Intelligent sleeping bed frame assembly process optimization method and system based on machine vision

Abstract

The present invention relates to the technical field of bed frame assembly process optimization, and in particular to an intelligent sleeping bed frame assembly process optimization method and system based on machine vision. Based on a preset equipment process plan, the target assembly equipment is controlled to perform assembly production, the assembly state image of the assembled bed frame is obtained at multiple preset angles, and the assembly state image is subjected to feature extraction processing, and a three-dimensional feature shape map of the assembled bed frame is constructed according to the points of interest of the assembly state image at each preset angle; the assembly state of the assembled bed frame is obtained according to the three-dimensional feature shape map of the assembled bed frame and the standard feature shape map; if the assembly state of the assembled bed frame is a defective assembly product, the cause of the defective assembly is traced and analyzed according to the defective assembly area, and equipment process plan optimization suggestions are generated according to the analysis results. By evaluating and analyzing the preset equipment process plan, the product quality control capability of the bed frame in the actual production process can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of bed frame assembly process optimization, and in particular to an intelligent sleeping bed frame assembly process optimization method and system based on machine vision. Background Art

[0002] With the continuous development of science and technology and the improvement of people's living standards, smart bed frames, as a new type of smart home product, have gradually been favored by consumers. Smart bed frames not only have the basic functions of traditional bed frames, but can also monitor and adjust the sleeping environment through sensors, controllers and other devices, providing users with a more comfortable and healthy sleeping experience. However, the assembly process of smart bed frames is relatively complex, requiring high-precision assembly technology and strict quality control to ensure the performance and reliability of the product. Therefore, how to improve the assembly efficiency and quality of smart bed frames has become an important issue facing current smart bed frame manufacturers. Machine vision technology, as a non-contact detection technology, has the advantages of high precision, high efficiency and high reliability, and has been widely used in industrial production. Applying machine vision technology to the assembly process of smart bed frames can realize real-time monitoring and regulation of the assembly process, thereby optimizing the production process. Therefore, studying the optimization method of the assembly process of smart bed frames based on machine vision has important theoretical significance and practical application value. Summary of the invention

[0003] The present invention overcomes the deficiencies of the prior art and provides a method and system for optimizing the assembly process of an intelligent sleeping bed frame based on machine vision.

[0004] To achieve the above-mentioned purpose, the technical solution adopted by the present invention is:

[0005] The first aspect of the present invention discloses a method for optimizing the assembly process of an intelligent sleeping bed frame based on machine vision, comprising the following steps:

[0006] Obtaining a preset equipment process plan for the smart sleeping bed frame, controlling the target assembly equipment to perform assembly production based on the preset equipment process plan, and obtaining assembly state images of the assembled bed frame at multiple preset angles after completing an assembly cycle;

[0007] Selecting each pixel point in the assembly state image as a candidate point, determining each candidate point as an interest point or a non-interest point according to the number of bright spots and dark spots in a circular area of ​​each candidate point, and extracting the interest points of the assembly state image at each preset angle;

[0008] Determine adjacent interest points according to the Euclidean distance between each interest point, connect each adjacent interest point to form a plurality of curves, import the plurality of curves formed by connecting each adjacent interest point into the voxel grid space for singularity analysis and smoothing, and generate a three-dimensional characteristic shape diagram of the assembled bed frame according to each smoothed curve;

[0009] Obtaining a standard characteristic shape diagram of the assembled bed frame, calculating the overlap degree of the mutually overlapping surfaces between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram, and determining the assembly state of the assembled bed frame according to the overlap degree;

[0010] If the assembly status of the assembled bed frame is a defective assembly product, the defective assembly area of ​​the assembled bed frame is obtained, the cause of the defective assembly is traced and analyzed according to the defective assembly area, and an optimization suggestion for the equipment process plan is generated according to the analysis result;

[0011] If the assembly status of the assembled bed frame is a good assembly product, the target assembly equipment is controlled to execute the next assembly cycle. If the assembly status of all assembled bed frames is a good assembly product after the target assembly equipment executes the preset assembly cycles, it means that the preset equipment process plan meets the mass production conditions, and there is no need to optimize the preset equipment process plan.

[0012] Preferably, each pixel point is selected as a candidate point in the assembly state image, and each candidate point is determined as an interest point or a non-interest point according to the number of bright spots and dark spots in the circular area of ​​each candidate point, and the interest points of the assembly state image at each preset angle are extracted, specifically:

[0013] Performing gray value conversion and denoising processing on the assembly state image, randomly selecting a pixel point in the assembly state image as a candidate point, and obtaining a brightness value of the candidate point;

[0014] A circular area with a preset radius is defined around the candidate point, the remaining pixels in the circular area are retrieved, and the brightness values ​​of the remaining pixels are obtained;

[0015] Compare the brightness value of each remaining pixel with the brightness value of the candidate point respectively; if the brightness value of a remaining pixel is greater than the brightness value of the candidate point, mark the remaining pixel as a bright spot; if the brightness value of a remaining pixel is not greater than the brightness value of the candidate point, mark the remaining pixel as a dark spot;

[0016] Statistically analyzing the number of bright spots and dark spots in the circular area of ​​the candidate point, if the number of bright spots or dark spots in the circular area of ​​the candidate point is greater than a preset value, the candidate point is marked as an interest point; otherwise, the candidate point is marked as a non-interest point;

[0017] The next pixel point is selected as a candidate point, and steps S204 to S208 are repeated until all the pixel points are traversed, and a list of interest points in the assembly state image is output.

[0018] Preferably, adjacent interest points are determined according to the Euclidean distance between each interest point, each adjacent interest point is connected to form a plurality of curves, the plurality of curves formed by connecting each adjacent interest point is introduced into the voxel grid space for singularity analysis and smoothing, and a three-dimensional characteristic shape diagram of the assembled bed frame is generated according to each smoothed curve, specifically:

[0019] Obtain the coordinate values ​​of the points of interest in the corresponding images of the assembly state images at each preset angle, and obtain the calibration parameters of the camera;

[0020] Constructing a Cartesian coordinate system, and converting the coordinates of the points of interest in the images of the assembly state at each preset angle into the Cartesian coordinate system according to the calibration parameters of the camera, so as to ensure that the coordinates of the points of interest in all the images of the assembly state at the preset angle are in the same coordinate system;

[0021] Calculating the Euclidean distance between each interest point in the Cartesian coordinate system, marking the interest point with the shortest Euclidean distance as an adjacent interest point, and connecting each adjacent interest point to form a plurality of curves;

[0022] Construct a voxel grid space, import a number of curves formed by connecting adjacent interest points into the voxel grid space, obtain the intersection points of each curve and the voxel grid, define them as second-order difference nodes, and obtain the second-order difference nodes of each curve;

[0023] Perform a second-order difference operation on the second-order difference nodes of each curve to obtain the difference value of each curve, and construct a difference value matrix according to the difference value of each curve; wherein the rows in the matrix represent each curve, the columns represent the positions of the second-order difference nodes, and each element represents the difference value of the corresponding curve at the corresponding second-order difference node;

[0024] Traversing each element in the difference value matrix, comparing each element in the difference value matrix with a preset threshold, and if an element in the difference value matrix is ​​greater than the preset threshold, marking the curve node corresponding to the element in the difference value matrix as a singular position node;

[0025] Repeat step S312 until all elements in the differential value matrix are traversed to obtain the singular position nodes of each curve, and perform smoothing on the singular position nodes of each curve to obtain each smoothed curve;

[0026] A three-dimensional characteristic shape diagram of the assembled bed frame is generated according to each smoothed curve.

[0027] Preferably, the overlap degree of the mutually overlapping surfaces between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is calculated, and the assembly state of the assembled bed frame is determined according to the overlap degree, specifically:

[0028] Aligning the three-dimensional feature shape image with the standard feature shape image based on the SURF algorithm, and after the alignment is completed, obtaining the overlapping area value of the overlapping surface between the three-dimensional feature shape image and the standard feature shape image;

[0029] Calculate the total area value of each face in the standard characteristic shape image, divide the overlapping area value by the total area value, and obtain the degree of overlap between the three-dimensional characteristic shape image and the standard characteristic shape image;

[0030] If the overlap between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is greater than a preset overlap, the assembled bed frame is marked as a good assembly product;

[0031] If the degree of overlap between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is not greater than the preset degree of overlap, the assembled bed frame is marked as a defective assembly product, and the area of ​​the non-overlapping surface between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is obtained and defined as the defective assembly area.

[0032] Preferably, if the assembly state of the assembled bed frame is a defective assembly product, the defective assembly area of ​​the assembled bed frame is obtained, and the cause of the defective assembly is traced and analyzed according to the defective assembly area, and an optimization suggestion for the equipment process solution is generated according to the analysis result, specifically:

[0033] If the assembly state of the assembled bed frame is a defective assembly product, obtaining the defective assembly area of ​​the assembled bed frame and obtaining the work log of the target assembly equipment;

[0034] According to the work log, various actual assembly parameters at various preset assembly time nodes when the target assembly equipment assembles the poorly assembled area are acquired to obtain various actual assembly parameter sets;

[0035] and obtaining a preset equipment process scheme for the smart sleeping bed frame, and obtaining various preset assembly parameters at various preset assembly time nodes when the target assembly equipment assembles the poorly assembled area according to the preset equipment process scheme, to obtain various preset assembly parameter sets;

[0036] A dynamic time warping algorithm is introduced to calculate the similarity between each actual assembly parameter set and the corresponding preset assembly parameter set based on the dynamic time warping algorithm;

[0037] If the similarities between the actual assembly parameter sets and the corresponding preset assembly parameter sets when the target assembly equipment assembles the poor assembly area are greater than the preset similarity threshold, it means that the assembly parameters of the target assembly equipment when assembling the poor assembly area are normal, which means that the production of defective assembly products is caused by defects in the equipment process plan itself, and then an equipment process plan optimization suggestion is generated based on the poor assembly area.

[0038] The following steps are also included:

[0039] If there is a situation where the similarity between at least one actual assembly parameter set and the corresponding preset assembly parameter set is not greater than a preset similarity threshold when the target assembly equipment assembles the poorly assembled area, then obtaining the actual assembly parameter set corresponding to the similarity not greater than the preset similarity threshold;

[0040] Determine abnormal assembly parameters in the target assembly equipment according to the actual assembly parameter set corresponding to the similarity not greater than the preset similarity threshold, and perform correction processing on the abnormal assembly parameters;

[0041] After the correction processing of the abnormal assembly parameters is completed, the target assembly equipment is controlled to execute the next assembly cycle, and the assembly state of the assembled bed frame assembled after the execution of the next assembly cycle is obtained; if the assembly state of the assembled bed frame assembled after the execution of the next assembly cycle is a good assembly product, it means that the generation of defective assembly products is caused by the drift of the assembly parameters of the target assembly equipment, and there is no need to optimize the equipment process plan at this time;

[0042] If the assembly status of the assembled bed frame obtained after executing the next assembly cycle is still a defective assembly product, it means that the generation of the defective assembly product is also caused by defects in the equipment process solution itself, and then an equipment process solution optimization suggestion is generated based on the defective assembly area.

[0043] The second aspect of the present invention discloses a machine vision-based intelligent sleeping bed frame assembly process optimization system, the intelligent sleeping bed frame assembly process optimization system includes a memory and a processor, the memory stores an intelligent sleeping bed frame assembly process optimization method program, when the intelligent sleeping bed frame assembly process optimization method program is executed by the processor, any step of the intelligent sleeping bed frame assembly process optimization method is implemented. The present invention solves the technical defects existing in the background technology, and has the following beneficial effects: based on a preset equipment process scheme, a target assembly equipment is controlled to perform assembly production, assembly state images of the assembled bed frame are obtained at multiple preset angles, and feature extraction processing is performed on the assembly state images, and a three-dimensional feature shape map of the assembled bed frame is constructed according to the interest points of the assembly state images at each preset angle; the assembly state of the assembled bed frame is obtained according to the three-dimensional feature shape map of the assembled bed frame and the standard feature shape map; if the assembly state of the assembled bed frame is an assembly defective product, the cause of the assembly defect is traced and analyzed according to the assembly defective area, and equipment process scheme optimization suggestions are generated according to the analysis results; if the assembly state of the assembled bed frame is an assembly good product, the target assembly equipment is controlled to execute the next assembly cycle, and if the assembly states of all assembled bed frames are assembly good products after the target assembly equipment executes the preset assembly cycles, it means that the preset equipment process scheme meets the mass production conditions, and there is no need to optimize the preset equipment process scheme. Through automation and intelligent means, the evaluation and analysis of preset equipment process plans can effectively improve the product quality control capabilities of bed frames in the actual production process, and provide strong support for the efficient and high-quality mass production of smart sleeping bed frames. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0045] Figure 1 This is a flow chart of the first method of the intelligent sleeping bed frame assembly process optimization method;

[0046] Figure 2 A flow chart of the second method of the intelligent sleeping bed frame assembly process optimization method;

[0047] Figure 3 This is the system block diagram of the intelligent sleeping bed frame assembly process optimization system. DETAILED DESCRIPTION

[0048] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0049] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0050] like Figure 1 As shown, the first aspect of the present invention discloses a method for optimizing the assembly process of an intelligent sleeping bed frame based on machine vision, comprising the following steps:

[0051] S102: obtaining a preset equipment process scheme for the smart sleeping bed frame, controlling the target assembly equipment to perform assembly production based on the preset equipment process scheme, and obtaining assembly state images of the assembled bed frame at multiple preset angles after completing an assembly cycle;

[0052] S104: selecting each pixel point in the assembly state image as a candidate point, determining each candidate point as an interest point or a non-interest point according to the number of bright spots and dark spots in a circular area of ​​each candidate point, and extracting interest points of the assembly state image at each preset angle;

[0053] S106: Determine adjacent interest points according to the Euclidean distance between each interest point, connect each adjacent interest point to form a plurality of curves, import the plurality of curves formed by connecting each adjacent interest point into a voxel grid space for singularity analysis and smoothing, and generate a three-dimensional characteristic shape diagram of the assembled bed frame according to each smoothed curve;

[0054] S108: obtaining a standard characteristic shape diagram of the assembled bed frame, calculating the overlap degree of the overlapping surfaces between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram, and determining the assembly state of the assembled bed frame according to the overlap degree;

[0055] S110: If the assembly status of the assembled bed frame is a defective assembly product, the defective assembly area of ​​the assembled bed frame is obtained, the cause of the defective assembly is traced and analyzed according to the defective assembly area, and an equipment process scheme optimization suggestion is generated according to the analysis result;

[0056] S112: If the assembly status of the assembled bed frame is a good assembly product, the target assembly equipment is controlled to execute the next assembly cycle. If the assembly status of all assembled bed frames is a good assembly product after the target assembly equipment executes the preset assembly cycles, it means that the preset equipment process plan meets the mass production conditions, and there is no need to optimize the preset equipment process plan.

[0057] The preset equipment process plan for the smart sleeping bed frame refers to a series of standardized processes and technical specifications pre-established by relevant technical personnel during the design and manufacturing of the smart sleeping bed frame, including the assembly sequence, assembly method, quality inspection standards and corresponding adjustment and optimization measures of the bed frame components. The preset equipment process plan usually involves the mechanical structure design, material selection, manufacturing process, and subsequent inspection and testing processes of the bed frame to achieve efficient production and quality control of the bed frame.

[0058] It should be noted that the target assembly equipment is obtained and controlled to produce according to the preset equipment process plan, and then the assembly state images of the assembled bed frame are obtained at multiple preset angles. Through machine vision technology, these images are feature extracted, and the key points of interest at each preset angle are identified and marked, so as to construct a three-dimensional feature shape map of the assembled bed frame. Next, by comparing the three-dimensional feature shape map of the assembled bed frame with the pre-set standard feature shape map, it is analyzed and judged whether the assembled bed frame meets the quality standards. If the judgment result is a defective assembly product, the process will automatically locate and mark the poor assembly area, and through in-depth analysis of these areas, find out the cause of the poor assembly, and then put forward targeted process optimization suggestions. This process realizes real-time monitoring and dynamic optimization of the assembly process, and effectively improves product quality. On the contrary, if all assembled bed frames are judged to be good assembly products, this indicates that the current preset equipment process plan performs well in mass production and no further optimization is required. The entire process uses automation and intelligent means to evaluate and analyze the preset equipment process solutions, which can effectively improve the product quality control capabilities of the bed frame in the actual production process, and provide strong support for the efficient and high-quality mass production of smart sleeping bed frames.

[0059] Preferably, if Figure 2 As shown, each pixel point is selected as a candidate point in the assembly state image, and each candidate point is determined as an interest point or a non-interest point according to the number of bright spots and dark spots in the circular area of ​​each candidate point, and the interest points of the assembly state image at each preset angle are extracted, specifically:

[0060] S202: graying and denoising the assembly state image, randomly selecting a pixel point in the assembly state image as a candidate point, and obtaining a brightness value of the candidate point;

[0061] S204: defining a circular area with a preset radius around the candidate point, retrieving the remaining pixels in the circular area, and obtaining the brightness values ​​of the remaining pixels;

[0062] S206: respectively comparing the brightness value of each remaining pixel with the brightness value of the candidate point; if the brightness value of a remaining pixel is greater than the brightness value of the candidate point, marking the remaining pixel as a bright spot; if the brightness value of a remaining pixel is not greater than the brightness value of the candidate point, marking the remaining pixel as a dark spot;

[0063] S208: Statistically analyzing the number of bright spots and dark spots in the circular area of ​​the candidate point, if the number of bright spots or dark spots in the circular area of ​​the candidate point is greater than a preset value, marking the candidate point as an interest point; otherwise, marking the candidate point as a non-interest point;

[0064] S210: Select the next pixel point as a candidate point, repeat steps S204 to S208 until all the pixels are traversed, and output a list of interest points in the assembly state image.

[0065] It should be noted that the assembly state image is grayscaled and denoised to enhance image features and reduce noise interference. Next, a pixel point is randomly selected in the image as a candidate point, and its brightness value is obtained. Then, a circular field is defined around the candidate point, and other pixels in the field are retrieved and their brightness values ​​are obtained. By comparing the brightness values ​​of the candidate point with other pixels, pixels with brightness values ​​greater than the candidate point are marked as bright spots, and pixels with brightness values ​​not greater than the candidate point are marked as dark spots. This step helps to identify significant change areas in the image, which may be related to assembly quality. Subsequently, the number of bright spots and dark spots in the circular field of the candidate point is statistically analyzed. If the number of bright spots or dark spots exceeds a preset threshold, the candidate point is marked as a point of interest, otherwise it is marked as a point of non-interest. Finally, steps S204 to S208 are repeated to traverse all pixels, and finally a list of points of interest in the assembly state image is output. This method can efficiently identify points of potential importance from the image, providing key information for subsequent quality analysis and process optimization.

[0066] Preferably, adjacent interest points are determined according to the Euclidean distance between each interest point, each adjacent interest point is connected to form a plurality of curves, the plurality of curves formed by connecting each adjacent interest point is introduced into the voxel grid space for singularity analysis and smoothing, and a three-dimensional characteristic shape diagram of the assembled bed frame is generated according to each smoothed curve, specifically:

[0067] S302: Obtaining coordinate values ​​of interest points in corresponding images of assembly status images at various preset angles, and obtaining calibration parameters of the camera;

[0068] S304: constructing a Cartesian coordinate system, and converting the coordinates of the interest points in the assembly state images of each preset angle into the Cartesian coordinate system according to the calibration parameters of the camera, so as to ensure that the coordinates of the interest points in all the assembly state images of the preset angles are in the same coordinate system;

[0069] S306: calculating the Euclidean distance between each interest point in the Cartesian coordinate system, marking the interest point with the shortest Euclidean distance as an adjacent interest point, and connecting each adjacent interest point to form a plurality of curves;

[0070] S308: construct a voxel grid space, import a number of curves formed by connecting adjacent interest points into the voxel grid space, obtain the intersection points of each curve and the voxel grid, define them as second-order difference nodes, and obtain the second-order difference nodes of each curve;

[0071] S310: performing a second-order difference operation on the second-order difference nodes of each curve to obtain a difference value of each curve, and constructing a difference value matrix according to the difference values ​​of each curve; wherein the rows in the matrix represent each curve, the columns represent the positions of the second-order difference nodes, and each element represents the difference value of the corresponding curve at the corresponding second-order difference node;

[0072] S312: traverse each element in the difference value matrix, compare each element in the difference value matrix with a preset threshold, and if an element in the difference value matrix is ​​greater than the preset threshold, mark the curve node corresponding to the element in the difference value matrix as a singular position node;

[0073] S314: repeating step S312 until all elements in the differential value matrix are traversed, obtaining the singular position nodes of each curve, and performing smoothing on the singular position nodes of each curve to obtain each smoothed curve;

[0074] S316: Generate a three-dimensional characteristic shape diagram of the assembled bed frame according to each smoothed curve.

[0075] It should be noted that the coordinate values ​​of the points of interest are extracted from the assembly state images at each preset angle, and combined with the calibration parameters of the camera, the coordinates of the points of interest in all images are converted into a unified Cartesian coordinate system to achieve coordinate consistency across images. Subsequently, the Euclidean distance between these points of interest is calculated, the points with the closest distance are paired to form adjacent points of interest, and these adjacent points are connected to form curves. These curves are further processed in the voxel grid space, and the second-order difference nodes are defined by the intersection with the voxel grid, so as to obtain the second-order difference information of the curve. By performing second-order difference operations on these second-order difference nodes, a difference value matrix is ​​constructed, in which the rows represent the curves, the columns represent the node positions, and the elements represent the difference values. This matrix is ​​used to analyze the local variation characteristics of the curve. By setting a threshold, the elements in the difference value matrix are traversed and compared, and the curve nodes corresponding to the elements exceeding the threshold are marked as singular position nodes, which may represent abnormalities or key feature points in the assembly. Finally, these singular position nodes are smoothed to reduce the influence of noise and outliers, and a smoother and more accurate curve is obtained. Based on these smoothed curves, a three-dimensional characteristic shape diagram of the assembled bed frame is generated, providing an intuitive and quantitative basis for subsequent quality evaluation and process optimization. Through second-order difference operations and threshold comparisons, anomalies or key feature points in the assembly process can be effectively identified, which is helpful for quality control. The entire process reduces manual intervention and improves the level of automation. At the same time, the use of machine vision and data analysis technology improves the degree of intelligence. In summary, this method based on machine vision and geometric analysis provides an efficient and accurate technical means for optimizing the assembly process of intelligent bed frames, which helps to improve product quality and production efficiency.

[0076] Preferably, the overlap degree of the mutually overlapping surfaces between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is calculated, and the assembly state of the assembled bed frame is determined according to the overlap degree, specifically:

[0077] Aligning the three-dimensional feature shape image with the standard feature shape image based on the SURF algorithm, and after the alignment is completed, obtaining the overlapping area value of the overlapping surface between the three-dimensional feature shape image and the standard feature shape image;

[0078] Calculate the total area value of each face in the standard characteristic shape image, divide the overlapping area value by the total area value, and obtain the degree of overlap between the three-dimensional characteristic shape image and the standard characteristic shape image;

[0079] If the overlap between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is greater than a preset overlap, the assembled bed frame is marked as a good assembly product;

[0080] If the degree of overlap between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is not greater than the preset degree of overlap, the assembled bed frame is marked as a defective assembly product, and the area of ​​the non-overlapping surface between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is obtained and defined as the defective assembly area.

[0081] It should be noted that the SURF (Speeded Robust Features) algorithm is used to process the three-dimensional feature shape map of the smart sleeping bed frame so as to align it with the pre-set standard feature shape map. Through the SURF algorithm, the system can effectively identify and match the key feature points in the two shape maps, thereby achieving accurate alignment. After the alignment is completed, the system calculates the area of ​​the overlapping area between the two, and this area value reflects the consistency of the two in terms of shape, size, etc. Next, the system calculates the total area value of each face in the standard feature shape map, and divides the overlapping area value by the total area value to obtain the overlap. The overlap is an indicator to measure the degree of matching between the assembled bed frame and the standard shape, which reflects the quality of assembly. If the overlap is greater than the preset overlap threshold, it indicates that the assembled bed frame is highly consistent with the standard shape, and the system marks it as "good assembly". On the contrary, if the overlap is lower than or equal to the preset threshold, it means that there is a large difference between the assembled bed frame and the standard shape, and the system marks it as "bad assembly", and further determines the non-overlapping area, which is considered to be a poorly assembled part. For bed frames marked as "defective assembly products", clearly pointing out the non-overlapping areas will help to quickly locate the problem and facilitate subsequent process optimization work.

[0082] Preferably, if the assembly state of the assembled bed frame is a defective assembly product, the defective assembly area of ​​the assembled bed frame is obtained, and the cause of the defective assembly is traced and analyzed according to the defective assembly area, and an optimization suggestion for the equipment process solution is generated according to the analysis result, specifically:

[0083] If the assembly state of the assembled bed frame is a defective assembly product, obtaining the defective assembly area of ​​the assembled bed frame and obtaining the work log of the target assembly equipment;

[0084] According to the work log, various actual assembly parameters at various preset assembly time nodes when the target assembly equipment assembles the poorly assembled area are acquired to obtain various actual assembly parameter sets;

[0085] and obtaining a preset equipment process scheme for the smart sleeping bed frame, and obtaining various preset assembly parameters at various preset assembly time nodes when the target assembly equipment assembles the poorly assembled area according to the preset equipment process scheme, to obtain various preset assembly parameter sets;

[0086] A dynamic time warping algorithm is introduced to calculate the similarity between each actual assembly parameter set and the corresponding preset assembly parameter set based on the dynamic time warping algorithm;

[0087] If the similarities between the actual assembly parameter sets and the corresponding preset assembly parameter sets when the target assembly equipment assembles the poor assembly area are greater than the preset similarity threshold, it means that the assembly parameters of the target assembly equipment when assembling the poor assembly area are normal, which means that the production of defective assembly products is caused by defects in the equipment process plan itself, and then an equipment process plan optimization suggestion is generated based on the poor assembly area.

[0088] It should be noted that when a defective bed frame assembly is detected, the system will automatically obtain the defective assembly area of ​​the bed frame and its related work log. The work log contains detailed parameter records of the target assembly equipment when operating in the defective assembly area. These parameters usually include but are not limited to speed, pressure, temperature, etc. At the same time, the system will also call the pre-set assembly process plan, that is, under ideal conditions, the parameter set that the target assembly equipment should follow at the same assembly node. Next, by introducing the dynamic time warping (DTW) algorithm, the system can compare the actual assembly parameter set with the preset assembly parameter set and calculate the similarity between the two. The DTW algorithm allows nonlinear alignment on the time axis, so that even if the parameter sequence deviates in time, it can be effectively compared, which is particularly suitable for processing time delays or advances that may occur during the assembly process. If the similarity between the actual assembly parameter set and the preset assembly parameter set is higher than the preset threshold, it indicates that the operating parameters of the target assembly equipment in the defective assembly area are basically in line with expectations, and there is no obvious parameter error. At this point, the system will further analyze and determine whether the cause of the defective assembly product is the defect in the assembly process plan itself, rather than improper equipment operation. If it is confirmed that the problem is with the process solution, the system will generate corresponding optimization suggestions to improve the process flow and reduce or avoid the production of defective assembly products in the future.

[0089] The following steps are also included:

[0090] If there is a situation where the similarity between at least one actual assembly parameter set and the corresponding preset assembly parameter set is not greater than a preset similarity threshold when the target assembly equipment assembles the poorly assembled area, then obtaining the actual assembly parameter set corresponding to the similarity not greater than the preset similarity threshold;

[0091] Determine abnormal assembly parameters in the target assembly equipment according to the actual assembly parameter set corresponding to the similarity not greater than the preset similarity threshold, and perform correction processing on the abnormal assembly parameters;

[0092] After the correction processing of the abnormal assembly parameters is completed, the target assembly equipment is controlled to execute the next assembly cycle, and the assembly state of the assembled bed frame assembled after the execution of the next assembly cycle is obtained; if the assembly state of the assembled bed frame assembled after the execution of the next assembly cycle is a good assembly product, it means that the generation of defective assembly products is caused by the drift of the assembly parameters of the target assembly equipment, and there is no need to optimize the equipment process plan at this time;

[0093] If the assembly status of the assembled bed frame obtained after executing the next assembly cycle is still a defective assembly product, it means that the generation of the defective assembly product is also caused by defects in the equipment process solution itself, and then an equipment process solution optimization suggestion is generated based on the defective assembly area.

[0094] It should be noted that if the target assembly equipment assembles the poorly assembled area and there is a situation where the similarity between at least one actual assembly parameter set and the corresponding preset assembly parameter set is not greater than the preset similarity threshold, the system will identify these actual assembly parameter sets with insufficient similarity and further determine the abnormal assembly parameters that cause the assembly failure. These abnormal parameters will be corrected in order to restore them to the preset standard range. After the correction is completed, the system controls the assembly equipment to enter the next assembly cycle and re-acquires the assembly state of the bed frame after assembly. If the assembly result of the new cycle is a good product, this indicates that the previous assembly defective product was caused by the temporary drift of the equipment parameters, rather than the defects of the process solution itself, so there is no need to optimize the process solution. On the contrary, if the assembly result of the new cycle is still a defective product, this suggests that there are defects in the assembly process solution itself, and it is necessary to generate corresponding process solution optimization suggestions based on the bad area. It can be seen that for defective products caused by defects in the process solution, the system can provide targeted optimization suggestions, which helps to improve the process level and product quality. In summary, this method provides an effective fault diagnosis and optimization means for the assembly process of the smart sleeping bed frame through real-time monitoring, rapid response and data-driven decision-making, which helps to improve product quality and production efficiency.

[0095] The intelligent sleeping bed frame assembly process optimization method also includes the following steps:

[0096] Acquire user evaluation information on the smart sleeping bed frame, perform feature extraction on the evaluation information, obtain feature data of users' negative reviews on various parts of the smart sleeping bed frame, and construct a negative review feature heat map of the smart sleeping bed frame based on the feature data of users' negative reviews on various parts of the smart sleeping bed frame;

[0097] Acquire production defect data of the target assembly equipment during the actual assembly production process, and construct a defect feature heat map of the smart sleeping bed frame according to the production defect data;

[0098] Calculating the cosine similarity between the negative review feature heat map and the defect feature heat map based on the cosine similarity algorithm;

[0099] If the cosine similarity between the negative review feature heat map and the defect feature heat map is greater than a preset similarity threshold, extract the negative review parts of the smart sleeping bed frame from the negative review feature heat map, and obtain each sub-assembly module in the target assembly device;

[0100] Based on the grey correlation analysis method, the correlation between the negatively reviewed parts of the smart bed frame by users and the sub-assembly modules in the target assembly equipment is analyzed, and only the sub-assembly modules corresponding to the correlation greater than the preset correlation are extracted and defined as the sub-assembly modules to be analyzed;

[0101] Acquire the real-time operating parameters of the subassembly module to be analyzed, introduce the Markov chain Monte Carlo method and calculate the fault state transition probability of the subassembly module to be analyzed in combination with the real-time operating parameters of the subassembly module to be analyzed, and obtain the fault state transition probability of the subassembly module to be analyzed;

[0102] If the fault state transition probability of the sub-assembly module to be analyzed is greater than a preset probability value, the target assembly equipment is controlled to stop production and fault warning information is generated.

[0103] It should be noted that by collecting and analyzing the user's evaluation information on the smart sleep bed frame and the defect data found in the actual assembly production process, possible problem areas can be identified and located. First, by extracting the features of the user's evaluation information, the negative review feature data of each part of the smart sleep bed frame can be obtained, and then a negative review feature heat map is constructed, which intuitively shows the most frequently mentioned problem areas in user feedback. Similarly, by analyzing the defect data in the production process, a defect feature heat map is constructed to reflect the quality problems that may occur in actual production. Then, the cosine similarity algorithm is used to compare the two heat maps. If the similarity between the two exceeds the preset threshold, it means that the problem area reported by the user is highly correlated with the defects in the production process. This step helps to quickly locate the potential source of quality problems. Subsequently, the gray correlation analysis method is used to evaluate the degree of correlation between the problem area reported by the user and the various sub-assembly modules in the production process, and only the sub-assembly modules with a correlation higher than the preset threshold are selected as the objects to be analyzed. These sub-assembly modules may be the main causes of the problem. For these sub-assembly modules to be analyzed, the system further collects their real-time operating parameters and uses the Markov chain Monte Carlo method to predict their fault state transition probabilities. This probabilistic model can estimate the probability of a subassembly module changing from a normal state to a faulty state. Finally, if the fault state transition probability of a subassembly module exceeds the preset probability threshold, it means that the subassembly module has a high risk of failure. At this time, the system will automatically control the target assembly equipment to stop production to avoid potential production accidents, and generate fault warning information to prompt relevant personnel to take measures to solve the problem. Overall, this method integrates user feedback, production data and prediction models to achieve early identification and prevention of quality problems of smart sleeping bed frames, effectively improve the efficiency and accuracy of product quality control, and also ensure the safety of the production process.

[0104] The intelligent sleeping bed frame assembly process optimization method also includes the following steps:

[0105] Based on the big data network, the reliability index of each sub-assembly module in the target assembly equipment is obtained under various preset working steps and various preset production environment parameter conditions.

[0106] Obtaining the actual working step length of the target assembly equipment and obtaining the actual production environment parameters of the target assembly equipment;

[0107] Calculating mutual information values ​​between the actual working step length and the actual production environment parameter and various preset working step lengths and various preset production environment parameter condition combinations;

[0108] Obtain a preset working step length and a preset production environment parameter condition combination with a maximum mutual information value, and obtain a real-time reliability index of the target assembly equipment according to the preset working step length and the preset production environment parameter condition combination with a maximum mutual information value;

[0109] If the real-time reliability index is greater than a preset index threshold, the target assembly equipment is controlled to shut down production.

[0110] It should be noted that, firstly, through big data network technology, the system collects and analyzes the reliability index of each submodule of the target assembly equipment under different preset working steps and production environment parameters. This comprehensive data collection and analysis enables the performance of the equipment under different operating conditions to be quantified, thus providing a basis for subsequent decision-making. Next, the system obtains the actual working step and production environment parameters of the target assembly equipment, which is a key step in combining the theoretical model with actual operation. By comparing the actual operating conditions with the preset conditions, the system calculates the mutual information value between the two. The mutual information value reflects the degree of consistency between the actual operating conditions and the preset conditions. The larger the value, the higher the consistency, that is, the greater the possibility that the performance of the equipment in actual operation is consistent with expectations. Based on the preset working step and production environment parameter combination with the maximum mutual information value, the system calculates the real-time reliability index of the target assembly equipment. This index directly reflects the current operating status of the equipment and provides managers with a real-time and accurate assessment of the health status of the equipment. Finally, when the real-time reliability index exceeds the preset index threshold, the system automatically controls the target assembly equipment to stop production. This mechanism is designed to prevent the equipment from continuing to operate under potential high-risk conditions, prevent possible failures or accidents, and protect the safety of equipment, personnel and production processes. In general, this method achieves efficient management and prediction of the reliability of target assembly equipment through big data analysis, information theory and real-time monitoring, effectively improves the stability and safety of equipment operation, and helps optimize the production process and reduce losses caused by unplanned downtime.

[0111] like Figure 3 As shown, the second aspect of the present invention discloses a smart sleeping bed frame assembly process optimization system based on machine vision, the smart sleeping bed frame assembly process optimization system includes a memory 80 and a processor 100, the memory 80 stores a smart sleeping bed frame assembly process optimization method program, when the smart sleeping bed frame assembly process optimization method program is executed by the processor 100, any step of the smart sleeping bed frame assembly process optimization method is implemented.

[0112] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for optimizing the assembly process of an intelligent sleeping bed frame based on machine vision, characterized in that: The following steps are involved: Obtaining a preset equipment process plan for the smart sleeping bed frame, controlling the target assembly equipment to perform assembly production based on the preset equipment process plan, and obtaining assembly state images of the assembled bed frame at multiple preset angles after completing an assembly cycle; Selecting each pixel point in the assembly state image as a candidate point, determining each candidate point as an interest point or a non-interest point according to the number of bright spots and dark spots in a circular area of ​​each candidate point, and extracting the interest points of the assembly state image at each preset angle; Determine adjacent interest points according to the Euclidean distance between each interest point, connect each adjacent interest point to form a plurality of curves, import the plurality of curves formed by connecting each adjacent interest point into the voxel grid space for singularity analysis and smoothing, and generate a three-dimensional characteristic shape diagram of the assembled bed frame according to each smoothed curve; Obtaining a standard characteristic shape diagram of the assembled bed frame, calculating the overlap degree of the mutually overlapping surfaces between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram, and determining the assembly state of the assembled bed frame according to the overlap degree; If the assembly status of the assembled bed frame is a defective assembly product, the defective assembly area of ​​the assembled bed frame is obtained, the cause of the defective assembly is traced and analyzed according to the defective assembly area, and an optimization suggestion for the equipment process plan is generated according to the analysis result; If the assembly status of the assembled bed frame is a good assembly product, the target assembly equipment is controlled to execute the next assembly cycle. If the assembly status of all assembled bed frames is a good assembly product after the target assembly equipment executes the preset assembly cycles, it means that the preset equipment process plan meets the mass production conditions, and there is no need to optimize the preset equipment process plan.

2. According to the machine vision-based intelligent sleeping bed frame assembly process optimization method of claim 1, it is characterized in that: In the assembly state image, each pixel point is selected as a candidate point, and each candidate point is determined as an interest point or a non-interest point according to the number of bright spots and dark spots in the circular area of ​​each candidate point, and the interest points of the assembly state image at each preset angle are extracted, specifically: S202: graying and denoising the assembly state image, randomly selecting a pixel point in the assembly state image as a candidate point, and obtaining a brightness value of the candidate point; S204: defining a circular area with a preset radius around the candidate point, retrieving the remaining pixels in the circular area, and obtaining the brightness values ​​of the remaining pixels; S206: Compare the brightness value of each of the remaining pixels with the brightness value of the candidate point respectively; If the brightness value of any other pixel is greater than the brightness value of the candidate point, the other pixel is marked as a bright spot; If the brightness value of a certain remaining pixel is not greater than the brightness value of the candidate point, the remaining pixel is marked as a dark point; S208: Statistically analyzing the number of bright spots and dark spots in the circular area of ​​the candidate point, if the number of bright spots or dark spots in the circular area of ​​the candidate point is greater than a preset value, marking the candidate point as an interest point; otherwise, marking the candidate point as a non-interest point; S210: Select the next pixel point as a candidate point, repeat steps S204 to S208 until all the pixel points are traversed, and output a list of interest points in the assembly state image.

3. The method for optimizing the assembly process of an intelligent sleeping bed frame based on machine vision according to claim 1, characterized in that: Adjacent interest points are determined according to the Euclidean distance between each interest point, and each adjacent interest point is connected to form a number of curves. The curves formed by connecting each adjacent interest point are imported into the voxel grid space for singular analysis and smoothing. The three-dimensional characteristic shape diagram of the assembled bed frame is generated according to each smoothed curve, which is specifically: S302: Obtaining the coordinate values ​​of the points of interest in the corresponding images of the assembly state images at each preset angle, and obtaining the calibration parameters of the camera; S304: constructing a Cartesian coordinate system, and converting the coordinates of the interest points in the assembly state images of each preset angle into the Cartesian coordinate system according to the calibration parameters of the camera, so as to ensure that the coordinates of the interest points in all the assembly state images of the preset angles are in the same coordinate system; S306: calculating the Euclidean distance between each interest point in the Cartesian coordinate system, marking the interest point with the shortest Euclidean distance as an adjacent interest point, and connecting each adjacent interest point to form a plurality of curves; S308: construct a voxel grid space, import a number of curves formed by connecting adjacent interest points into the voxel grid space, obtain the intersection points of each curve and the voxel grid, define them as second-order difference nodes, and obtain the second-order difference nodes of each curve; S310: performing a second-order difference operation on the second-order difference nodes of each curve to obtain a difference value of each curve, and constructing a difference value matrix according to the difference values ​​of each curve; wherein the rows in the matrix represent each curve, the columns represent the positions of the second-order difference nodes, and each element represents the difference value of the corresponding curve at the corresponding second-order difference node; S312: traverse each element in the difference value matrix, compare each element in the difference value matrix with a preset threshold, and if an element in the difference value matrix is ​​greater than the preset threshold, mark the curve node corresponding to the element in the difference value matrix as a singular position node; S314: repeating step S312 until all elements in the differential value matrix are traversed, obtaining the singular position nodes of each curve, and performing smoothing on the singular position nodes of each curve to obtain each smoothed curve; S316: Generate a three-dimensional characteristic shape diagram of the assembled bed frame according to each smoothed curve.

4. The method for optimizing the assembly process of an intelligent sleeping bed frame based on machine vision according to claim 1, characterized in that: The overlap degree of the mutually overlapping surfaces between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is calculated, and the assembly state of the assembled bed frame is determined according to the overlap degree, specifically: Aligning the three-dimensional feature shape image with the standard feature shape image based on the SURF algorithm, and after the alignment is completed, obtaining the overlapping area value of the overlapping surface between the three-dimensional feature shape image and the standard feature shape image; Calculate the total area value of each face in the standard characteristic shape image, divide the overlapping area value by the total area value, and obtain the degree of overlap between the three-dimensional characteristic shape image and the standard characteristic shape image; If the overlap between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is greater than a preset overlap, the assembled bed frame is marked as a good assembly product; If the degree of overlap between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is not greater than the preset degree of overlap, the assembled bed frame is marked as a defective assembly product, and the area of ​​the non-overlapping surface between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is obtained and defined as the defective assembly area.

5. The method for optimizing the assembly process of an intelligent sleeping bed frame based on machine vision according to claim 1, characterized in that: If the assembly status of the assembled bed frame is a defective assembly, the defective assembly area of ​​the assembled bed frame is obtained, and the cause of the defective assembly is traced and analyzed according to the defective assembly area. The equipment process optimization proposal is generated according to the analysis result, which is specifically as follows: If the assembly state of the assembled bed frame is a defective assembly product, obtaining the defective assembly area of ​​the assembled bed frame and obtaining the work log of the target assembly equipment; According to the work log, various actual assembly parameters at various preset assembly time nodes when the target assembly equipment assembles the poorly assembled area are acquired to obtain various actual assembly parameter sets; and obtaining a preset equipment process scheme for the smart sleeping bed frame, and obtaining various preset assembly parameters at various preset assembly time nodes when the target assembly equipment assembles the poorly assembled area according to the preset equipment process scheme, to obtain various preset assembly parameter sets; A dynamic time warping algorithm is introduced to calculate the similarity between each actual assembly parameter set and the corresponding preset assembly parameter set based on the dynamic time warping algorithm; If the similarities between the actual assembly parameter sets and the corresponding preset assembly parameter sets when the target assembly equipment assembles the poor assembly area are greater than the preset similarity threshold, it means that the assembly parameters of the target assembly equipment when assembling the poor assembly area are normal, which means that the production of defective assembly products is caused by defects in the equipment process plan itself, and then an equipment process plan optimization suggestion is generated based on the poor assembly area.

6. The method for optimizing the assembly process of an intelligent sleeping bed frame based on machine vision according to claim 5 further comprises the following steps: If there is a situation where the similarity between at least one actual assembly parameter set and the corresponding preset assembly parameter set is not greater than a preset similarity threshold when the target assembly equipment assembles the poorly assembled area, then obtaining the actual assembly parameter set corresponding to the similarity not greater than the preset similarity threshold; Determine abnormal assembly parameters in the target assembly equipment according to the actual assembly parameter set corresponding to the similarity not greater than the preset similarity threshold, and perform correction processing on the abnormal assembly parameters; After the correction process of the abnormal assembly parameters is completed, the target assembly equipment is controlled to execute the next assembly cycle, and the assembly state of the assembled bed frame obtained after the execution of the next assembly cycle is obtained; If the assembly status of the assembled bed frame obtained after the next assembly cycle is good assembly, it means that the generation of defective assembly products is caused by the drift of assembly parameters of the target assembly equipment, and there is no need to optimize the equipment process plan at this time; If the assembly status of the assembled bed frame obtained after executing the next assembly cycle is still a defective assembly product, it means that the generation of the defective assembly product is also caused by defects in the equipment process solution itself, and then an equipment process solution optimization suggestion is generated based on the defective assembly area.

7. An intelligent sleeping bed frame assembly process optimization system based on machine vision, characterized in that: The intelligent sleeping bed frame assembly process optimization system includes a memory and a processor, wherein the memory stores an intelligent sleeping bed frame assembly process optimization method program, and when the intelligent sleeping bed frame assembly process optimization method program is executed by the processor, the following steps are implemented: Obtaining a preset equipment process plan for the smart sleeping bed frame, controlling the target assembly equipment to perform assembly production based on the preset equipment process plan, and obtaining assembly state images of the assembled bed frame at multiple preset angles after completing an assembly cycle; Selecting each pixel point in the assembly state image as a candidate point, determining each candidate point as an interest point or a non-interest point according to the number of bright spots and dark spots in a circular area of ​​each candidate point, and extracting the interest points of the assembly state image at each preset angle; Determine adjacent interest points according to the Euclidean distance between each interest point, connect each adjacent interest point to form a plurality of curves, import the plurality of curves formed by connecting each adjacent interest point into the voxel grid space for singularity analysis and smoothing, and generate a three-dimensional characteristic shape diagram of the assembled bed frame according to each smoothed curve; Obtaining a standard characteristic shape diagram of the assembled bed frame, calculating the overlap degree of the mutually overlapping surfaces between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram, and determining the assembly state of the assembled bed frame according to the overlap degree; If the assembly status of the assembled bed frame is a defective assembly product, the defective assembly area of ​​the assembled bed frame is obtained, the cause of the defective assembly is traced and analyzed according to the defective assembly area, and an optimization suggestion for the equipment process plan is generated according to the analysis result; If the assembly status of the assembled bed frame is a good assembly product, the target assembly equipment is controlled to execute the next assembly cycle. If the assembly status of all assembled bed frames is a good assembly product after the target assembly equipment executes the preset assembly cycles, it means that the preset equipment process plan meets the mass production conditions, and there is no need to optimize the preset equipment process plan.

8. The intelligent sleeping bed frame assembly process optimization system based on machine vision according to claim 7 is characterized in that: The overlap degree of the mutually overlapping surfaces between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is calculated, and the assembly state of the assembled bed frame is determined according to the overlap degree, specifically: Aligning the three-dimensional feature shape image with the standard feature shape image based on the SURF algorithm, and after the alignment is completed, obtaining the overlapping area value of the overlapping surface between the three-dimensional feature shape image and the standard feature shape image; Calculate the total area value of each face in the standard characteristic shape image, divide the overlapping area value by the total area value, and obtain the degree of overlap between the three-dimensional characteristic shape image and the standard characteristic shape image; If the overlap between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is greater than a preset overlap, the assembled bed frame is marked as a good assembly product; If the degree of overlap between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is not greater than the preset degree of overlap, the assembled bed frame is marked as a defective assembly product, and the area of ​​the non-overlapping surface between the three-dimensional characteristic shape diagram and the standard characteristic shape diagram is obtained and defined as the defective assembly area.

9. The intelligent sleeping bed frame assembly process optimization system based on machine vision according to claim 7, characterized in that: If the assembly status of the assembled bed frame is a defective assembly, the defective assembly area of ​​the assembled bed frame is obtained, and the cause of the defective assembly is traced and analyzed according to the defective assembly area. The equipment process optimization proposal is generated according to the analysis result, which is specifically as follows: If the assembly state of the assembled bed frame is a defective assembly product, obtaining the defective assembly area of ​​the assembled bed frame and obtaining the work log of the target assembly equipment; According to the work log, various actual assembly parameters at various preset assembly time nodes when the target assembly equipment assembles the poorly assembled area are acquired to obtain various actual assembly parameter sets; and obtaining a preset equipment process scheme for the smart sleeping bed frame, and obtaining various preset assembly parameters at various preset assembly time nodes when the target assembly equipment assembles the poorly assembled area according to the preset equipment process scheme, to obtain various preset assembly parameter sets; A dynamic time warping algorithm is introduced to calculate the similarity between each actual assembly parameter set and the corresponding preset assembly parameter set based on the dynamic time warping algorithm; If the similarities between the actual assembly parameter sets and the corresponding preset assembly parameter sets when the target assembly equipment assembles the poor assembly area are greater than the preset similarity threshold, it means that the assembly parameters of the target assembly equipment when assembling the poor assembly area are normal, which means that the production of defective assembly products is caused by defects in the equipment process plan itself, and then an equipment process plan optimization suggestion is generated based on the poor assembly area.

10. The intelligent sleeping bed frame assembly process optimization system based on machine vision according to claim 9 further comprises the following steps: If there is a situation where the similarity between at least one actual assembly parameter set and the corresponding preset assembly parameter set is not greater than a preset similarity threshold when the target assembly equipment assembles the poorly assembled area, then obtaining the actual assembly parameter set corresponding to the similarity not greater than the preset similarity threshold; Determine abnormal assembly parameters in the target assembly equipment according to the actual assembly parameter set corresponding to the similarity not greater than the preset similarity threshold, and perform correction processing on the abnormal assembly parameters; After the correction process of the abnormal assembly parameters is completed, the target assembly equipment is controlled to execute the next assembly cycle, and the assembly state of the assembled bed frame obtained after the execution of the next assembly cycle is obtained; If the assembly status of the assembled bed frame obtained after the next assembly cycle is good assembly, it means that the generation of defective assembly products is caused by the drift of assembly parameters of the target assembly equipment, and there is no need to optimize the equipment process plan at this time; If the assembly status of the assembled bed frame obtained after executing the next assembly cycle is still a defective assembly product, it means that the generation of the defective assembly product is also caused by defects in the equipment process solution itself, and then an equipment process solution optimization suggestion is generated based on the defective assembly area.

Images