​​the backrest, and excessive ironing can damage the surface texture of the genuine leather. Summary of the Invention

[0004] This invention provides a control method and system for an automotive seat ironing robot to solve existing problems.

[0005] The present invention provides a control method and system for an automotive seat ironing robot, which adopts the following technical solution: One embodiment of the present invention provides a control method for an automotive seat ironing robot, the method comprising the following steps: Obtain the ironing areas for each car seat, as well as the quantitative values ​​of the inherent texture of the fabric, the theoretical standard curvature, and the basic number of ironing cycles for each ironing area. Based on the grayscale and depth performance of the same ironing area of ​​the current car seat and several historical car seats of the same model that have passed ironing, determine the initial comprehensive wrinkle score of each ironing area of ​​the current car seat. Based on the quantitative value of the inherent texture of the fabric and the magnitude of the theoretical standard curvature of each ironing area of ​​the current car seat, the fabric-curvature correction coefficient of each ironing area of ​​the current car seat is determined; the initial wrinkle comprehensive score is corrected using the fabric-curvature correction coefficient to obtain the corrected wrinkle comprehensive score of each ironing area of ​​the current car seat; the final wrinkle score of each ironing area of ​​the current car seat is determined based on the area of ​​each ironing area of ​​the current car seat and the magnitude of the corrected wrinkle comprehensive score. Based on the difference in the final wrinkle score of the same ironing area between the current car seat and the historical car seat, and in conjunction with the basic number of ironing operations, the number of times to update the ironing operation for each ironing area of ​​the current car seat is determined; using the number of updated ironing operations, the car seat ironing robot is controlled to perform ironing operations on the current car seat.

[0006] Furthermore, the specific steps for determining the initial wrinkle composite score for each ironing area of ​​the current car seat are as follows: For the current car seat, obtain the entropy and contrast of the gray-level co-occurrence matrix of each ironing area, and record the mean of the normalized entropy and the normalized contrast as the gray-level texture feature value; obtain the normalized value of the standard deviation of the gray-level values ​​of all pixels in each ironing area, and record it as the gray-level difference feature value; obtain the normalized value of the standard deviation of the depth values ​​of all pixels in each ironing area, and record it as the concavity and convexity representation feature value. Based on the grayscale texture feature value, grayscale difference feature value, and concavity / convexity performance feature value of each ironed area of ​​each historical car seat of the same model as the current car seat before and after ironing, determine the grayscale texture feature weight, grayscale difference feature weight, and concavity / convexity performance feature weight. Based on the weights of grayscale texture features, grayscale difference features, and convexity / concave appearance features, the grayscale texture feature values, grayscale difference feature values, and convexity / concave appearance feature values ​​of each ironing area of ​​the current car seat are weighted and averaged to obtain the initial comprehensive wrinkle score of each ironing area of ​​the current car seat.

[0007] Furthermore, the specific steps for determining the grayscale texture feature weights, grayscale difference feature weights, and concavity / convexity representation feature weights are as follows: For each ironed area of ​​any historical car seat of the same model as the current car seat that has passed ironing, calculate the ratio of the grayscale texture feature value before ironing to the grayscale texture feature value after ironing, and record it as the grayscale texture feature change rate; calculate the ratio of the grayscale difference feature value before ironing to the grayscale difference feature value after ironing, and record it as the grayscale difference feature change rate; calculate the ratio of the concavity and convexity feature value before ironing to the concavity and convexity feature value after ironing, and record it as the concavity and convexity feature change rate. For each ironed area of ​​all historical car seats of the same model as the current car seat that have passed ironing, the mean of the grayscale texture feature change rate is recorded as the grayscale texture feature weight; the mean of the grayscale difference feature change rate is recorded as the grayscale difference feature weight; and the mean of the concavity and convexity performance feature change rate is recorded as the concavity and convexity performance feature weight.

[0008] Furthermore, the specific steps for determining the fabric-curvature correction factor for each ironing area of ​​the current car seat are as follows: Based on the quantitative values ​​of the inherent texture of the fabric and the absolute value of the theoretical standard curvature corresponding to all ironing areas of the current car seat, determine the fabric reference value and the curvature reference value. For current car seats, the ratio of the fabric reference value to the quantitative value of the inherent texture of the fabric corresponding to each ironing area is recorded as the fabric correction coefficient. The ratio of the curvature reference value to the absolute value of the theoretical standard curvature corresponding to each ironing area is recorded as the curvature correction coefficient. The sum of the fabric correction coefficient and the curvature correction coefficient is recorded as the fabric-curvature correction coefficient for each ironing area.

[0009] Furthermore, the specific steps for determining the fabric reference value and curvature reference value are as follows: For the current car seat, obtain the minimum value among the quantitative values ​​of the inherent texture of the fabric corresponding to all ironing areas, and record it as the fabric reference value. Obtain the minimum non-zero value among the absolute values ​​of the theoretical standard curvature corresponding to all ironing areas, and record it as the curvature reference value.

[0010] Furthermore, the specific steps for obtaining the comprehensive score of corrected wrinkles for each ironing area of ​​the current car seat are as follows: For the current car seat, the product of the fabric curvature correction factor for each ironing area and the initial wrinkle comprehensive score is denoted as the corrected wrinkle comprehensive score.

[0011] Furthermore, the specific steps for determining the final wrinkle score for each ironing area of ​​the current car seat are as follows: The global adjustment factor is determined based on the combined score of the area of ​​all ironed areas of the current car seat and the correction of wrinkles; The final wrinkle score is the product of the global adjustment factor and the overall score of the corrected wrinkles for each ironing area of ​​the current car seat.

[0012] Furthermore, the specific steps involved in determining the global adjustment factor are as follows: For the current car seat, the area of ​​each ironing area is used as the weight, and the weighted average of the correction wrinkle scores of all ironing areas is calculated to obtain the overall wrinkle score of the current car seat. Obtain the average score of the correction wrinkles for all ironing areas of the current car seat, and record the ratio of the overall wrinkle score of the current car seat to the average score as the global adjustment factor.

[0013] Furthermore, the specific steps for determining the number of times each ironing area of ​​the current car seat needs to be updated include the following: Obtain the average of the final wrinkle scores before ironing for each ironing area of ​​all historical car seats of the same model as the current car seat that have passed ironing, and record it as the baseline wrinkle score for each ironing area. For the current car seat, calculate the relative deviation of the final wrinkle score of each ironing area relative to the baseline wrinkle score, and round the sum of the base number of ironing times for each ironing area and the relative deviation to the nearest integer value, which is recorded as the updated number of ironing times for each ironing area.

[0014] The present invention also proposes a control system for an automotive seat ironing robot, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the aforementioned control method for an automotive seat ironing robot.

[0015] The beneficial effects of the technical solution of the present invention are: In this embodiment of the invention, based on the grayscale and depth performance of the same ironing area of ​​the current car seat and several historical car seats of the same model that have passed ironing, the initial comprehensive wrinkle score of each ironing area of ​​the current car seat is determined. The fabric-curvature correction coefficient of each ironing area is obtained, and the initial comprehensive wrinkle score is corrected to obtain the corrected comprehensive wrinkle score. Combined with the area of ​​each ironing area, the final wrinkle score is determined. Based on the difference between the final wrinkle scores of the same ironing area of ​​the current car seat and historical car seats, and combined with the basic ironing number, the update ironing number of each ironing area of ​​the current car seat is determined. Using the update ironing number, the car seat ironing robot is controlled to perform ironing operations on the current car seat. This invention combines visual perception, big data analysis, and automated control to dynamically match the number of ironing cycles to the wrinkle differences in different individuals and areas of the same model of seat. This avoids under-ironing in severely wrinkled areas and over-ironing in lightly wrinkled areas, preventing damage to the surface texture of leather and other fabrics. It integrates multi-dimensional information such as 2D image texture, grayscale fluctuations, and 3D convexity / concaveness, automatically calculating feature weights based on historical samples without manual weight setting. Multiple features compensate for each other, reducing misjudgments of wrinkles caused by noise in a single image. A fabric correction coefficient is set to offset the influence of the inherent weave, embossing, and grain texture of the seat fabric, avoiding misidentification of the original fabric texture as manufactured wrinkles. The curvature correction coefficient differentiates between the permissible process differences in planar and large-curvature corner and wing areas, aligning with the actual quality inspection logic in seat production. It calculates the overall seat wrinkle score using area weighting and performs global alignment correction on the zone scores, preventing small-area local defects from excessively dominating the overall seat evaluation, maintaining consistency with the global judgment approach of manual quality inspection. The algorithm has low computational load and can be directly adapted to real-time production line control. It includes boundary condition handling to avoid abnormal conditions such as division by zero, demonstrating good algorithm robustness. The benchmark is taken from the seat's own data, reducing the workload of manual calibration on the production line, providing an efficient and accurate automated ironing solution for automotive seat production, and improving the ironing effect of automotive seats. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the steps of a control method for an automotive seat ironing robot according to the present invention. Figure 2 A schematic diagram of the ironing process for car seats. Detailed Implementation

[0017] The following description, in conjunction with the accompanying drawings, details the specific solution of the control method and system for an automotive seat ironing robot provided by this invention.

[0018] Please see Figure 1 The diagram illustrates a flowchart of a control method for an automotive seat ironing robot according to an embodiment of the present invention. The method includes the following steps: Step S001: Obtain the ironing area of ​​each car seat, as well as the quantitative value of the inherent texture of the fabric, the theoretical standard curvature, and the basic number of ironing times for each ironing area.

[0019] In this embodiment, the car seat ironing process is as follows: (1) When the chair arrives at the ironing station, the chair model is determined by scanning the barcode with a barcode scanner, and the line system sends a chair arrival signal.

[0020] (2) The ironing system PLC (Programmable Logic Controller) reads the seat model and the positioning signal, the positioning fork extends, and the system automatically irons according to the seat model.

[0021] (3) After ironing is completed, the positioning fork retracts and the ironing system PLC sends a seat release signal to the production line system.

[0022] (4) Line system release seats.

[0023] (5) The ironing system PLC resets the model of the seat that was just released to zero, and at the same time, the ironing system PLC resets the seat release signal.

[0024] The diagram illustrates the process of ironing car seats, as shown below. Figure 2 As shown.

[0025] The system automatically irons the seats according to their model, specifically: First, for each model of standard car seat, obtain the standard model drawing of each model seat, and manually pre-divide each model seat into several ironing areas and mark them with serial numbers. The division method is as follows: the seat cover is composed of multiple pieces of fabric spliced ​​together, and is manually divided according to function into parts such as the human body contact surface, side wing support area, side wall, backrest back, headrest surfaces, airbag positions, etc. During the division, areas that do not need to be ironed are excluded, that is, each ironing area corresponds to a fabric label (e.g., genuine leather, imitation leather, fabric).

[0026] For each seat model, the fabric panels of the seat cover come with pre-formed weave, embossing, and grain texture at the factory. A non-contact 3D optical profilometer or laser confocal microscope is used to collect the quantified values ​​of the inherent texture of each fabric panel. Specifically, with the fabric flat and wrinkle-free, a wrinkle-free area of ​​a standard fabric sample is scanned and inspected, and the arithmetic mean roughness is output as a quantified representation of the inherent texture of that fabric panel—this is a known technique. Simultaneously, for each seat model, the standard digital CAD model of the seat is read to obtain the theoretical standard curvature of each ironing area—this is a known technique. Based on the fabric label, the quantified value of the inherent texture, and the theoretical standard curvature of each ironing area, a basic number of ironing cycles is manually set for each ironing area. In this embodiment, the basic number of ironing cycles is at least 1 and at most 4.

[0027] Therefore, several ironing areas are defined in the standard model diagram of each seat model, along with the quantitative value of the inherent fabric texture, theoretical standard curvature, and basic number of ironing cycles for each ironing area. These values ​​are stored in a database for later retrieval.

[0028] For each car seat, the seat model is determined by scanning the code, and the model information is transmitted to the PLC controller. The PLC then calls up the standard model diagram of that model in advance.

[0029] The car seat ironing robot is equipped with two industrial cameras to collect 2D visual images and 3D depth data of the seats.

[0030] After each car seat is in place, the PLC triggers industrial cameras (one each for the front and side) to capture 2D visual images and 3D depth data of the seat. The capture resolution is 1920×1080, and the 3D depth data accuracy is ±0.1mm. Before data acquisition, the system automatically performs adaptive white balance adjustment to eliminate the influence of uneven lighting in the workshop (such as local shadows and strong light reflections) on the image. Furthermore, after ironing is completed, the 2D visual images and 3D depth data of the seat are captured again.

[0031] For each car seat, 2D visual images taken before or after ironing are matched with the corresponding model image to obtain the ironing areas for each car seat. Specifically, for each car seat's standard model image, key feature points (such as seat back edges, side wing seams, headrest interfaces, etc., approximately 50-80 feature points per model) are manually preset. The ORB feature point matching algorithm is used to prioritize matching these preset key feature points, improving matching efficiency and accuracy. After matching, the coordinate transformation matrix between the photographed image and the model image is calculated, mapping the coordinates of the ironing areas in the model image to the photographed image, outputting the precise coordinate range of each ironing area. This is a well-known technique, and the specific method will not be described here.

[0032] This yields the ironing areas for each car seat before and after ironing, with each pixel within each ironing area corresponding to a grayscale value and a depth value. It also includes the corresponding marker number, fabric texture quantification value, theoretical standard curvature, and basic ironing count for each ironing area retrieved from the corresponding model.

[0033] Step S002: Based on the grayscale and depth performance of the same ironing area of ​​the current car seat and several historical car seats of the same model that have passed ironing, determine the initial comprehensive wrinkle score of each ironing area of ​​the current car seat.

[0034] It should be noted that wrinkles are a common defect in car seats after upholstery assembly. These wrinkles originate from several factors: the fabric itself has anisotropic stretching properties, and residual stress is introduced during cutting and sewing; the seat's complex free-form surface makes it difficult to achieve uniform surface tension distribution during upholstery, leading to fabric stacking in high-curvature areas such as side wings and corners; deviations in foam size, mismatch between the cover and foam profiles, and stress release caused by changes in workshop temperature and humidity can all contribute to wrinkles on the finished seat surface. Therefore, even for the same model of car seat in the same ironing area, different car seats produced may exhibit different wrinkle patterns within that area. Thus, it is necessary to dynamically adjust the number of ironing cycles to avoid over- or under-ironing.

[0035] Preferably, in one embodiment of the present invention, the method for obtaining the initial wrinkle composite score of each ironing area of ​​the current car seat includes: For the current car seat (not ironed, only data collected before ironing exists), obtain the first... The entropy value of the gray-level co-occurrence matrix of the ironing area With contrast , entropy value normalized value With contrast normalized value The mean value is denoted as the grayscale texture feature value. .in, This is a function that takes the minimum value.

[0036] It should be noted that the construction of the gray-level co-occurrence matrix, as well as the calculation of its entropy and contrast, are well-known techniques, and the specific methods will not be described here. A larger entropy value indicates higher texture complexity and more severe wrinkles; a larger contrast value indicates more obvious differences in texture brightness and deeper wrinkles; the combination of the two yields the gray-level texture feature value.

[0037] For the current car seat, obtain the first Standard deviation of grayscale values ​​of all pixels within an ironing area normalized value , denoted as the grayscale difference feature value .

[0038] The larger the grayscale standard deviation, the more intense the fluctuation of light and dark within the area, and the more obvious the wrinkles.

[0039] For the current car seat, obtain the first Standard deviation of depth values ​​of all pixels within an ironing area normalized value , denoted as the eigenvalue of the concavity / convexity representation. .

[0040] The larger the standard deviation of depth, the more obvious the undulations and the more severe the wrinkles in the region.

[0041] Select 100 historical car seats of the same model as the current car seats that have passed ironing, and retrieve their before and after ironing images. For the same ironing area, based on the before-ironing image information, obtain the first-to-last image of all historical car seats before ironing. The maximum value among the entropy values ​​of the gray-level co-occurrence matrix of each ironing area. Get all historical car seat ironing before The maximum contrast value in the gray-level co-occurrence matrix of each ironing area Get all historical car seat ironing before The maximum value among the standard deviations of the grayscale values ​​of all pixels within the ironing area. Get all historical car seat ironing before The maximum value among the standard deviations of the depth values ​​of all pixels within the ironing area. .

[0042] It should be noted that in this embodiment, by increasing the number of historical car seats, the following is ensured: , , , All values ​​are greater than 0. Historically, each ironing area of ​​the car seat was ironed using the corresponding base number of ironing cycles.

[0043] According to the current car seat... Grayscale texture feature value of each ironing area Gray-scale difference feature value Concave and convex characteristic values The acquisition method involves obtaining grayscale texture feature values, grayscale difference feature values, and concavity / convexity performance feature values ​​for each ironing area of ​​each historical car seat based on image information before and after ironing.

[0044] For each historical car seat, the first For each ironing area, calculate the ratio of the grayscale texture feature value before ironing to the grayscale texture feature value after ironing, denoted as the grayscale texture feature change rate; calculate the ratio of the grayscale difference feature value before ironing to the grayscale difference feature value after ironing, denoted as the grayscale difference feature change rate; calculate the ratio of the concavity / convexity feature value before ironing to the concavity / convexity feature value after ironing, denoted as the concavity / convexity feature change rate.

[0045] For all historical car seats, the first For each ironing area, the average rate of change of grayscale texture features is denoted as the grayscale texture feature weight. The mean of the rate of change of gray-level difference features is denoted as the gray-level difference feature weight. The mean of the rate of change of the concavity / convexity performance characteristics is denoted as the weight of the concavity / convexity performance characteristics. .

[0046] It should be noted that: the feature change rate characterizes the degree to which the feature decays with ironing; the larger the change rate value, the greater the decrease in the feature after ironing, the higher the sensitivity of the feature to wrinkle removal, and the greater the corresponding weight. In calculating the change rate of each feature, if the feature value is 0 before or after ironing, the invalid sample data is removed; and by expanding the number of historical seat samples, it is ensured that each ironing area has at least 50 valid feature change rate samples, thereby ensuring the weight of grayscale texture features. Gray-scale difference feature weights Concave-convexity representation feature weights All are greater than 0.

[0047] Based on grayscale texture feature weights Gray-scale difference feature weights Concave-convexity representation feature weights Regarding the current car seat's first Grayscale texture feature value of each ironing area Gray-scale difference feature value Concave and convex characteristic values Perform a weighted average to obtain the current car seat's number. Overall score of initial wrinkles in each ironing area The specific calculation formula is as follows: Therefore, based on the statistical analysis of the feature change rate before and after ironing of historical ironed seat samples, the weights of each feature are obtained. The initial comprehensive wrinkle score is obtained by weighted averaging of grayscale texture feature values, grayscale difference feature values, and concavity / convexity performance feature values. This method eliminates the need for manual, experience-based weight setting; it integrates multi-dimensional information from two-dimensional image texture, grayscale fluctuations, and three-dimensional concavity / convexity, allowing multiple features to compensate for each other and reducing misjudgments caused by noise from a single feature; sensitive features receive a higher contribution ratio, resulting in better performance of the actual bonding wrinkle process; the weighted average is normalized, ensuring stable output values, facilitating subsequent fabric and curvature correction calculations, minimizing computational overhead, and adapting to real-time production line control scenarios.

[0048] Step S003: Based on the quantitative value of the inherent texture of the fabric and the magnitude of the theoretical standard curvature corresponding to each ironing area of ​​the current car seat, determine the fabric-curvature correction coefficient for each ironing area of ​​the current car seat; use the fabric-curvature correction coefficient to correct the initial wrinkle comprehensive score to obtain the corrected wrinkle comprehensive score for each ironing area of ​​the current car seat; determine the final wrinkle score for each ironing area of ​​the current car seat based on the area of ​​each ironing area and the magnitude of the corrected wrinkle comprehensive score.

[0049] It's important to note that different seat fabrics inherently possess textures, embossing, and grain. Even without manufactured wrinkles, visual imaging will still output a certain wrinkle feature score. Directly calculating the degree of wrinkles in a region using the original visual features will misjudge the fabric's inherent texture as manufactured wrinkles. For example, perforated leather, fine-textured imitation leather, and coarse-textured fabrics will show significant differences in visual detection scores for the same genuine wrinkles. Therefore, a fabric correction coefficient is needed to offset the inherent texture baseline differences. Smooth fabrics require no correction; fabrics with rich textures and higher original visual scores require attenuation correction.

[0050] It should be further noted that the same visual wrinkle feature has different permissible levels of processing tolerance when placed in areas with high curvature at the edges and corners versus flat areas. Areas with high forming stress at the edges and corners allow for slight wrinkles, while flat areas have stricter requirements. Therefore, a regional curvature correction coefficient needs to be introduced simultaneously. Ironing areas are approximately flat with very little curvature, requiring no curvature correction; ironing areas with high curvature will have their surface geometry inflate the original visual score, requiring attenuation.

[0051] Preferably, in one embodiment of the present invention, the method for obtaining the final wrinkle score of each ironing area of ​​the current car seat includes: For the current car seat, obtain the minimum value among the quantitative values ​​of the inherent texture of the fabric corresponding to all ironing areas, and record it as the fabric reference value. Obtain the minimum non-zero absolute value of the theoretical standard curvature corresponding to all ironing areas, and denote it as the curvature reference value. .

[0052] It should be noted that the quantitative value of the inherent texture of the fabric collected must be greater than 0.

[0053] For current car seats, the fabric benchmark value With the Quantitative value of the inherent texture of the fabric corresponding to each ironing area ratio , denoted as the fabric correction factor , curvature reference value With the The absolute value of the theoretical standard curvature corresponding to each ironing area ratio , denoted as curvature correction factor .

[0054] It should be noted that: if Explanation of the first The ironing area is approximately flat and requires no correction, so let In this embodiment, the fabric with the weakest internal texture of the current seat is used as the fabric benchmark. This normalizes the texture baseline between different fabric blocks within the same seat, eliminating artificially high wrinkle scores caused by the inherent weave, embossing, and grain textures of different fabric panels, and preventing the fabric's inherent texture from being misidentified as process wrinkles. For smooth fabric areas, the fabric correction coefficient approaches 1, hardly interfering with the score; for textured fabric areas, the fabric correction coefficient is less than 1, automatically attenuating and correcting the initial wrinkle score, conforming to objective process rules, and eliminating the need for manually setting thresholds for each type of fabric. Furthermore, the benchmark is taken from the current seat's own data, adapting to the working conditions of multiple fabric panels on a single seat, without relying on external standardized fabrics, reducing the workload of production line calibration. Meanwhile, using the minimum non-zero curvature of the current seat as the curvature benchmark, the algorithm distinguishes the geometric differences between the flat planar area, the large curvature side wing area, and the corner transition area. The larger the absolute value of curvature in the large curvature area, the smaller the curvature correction coefficient, automatically attenuating wrinkle scores. This aligns with common manufacturing practices: the corner and side wing areas experience high forming stress, allowing for slight wrinkles, while the planar main body area requires stringent wrinkle control. For the ideal planar area, the boundary condition is directly assigned a value of 1, without curvature compensation, avoiding over-correction of the planar area; this also avoids the division-by-zero calculation anomalies caused by a curvature of 0, improving the algorithm's robustness.

[0055] For the current car seat, calculate the first... Fabric correction factor for each ironing area With curvature correction factor The sum of these values ​​is denoted as the fabric-curvature correction coefficient. The fabric-curvature correction coefficient is then combined with the first... Overall score of initial wrinkles in each ironing area The product of these is denoted as the corrected wrinkle composite score. The specific calculation formula is as follows: It should be noted that: This yields the comprehensive score for corrected wrinkles after joint fabric-curvature correction. The summation and fusion method can simultaneously reflect two types of interference factors: fabric texture deviation and surface curvature deviation. If either the fabric or curvature is too large, the sum value decreases, thus attenuating the initial wrinkle score, while simultaneously achieving fabric baseline compensation and surface curvature compensation.

[0056] For the current car seat, the overall wrinkle score is obtained by weighting the area (number of pixels in each ironing area) of each ironing region and taking a weighted average of the wrinkle correction scores of all ironing regions. The specific calculation formula is as follows: In the formula, This represents the number of ironing areas on the current car seats. For the current car seat The area of ​​the ironing zone.

[0057] It should be noted that the overall seat wrinkle level is calculated using the area of ​​the ironed area as the weight. Larger areas such as the seat cushion and main backrest contribute more to the overall wrinkle score, while smaller areas like corners, airbag locations, and edging have a less significant impact on the overall seat. This approach aligns more closely with the vehicle manufacturer's actual evaluation logic for the overall appearance quality of the seat, preventing small, localized defects from excessively dominating the evaluation results. A weighted average is used to synthesize the wrinkle information after correction from all ironed areas, resulting in a global reference value that characterizes the entire seat, providing a reliable benchmark for subsequent global adjustments to the regional scores.

[0058] The current car seat Final wrinkle score for each ironing area The calculation formula is: In the formula, This is the average score of the wrinkle correction for all ironed areas of the current car seat. If... Then directly order When the wrinkle score for all regions is 0 (completely qualified with no wrinkles), the global scaling adjustment is skipped to prevent calculation failure and ensure the algorithm runs stably under all working conditions. As a global adjustment factor, molecule Dominated by a large area of ​​main regions, the denominator It's a simple arithmetic average, with equal weighting for each region. A global adjustment factor is used to align and correct the score of each sub-region to the overall wrinkle level of the entire seat, adjusting the wrinkle performance of each sub-region using the overall wrinkle score of the entire seat. The global adjustment factor is used to assess the skewness of a large area relative to the overall average. When wrinkles are more pronounced in a large main area, the numerator is more significantly affected by the area weight, and the global adjustment factor increases. Based on this increased factor, the scores of each sub-region are multiplied and amplified, thereby improving the overall detection sensitivity of the system when the main performance is poor. That is, when the wrinkles of the entire seat are generally heavy, the scores of each sub-region are moderately increased; when the overall quality of the entire seat is good, the scores of each sub-region are decreased, avoiding the isolation of individual area wrinkles and maintaining consistency with the global evaluation thinking of manual quality inspection.

[0059] Step S004: Based on the difference in the final wrinkle score of the same ironing area between the current car seat and the historical car seat, and in conjunction with the basic ironing count, determine the update ironing count for each ironing area of ​​the current car seat; using the update ironing count, control the car seat ironing robot to perform ironing operations on the current car seat.

[0060] Preferably, in one embodiment of the present invention, the method for obtaining the number of times each ironing area of ​​the current car seat has been updated includes: For 100 historical car seats of the same model as the current car seats that have passed ironing, based on the image information before ironing, obtain the number of all historical car seats. The average of the final wrinkle scores of each ironing area before ironing is denoted as the i-th. The baseline wrinkle score for each ironing area .

[0061] It should be noted that: the number of seats in historical vehicles can be increased to ensure... Greater than 0.

[0062] Get the current car seat number Basic number of ironing passes for each ironing area Then the current car seat's number Update the number of ironing cycles for each ironing area. The calculation formula is: In the formula, This indicates rounding to the nearest integer. If Then let ;like Then let . for Compared to The relative deviation.

[0063] It should be noted that the minimum number of ironing passes is 1, and the maximum is 4. Because historically, each ironing area of ​​a car seat was ironed using the corresponding base number of ironing passes, when… When, it indicates the current car seat's number of... If the wrinkles in the ironing area are relatively small, reduce the number of ironing passes to avoid over-ironing. When When, it indicates the current car seat's number of... If the wrinkles in the ironing area are relatively large, increase the number of ironing cycles to avoid under-ironing.

[0064] The PLC sends the precise coordinates of each ironing area of ​​the car seat and the number of ironing cycles to the car seat ironing robot, which then performs the ironing operation.

[0065] The present invention also provides a control system for an automotive seat ironing robot, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program stored in the memory to implement the steps of the aforementioned control method for an automotive seat ironing robot.

[0066] This invention is now complete.

[0067] In summary, in this embodiment of the invention, based on the grayscale and depth characteristics of the same ironing area of ​​the current car seat and several historical car seats of the same model that have passed ironing, an initial comprehensive wrinkle score for each ironing area of ​​the current car seat is determined. A fabric-curvature correction coefficient for each ironing area is obtained, and the initial comprehensive wrinkle score is corrected to obtain a corrected comprehensive wrinkle score. Combined with the area of ​​each ironing area, a final wrinkle score is determined. Based on the difference in the final wrinkle scores of the same ironing area between the current car seat and historical car seats, and combined with the basic number of ironing passes, the number of update ironing passes for each ironing area of ​​the current car seat is determined. Using these update ironing passes, the car seat ironing robot is controlled to perform ironing operations on the current car seat. This invention can improve the ironing effect of car seats.

Claims

1. A control method for a car seat ironing robot, characterized in that, The method includes the following steps: Obtain the ironing areas for each car seat, as well as the quantitative values ​​of the inherent texture of the fabric, the theoretical standard curvature, and the basic number of ironing cycles for each ironing area. Based on the grayscale and depth performance of the current car seat and several historical car seats of the same model that passed ironing, determine the initial comprehensive wrinkle score of each ironing area of ​​the current car seat. Based on the quantitative value of the inherent texture of the fabric and the magnitude of the theoretical standard curvature of each ironing area of ​​the current car seat, the fabric-curvature correction coefficient of each ironing area of ​​the current car seat is determined; the initial wrinkle comprehensive score is corrected using the fabric-curvature correction coefficient to obtain the corrected wrinkle comprehensive score of each ironing area of ​​the current car seat; the final wrinkle score of each ironing area of ​​the current car seat is determined based on the area of ​​each ironing area of ​​the current car seat and the magnitude of the corrected wrinkle comprehensive score. Based on the difference in the final wrinkle score of the same ironing area between the current car seat and the historical car seat, and in conjunction with the basic number of ironing operations, the number of times to update the ironing operation for each ironing area of ​​the current car seat is determined; using the number of updated ironing operations, the car seat ironing robot is controlled to perform ironing operations on the current car seat.

2. The control method for an automotive seat ironing robot according to claim 1, characterized in that, The specific steps for determining the initial wrinkle score for each ironing area of ​​the current car seat are as follows: For the current car seat, obtain the entropy value and contrast of the gray-level co-occurrence matrix of each ironing area, and record the mean of the normalized value of the entropy value and the normalized value of the contrast as the gray-level texture feature value. Obtain the normalized value of the standard deviation of the grayscale values ​​of all pixels in each ironing area, and record it as the grayscale difference feature value; Obtain the normalized value of the standard deviation of the depth values ​​of all pixels in each ironing area, and denote it as the concavity and convexity feature value; Based on the grayscale texture feature value, grayscale difference feature value, and concavity / convexity performance feature value of each ironed area of ​​each historical car seat of the same model as the current car seat before and after ironing, determine the grayscale texture feature weight, grayscale difference feature weight, and concavity / convexity performance feature weight. Based on the weights of grayscale texture features, grayscale difference features, and convexity / concave appearance features, the grayscale texture feature values, grayscale difference feature values, and convexity / concave appearance feature values ​​of each ironing area of ​​the current car seat are weighted and averaged to obtain the initial comprehensive wrinkle score of each ironing area of ​​the current car seat.

3. The control method for an automotive seat ironing robot according to claim 2, characterized in that, The specific steps for determining the grayscale texture feature weights, grayscale difference feature weights, and concavity / convexity representation feature weights are as follows: For each ironing area of ​​any historical car seat of the same model as the current car seat that has passed ironing, calculate the ratio of the grayscale texture feature value before ironing to the grayscale texture feature value after ironing, and record it as the grayscale texture feature change rate. The ratio of the grayscale difference characteristic value before ironing to the grayscale difference characteristic value after ironing is recorded as the grayscale difference characteristic change rate; the ratio of the concave-convex appearance characteristic value before ironing to the concave-convex appearance characteristic value after ironing is recorded as the concave-convex appearance characteristic change rate. For each ironed area of ​​all historical car seats of the same model as the current car seat that have passed ironing, the mean of the grayscale texture feature change rate is recorded as the grayscale texture feature weight. The mean of the rate of change of grayscale difference features is denoted as the grayscale difference feature weight; the mean of the rate of change of concavity / convexity representation features is denoted as the concavity / convexity representation feature weight.

4. The control method for an automotive seat ironing robot according to claim 1, characterized in that, The specific steps for determining the fabric-curvature correction factor for each ironing area of ​​the current car seat are as follows: Based on the quantitative values ​​of the inherent texture of the fabric and the absolute value of the theoretical standard curvature corresponding to all ironing areas of the current car seat, determine the fabric reference value and the curvature reference value. For current car seats, the ratio of the fabric reference value to the quantitative value of the inherent texture of the fabric corresponding to each ironing area is recorded as the fabric correction coefficient. The ratio of the curvature reference value to the absolute value of the theoretical standard curvature corresponding to each ironing area is recorded as the curvature correction coefficient. The sum of the fabric correction coefficient and the curvature correction coefficient is recorded as the fabric-curvature correction coefficient for each ironing area.

5. The control method for an automotive seat ironing robot according to claim 4, characterized in that, The specific steps for determining the fabric reference value and curvature reference value are as follows: For the current car seat, obtain the minimum value among the quantitative values ​​of the inherent texture of the fabric corresponding to all ironing areas, and record it as the fabric reference value. Obtain the minimum non-zero value among the absolute values ​​of the theoretical standard curvature corresponding to all ironing areas, and record it as the curvature reference value.

6. The control method for an automotive seat ironing robot according to claim 1, characterized in that, The specific steps for obtaining the comprehensive score of corrected wrinkles for each ironing area of ​​the current car seat are as follows: For the current car seat, the product of the fabric curvature correction factor for each ironing area and the initial wrinkle comprehensive score is denoted as the corrected wrinkle comprehensive score.

7. The control method for an automotive seat ironing robot according to claim 1, characterized in that, The specific steps for determining the final wrinkle score for each ironing area of ​​the current car seat are as follows: The global adjustment factor is determined based on the combined score of the area of ​​all ironed areas of the current car seat and the correction of wrinkles; The final wrinkle score is the product of the global adjustment factor and the overall score of the corrected wrinkles for each ironing area of ​​the current car seat.

8. The control method for an automotive seat ironing robot according to claim 7, characterized in that, The specific steps involved in determining the global adjustment factor are as follows: For the current car seat, the area of ​​each ironing area is used as the weight, and the weighted average of the correction wrinkle scores of all ironing areas is calculated to obtain the overall wrinkle score of the current car seat. Obtain the average score of the correction wrinkles for all ironing areas of the current car seat, and record the ratio of the overall wrinkle score of the current car seat to the average score as the global adjustment factor.

9. The control method for an automotive seat ironing robot according to claim 1, characterized in that, The specific steps for determining the number of times each ironing area of ​​the current car seat needs to be updated are as follows: Obtain the average of the final wrinkle scores before ironing for each ironing area of ​​all historical car seats of the same model as the current car seat that have passed ironing, and record it as the baseline wrinkle score for each ironing area. For the current car seat, calculate the relative deviation of the final wrinkle score of each ironing area relative to the baseline wrinkle score, and round the sum of the base number of ironing times for each ironing area and the relative deviation to the nearest integer value, which is recorded as the updated number of ironing times for each ironing area.

10. A control system for an automotive seat ironing robot, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed by the processor, it implements the steps of the control method for an automotive seat ironing robot as described in any one of claims 1-9.

Images