Intelligent solder paste adding method for fully automatic visual printing machine based on visual interactive recognition

Abstract

The present invention relates to the field of visual recognition, and discloses a method for intelligently adding solder paste to a fully automatic visual printer based on visual interactive recognition, comprising the following steps: collecting visual images in a solder paste working area, and performing feature extraction and analysis on the solder paste in the image, determining the distribution of the solder paste in combination with the analysis results, and classifying the solder paste storage area, and then performing intelligent addition processing on the solder paste in combination with the real-time weight data of the solder paste, and finally generating a solder paste distribution sequence table for a solder paste spot welding head. The present invention can perform visual analysis on the solder paste in the solder paste storage area in the fully automatic visual printer, and generate a solder paste addition plan based on the visual analysis results, thereby preventing the printer from being unable to maintain normal printing due to a small amount of solder paste.

Description

Technical Field

[0001] The present invention relates to the field of visual recognition, and in particular to a method for intelligently adding solder paste to a fully automatic visual printer based on visual interactive recognition. Background Art

[0002] The fully automatic visual printing machine uses computer vision technology and automatic control technology to realize the automation and precision of the printing process, greatly improving production efficiency and printing quality. It captures the image of the printed product through a digital camera, processes and analyzes the image, determines the printing position and color, controls the amount of ink on the printing plate and the printing pressure, and achieves precise printing. The equipment is widely used in packaging, labels, books, advertising and other fields. It can print products of various materials and shapes, such as paper, plastic film, glass bottles, metal plates, etc., with high flexibility and adaptability.

[0003] In a fully automatic visual printer, solder paste is a key material used to form solder joints between electronic components and printed circuit boards. During the soldering process, the solder paste melts at the solder joints and forms a stable solder connection after cooling, fixing the electronic components on the circuit board. The active flux in the solder paste can effectively reduce oxidation on the metal surface and promote wetting during soldering. Since the amount of solder paste decreases during the spot welding process in the printing field, it is necessary to intelligently add solder paste to achieve the continuity of the solder connection. Therefore, an intelligent solder paste adding method for a fully automatic visual printer based on visual interactive recognition is proposed. Summary of the invention

[0004] The present invention overcomes the deficiencies of the prior art and provides a method for intelligently adding solder paste to a fully automatic visual printer based on visual interactive recognition.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] The first aspect of the present invention provides a method for intelligently adding solder paste to a fully automatic visual printer based on visual interactive recognition, comprising the following steps:

[0007] Collect visual images of the solder paste working area in the fully automatic visual printer, perform feature analysis on the collected visual images, determine the distribution of solder paste in the solder paste working area, and classify the solder paste storage area;

[0008] Based on the real-time distribution map of solder paste and the real-time weight data of solder paste, the solder paste adding time node is determined and the solder paste adding process is performed intelligently in the solder paste storage area;

[0009] After solder paste is added to the solder paste storage area, a solder paste distribution sequence table of the solder paste spot welding head is generated based on the spot welding path of the solder paste spot welding head.

[0010] Furthermore, in a preferred embodiment of the present invention, the visual image of the solder paste working area is collected in the fully automatic visual printer, and the characteristics of the collected visual image are analyzed to determine the distribution of the solder paste in the solder paste working area, and the solder paste storage area is classified, specifically:

[0011] The fully automatic visual printer that needs to perform solder paste addition analysis is calibrated as the target printer, and the working area for storing solder paste in the target printer is obtained and calibrated as the solder paste working area;

[0012] In the solder paste storage area, a visual image acquisition device and a visual image analysis device are introduced, and the visual image acquisition device is used to collect the visual image of the current solder paste storage area, and the real-time visual image of the solder paste storage area is calibrated;

[0013] The Sobel operator is introduced to grayscale all pixels in the real-time visual image of the solder paste storage area. After the grayscale processing, the contour features of the solder paste in the real-time visual image of the solder paste storage area are extracted based on the Sobel operator and calibrated as solder paste contour features.

[0014] Obtain a color histogram of solder paste contour features, perform color statistics on the real-time visual image of the solder paste based on the color histogram of the solder paste contour features, and generate solder paste color features;

[0015] A SIFT algorithm is introduced, and a Gaussian convolution kernel is constructed through the SIFT algorithm, and a scale space is constructed, wherein the scale space is obtained by convolving the Gaussian convolution kernel onto the real-time visual image of the solder paste storage area, and is used for feature smoothing of solder paste contour features and solder paste color features;

[0016] In the scale space, scale-invariant feature points are extracted from the solder paste contour feature and the solder paste color feature, and the scale-invariant feature points on the solder paste contour feature and the solder paste color feature are fused to obtain the solder paste contour smoothing feature and the solder paste color smoothing feature.

[0017] The gradient direction histogram of solder paste contour smoothing features and solder paste color smoothing features is performed. Figure 2 Secondary smoothing process to obtain secondary smoothing features of solder paste contour and secondary smoothing features of solder paste color;

[0018] The secondary smooth features of solder paste contour and secondary smooth features of solder paste color are fused to obtain real-time fusion features of solder paste. The real-time fusion features of solder paste are analyzed to determine the weight data and distribution of solder paste in the solder paste storage area, and the solder paste storage areas are classified.

[0019] Furthermore, in a preferred embodiment of the present invention, the solder paste real-time fusion feature is used for feature analysis to determine the weight data and distribution of the solder paste in the solder paste storage area, and to classify the solder paste storage area, specifically:

[0020] A weight detection sensor is installed in the solder paste storage area, and the weight detection sensor is operated to calculate the real-time weight data of the solder paste in the solder paste storage area, and calibrate it as the real-time weight data of the solder paste;

[0021] A big data network is introduced, and based on the big data network, standard color features and standard contour features of solder paste are retrieved to generate standard fusion features of solder paste, and a Euclidean distance algorithm is introduced to calculate the Euclidean distance between the standard fusion features of solder paste and the real-time fusion features of solder paste, which are calibrated as a type of Euclidean distance;

[0022] A Euclidean distance threshold is preset, and if a type of Euclidean distance is maintained within the Euclidean distance threshold, the solder paste in the solder paste storage area is marked as solder paste of qualified quality; if a type of Euclidean distance is not maintained within the Euclidean distance threshold, the solder paste in the solder paste storage area is marked as solder paste of unqualified quality;

[0023] If only solder paste of qualified quality exists in the solder paste storage area, the solder paste storage area is marked as a Class I solder paste storage area; if unqualified solder paste exists in the solder paste storage area, the solder paste storage area is marked as a Class II solder paste storage area;

[0024] Calculate the pixel values ​​of different pixels in the solder paste storage area, preset a standard pixel value, mark the pixel points with pixel values ​​less than the standard pixel value as a type of pixel points, and combine all the type of pixel points to obtain a sub-area with less solder paste in the solder paste storage area, and mark it as a sub-area with less solder paste;

[0025] Calculate the pixel value difference between the pixel value of the pixel in the sub-region with less solder paste and the standard pixel value, calibrate it as the target pixel value difference, and calculate the concentration of solder paste that needs to be added to the sub-region with less solder paste based on the target pixel value difference, calibrate it as the target solder paste concentration;

[0026] Calculate the volume of the area with less solder paste, and combine it with the target solder paste concentration to generate the amount of solder paste that needs to be added to the area with less solder paste;

[0027] Based on the solder paste low amount sub-region, a schematic diagram of the real-time distribution of solder paste in the solder paste storage area is constructed and calibrated as a real-time distribution map of solder paste.

[0028] Further, in a preferred embodiment of the present invention, the solder paste adding time node determination and solder paste intelligent adding processing are performed in the solder paste storage area based on the real-time distribution map of the solder paste and the real-time weight data of the solder paste, specifically:

[0029] Obtain a solder paste source, which is a device that provides solder paste to a target printer, and connect the solder paste source to the target printer so that the target printer can add solder paste to a solder paste storage area based on the solder paste source;

[0030] Determining a position of a solder paste spot welding head in a target printer, wherein the solder paste spot welding head is connected to a solder paste storage area, and the solder paste storage area conducts solder paste to the solder paste spot welding head to implement solder paste spot welding processing;

[0031] Preset standard solder paste weight data, obtain the spot welding path of the solder paste spot welding head, and monitor in real time whether the solder paste weight data in the solder paste storage area when the solder paste spot welding head executes the spot welding path is greater than the standard solder paste weight data, and obtain the solder paste adding timeline at the same time, wherein the solder paste adding timeline is equal to the working time of the target printer;

[0032] If yes, then based on the solder paste adding timeline, a class of solder paste adding time points is generated in the target printer, wherein the class of solder paste adding time points is a time point after the solder paste spot welding head completely executes the spot welding path;

[0033] If not, based on the solder paste adding timeline, the critical point at which the solder paste weight data in the solder paste storage area is not greater than the preset value is the corresponding time point, which is calibrated as the second-class solder paste adding time point;

[0034] If the solder paste storage area is a first-class solder paste storage area, then in the first-class solder paste storage area, the amount of solder paste that needs to be added in the sub-area with less solder paste is calculated based on the real-time distribution map of solder paste and the real-time weight data of solder paste, and the amount is marked as the amount of solder paste to be added in the first-class solder paste storage area;

[0035] The real-time weight data of the solder paste is analyzed. If the real-time weight data of the solder paste is greater than the standard solder paste weight data, when the solder paste adding timeline reaches the time point for adding a type of solder paste, a solder paste equal to the amount of solder paste to be added to the type of solder paste storage area is added to the type of solder paste storage area through the solder paste source;

[0036] If the real-time solder paste weight data is not greater than the standard solder paste weight data, when the solder paste adding timeline reaches the second-class solder paste adding time point, solder paste equal to the amount of solder paste to be added to the first-class solder paste storage area is added to the first-class solder paste storage area through the solder paste source;

[0037] If the solder paste storage area is a Class II solder paste storage area, intelligent solder paste addition processing is performed on the Class II solder paste storage area in combination with the real-time solder paste distribution map and the real-time solder paste weight data.

[0038] Furthermore, in a preferred embodiment of the present invention, if the solder paste storage area is a Class II solder paste storage area, then in combination with the real-time distribution map of the solder paste and the real-time weight data of the solder paste, the Class II solder paste storage area is subjected to intelligent solder paste addition processing, specifically:

[0039] When the solder paste storage area is a Class II solder paste storage area, the amount of solder paste that needs to be added to the sub-area with less solder paste is calibrated as the amount of solder paste to be added to the Class II solder paste storage area;

[0040] In the second-class solder paste storage area, locate the position of the unqualified solder paste, and remove the unqualified solder paste from the position of the unqualified solder paste, so that all the solder paste in the second-class solder paste storage area is qualified solder paste;

[0041] In the solder paste amount sub-area of ​​the second-class solder paste storage area, the real-time solder paste weight data is analyzed. If the real-time solder paste weight data is greater than the standard solder paste weight data, when the solder paste adding timeline reaches the first-class solder paste adding time point, solder paste equal to the amount of solder paste to be added to the second-class solder paste storage area is added to the solder paste amount sub-area of ​​the second-class solder paste storage area through the solder paste source;

[0042] If the real-time solder paste weight data is not greater than the standard solder paste weight data, when the solder paste adding timeline reaches the second-class solder paste adding time point, solder paste equal to the amount of solder paste to be added to the second-class solder paste storage area is added to the solder paste less sub-area of ​​the second-class solder paste storage area through the solder paste source;

[0043] After adding solder paste to the solder paste-less sub-area of ​​the second-class solder paste storage area, performing solder paste paving processing in the second-class solder paste storage area, and obtaining a visual image of the second-class solder paste storage area after adding solder paste;

[0044] The pixel value analysis is performed on the visual image of the Class II solder paste storage area after adding solder paste. When the pixel values ​​of all pixels in the visual image of the Class II solder paste storage area after adding solder paste are the same, the solder paste tiling process is stopped in the Class II solder paste storage area.

[0045] Further, in a preferred embodiment of the present invention, after adding solder paste to the solder paste storage area, a solder paste distribution sequence table of the solder paste spot welding head is generated based on the spot welding path of the solder paste spot welding head, specifically:

[0046] Calculate the solder paste usage rate of all solder paste spot welding heads on the target printer, and preliminarily sort the solder paste distribution order of the solder paste spot welding heads based on the descending principle and the solder paste usage rate, and generate a preliminary solder paste distribution order table for the solder paste spot welding heads;

[0047] Combining the spot welding paths of all solder paste spot welding heads to obtain a total spot welding path, analyzing the total spot welding path to divide the total spot welding path into a non-linear spot welding path and a linear spot welding path;

[0048] Analyze the spot welding paths of all solder paste spot welding heads, calculate the proportion of non-straight paths in the spot welding paths of the solder paste spot welding heads, and based on the descending sorting results of the proportion of non-straight paths in the spot welding paths of the solder paste spot welding heads, update the preliminary solder paste allocation sequence table of the solder paste spot welding heads for a second time to obtain the solder paste allocation sequence table of the solder paste spot welding heads;

[0049] Among them, in the solder paste distribution sequence table of the solder paste spot welding heads, the proportion of non-straight paths in the spot welding paths of the solder paste spot welding heads with an earlier solder paste distribution sequence is greater than the proportion of non-straight paths in the spot welding paths of the solder paste spot welding heads with a later solder paste distribution sequence;

[0050] Based on the solder paste distribution sequence table for solder paste spot welding heads, in the solder paste storage area of ​​the target printer, if the real-time solder paste weight data is not greater than the standard solder paste weight data, the target printer will distribute residual solder paste to different solder paste spot welding heads in the solder paste storage area based on the solder paste distribution sequence table for solder paste spot welding heads until the real-time solder paste weight data is greater than the standard solder paste weight data.

[0051] The second aspect of the present invention further provides a fully automatic visual printer solder paste intelligent adding system based on visual interactive recognition, the solder paste intelligent adding system comprising a memory and a processor, the memory storing a solder paste intelligent adding method, and when the solder paste intelligent adding method is executed by the processor, the following steps are implemented:

[0052] Collect visual images of the solder paste working area in the fully automatic visual printer, perform feature analysis on the collected visual images, determine the distribution of solder paste in the solder paste working area, and classify the solder paste storage area;

[0053] Based on the real-time distribution map of solder paste and the real-time weight data of solder paste, the solder paste adding time node is determined and the solder paste adding process is performed intelligently in the solder paste storage area;

[0054] After solder paste is added to the solder paste storage area, a solder paste distribution sequence table of the solder paste spot welding head is generated based on the spot welding path of the solder paste spot welding head.

[0055] The present invention solves the technical defects existing in the background technology, and has the following beneficial effects: collecting visual images in the solder paste working area, and performing feature extraction and analysis on the solder paste in the image, combining the analysis results to determine the distribution of the solder paste and regional classification of the solder paste storage area, and then combining the real-time weight data of the solder paste, performing intelligent addition processing on the solder paste, and finally generating a solder paste distribution sequence table for the solder paste spot welding head. The present invention can perform visual analysis on the solder paste in the solder paste storage area in the fully automatic visual printer, and generate a solder paste addition plan based on the visual analysis results, so as to prevent the printer from being unable to maintain normal printing due to a small amount of solder paste. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] 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.

[0057] Figure 1 A flow chart of a method for intelligently adding solder paste to a fully automatic visual printer based on visual interactive recognition is shown;

[0058] Figure 2 A flow chart of a method for intelligently adding solder paste is shown;

[0059] Figure 3 The program view of the solder paste intelligent adding system of the fully automatic visual printer based on visual interactive recognition is shown. DETAILED DESCRIPTION

[0060] 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.

[0061] 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.

[0062] Figure 1 A flow chart of a method for intelligently adding solder paste to a fully automatic visual printer based on visual interactive recognition is shown, comprising the following steps:

[0063] S102: collecting visual images of the solder paste working area in the fully automatic visual printer, and performing feature analysis on the collected visual images to determine the distribution of solder paste in the solder paste working area, and classifying the solder paste storage area;

[0064] S104: Based on the real-time distribution map of solder paste and the real-time weight data of solder paste, determining the time node for adding solder paste and performing intelligent solder paste adding processing in the solder paste storage area;

[0065] S106: After solder paste is added to the solder paste storage area, a solder paste distribution sequence table for the solder paste spot welding head is generated based on the spot welding path of the solder paste spot welding head.

[0066] Furthermore, in a preferred embodiment of the present invention, the visual image of the solder paste working area is collected in the fully automatic visual printer, and the characteristics of the collected visual image are analyzed to determine the distribution of the solder paste in the solder paste working area, and the solder paste storage area is classified, specifically:

[0067] The fully automatic visual printer that needs to perform solder paste addition analysis is calibrated as the target printer, and the working area for storing solder paste in the target printer is obtained and calibrated as the solder paste working area;

[0068] In the solder paste storage area, a visual image acquisition device and a visual image analysis device are introduced, and the visual image acquisition device is used to collect the visual image of the current solder paste storage area, and the real-time visual image of the solder paste storage area is calibrated;

[0069] The Sobel operator is introduced to grayscale all pixels in the real-time visual image of the solder paste storage area. After the grayscale processing, the contour features of the solder paste in the real-time visual image of the solder paste storage area are extracted based on the Sobel operator and calibrated as solder paste contour features.

[0070] Obtain a color histogram of solder paste contour features, perform color statistics on the real-time visual image of the solder paste based on the color histogram of the solder paste contour features, and generate solder paste color features;

[0071] A SIFT algorithm is introduced, and a Gaussian convolution kernel is constructed through the SIFT algorithm, and a scale space is constructed, wherein the scale space is obtained by convolving the Gaussian convolution kernel onto the real-time visual image of the solder paste storage area, and is used for feature smoothing of solder paste contour features and solder paste color features;

[0072] In the scale space, scale-invariant feature points are extracted from the solder paste contour feature and the solder paste color feature, and the scale-invariant feature points on the solder paste contour feature and the solder paste color feature are fused to obtain the solder paste contour smoothing feature and the solder paste color smoothing feature.

[0073] The gradient direction histogram of solder paste contour smoothing features and solder paste color smoothing features is performed. Figure 2 Secondary smoothing process to obtain secondary smoothing features of solder paste contour and secondary smoothing features of solder paste color;

[0074] The secondary smooth features of solder paste contour and secondary smooth features of solder paste color are fused to obtain real-time fusion features of solder paste. The real-time fusion features of solder paste are analyzed to determine the weight data and distribution of solder paste in the solder paste storage area, and the solder paste storage areas are classified.

[0075] It should be noted that the solder paste is stored in the solder paste storage area. Usually, the solder paste is solid, and the color of the solder paste is usually bright silver. Visual interactive recognition of solder paste is to perform image analysis on the solder paste, obtain the image of the solder paste in the solder paste working area, and extract the solder paste features for analysis. The solder paste features include contour features and color features. If the contour features are abnormal, that is, the solder paste part is solid, and the color has other discoloration reactions such as yellowing, it proves that the solder paste is abnormal. After extracting the solder paste features, since the extracted features have disadvantages such as noise and non-smoothness, it is necessary to optimize the extracted solder paste features to make the accuracy of the feature analysis of the solder paste higher. The introduction of the Sobel operator can extract the features of the solder paste. The principle is based on the gray value change of the similar end in the image, so as to calculate the image gradient and determine the edge position of the solder paste, that is, the image contour. At the same time, color features can also be obtained. The SIFT algorithm is an algorithm for scale-invariant feature transformation. It uses the convolution of the image and the Gaussian kernel to establish a scale space, and convolve the image in the space to obtain the smooth features of the solder paste. Among them, the gradient direction histogram of the solder paste contour smoothness feature and the solder paste color smoothness feature is performed. Figure 2 The secondary smoothing process can enhance the stability and robustness of the feature. By fusing the two features, the fusion feature of solder paste can be obtained.

[0076] Furthermore, in a preferred embodiment of the present invention, the solder paste real-time fusion feature is used for feature analysis to determine the weight data and distribution of the solder paste in the solder paste storage area, and to classify the solder paste storage area, specifically:

[0077] A weight detection sensor is installed in the solder paste storage area, and the weight detection sensor is operated to calculate the real-time weight data of the solder paste in the solder paste storage area, and calibrate it as the real-time weight data of the solder paste;

[0078] A big data network is introduced, and based on the big data network, standard color features and standard contour features of solder paste are retrieved to generate standard fusion features of solder paste, and a Euclidean distance algorithm is introduced to calculate the Euclidean distance between the standard fusion features of solder paste and the real-time fusion features of solder paste, which are calibrated as a type of Euclidean distance;

[0079] A Euclidean distance threshold is preset, and if a type of Euclidean distance is maintained within the Euclidean distance threshold, the solder paste in the solder paste storage area is marked as solder paste of qualified quality; if a type of Euclidean distance is not maintained within the Euclidean distance threshold, the solder paste in the solder paste storage area is marked as solder paste of unqualified quality;

[0080] If only solder paste of qualified quality exists in the solder paste storage area, the solder paste storage area is marked as a Class I solder paste storage area; if unqualified solder paste exists in the solder paste storage area, the solder paste storage area is marked as a Class II solder paste storage area;

[0081] Calculate the pixel values ​​of different pixels in the solder paste storage area, preset a standard pixel value, mark the pixel points with pixel values ​​less than the standard pixel value as a type of pixel points, and combine all the type of pixel points to obtain a sub-area with less solder paste in the solder paste storage area, and mark it as a sub-area with less solder paste;

[0082] Calculate the pixel value difference between the pixel value of the pixel in the sub-region with less solder paste and the standard pixel value, calibrate it as the target pixel value difference, and calculate the concentration of solder paste that needs to be added to the sub-region with less solder paste based on the target pixel value difference, calibrate it as the target solder paste concentration;

[0083] Calculate the volume of the area with less solder paste, and combine it with the target solder paste concentration to generate the amount of solder paste that needs to be added to the area with less solder paste;

[0084] Based on the solder paste low amount sub-region, a schematic diagram of the real-time distribution of solder paste in the solder paste storage area is constructed and calibrated as a real-time distribution map of solder paste.

[0085] It should be noted that the function of the weight detection sensor is to detect the weight of the solder paste in the solder paste storage area. If the weight of the solder paste is light, it proves that the amount of solder paste in the solder paste storage area is small and needs to be increased. The fusion features of the solder paste are analyzed to determine the difference between the outline and color of the solder paste and the standard value, that is, to determine the Euclidean distance between the features. The smaller the Euclidean distance, the higher the similarity. If the Euclidean distance is not maintained within the specified threshold, it proves that the quality of the solder paste is unqualified, otherwise it is qualified. If the solder paste storage area is full of qualified solder paste, the solder paste storage area is calibrated as a Class I solder paste storage area, otherwise it is a Class II solder paste storage area. In the solder paste storage area, there may be a part of the sub-area where the amount of solder paste is significantly less than that of other sub-areas. The analysis can be performed by analyzing the pixel values ​​of the pixels in the image. If the pixel value of the pixel is large, it proves that the amount of solder paste in the current area is large, otherwise the amount of solder paste is small, which is a sub-area with small amount of solder paste. Based on the pixel value difference between the sub-regions, the concentration difference of the solder paste can be calculated. The concentration difference combined with the volume can determine the amount difference of the solder paste between different sub-regions, thereby obtaining the amount of solder paste that needs to be replenished in the sub-region with less solder paste. Finally, based on the sub-region with less solder paste, a real-time distribution map of the solder paste can be constructed.

[0086] Further, in a preferred embodiment of the present invention, after adding solder paste to the solder paste storage area, a solder paste distribution sequence table of the solder paste spot welding head is generated based on the spot welding path of the solder paste spot welding head, specifically:

[0087] Calculate the solder paste usage rate of all solder paste spot welding heads on the target printer, and preliminarily sort the solder paste distribution order of the solder paste spot welding heads based on the descending principle and the solder paste usage rate, and generate a preliminary solder paste distribution order table for the solder paste spot welding heads;

[0088] Combining the spot welding paths of all solder paste spot welding heads to obtain a total spot welding path, analyzing the total spot welding path to divide the total spot welding path into a non-linear spot welding path and a linear spot welding path;

[0089] Analyze the spot welding paths of all solder paste spot welding heads, calculate the proportion of non-straight paths in the spot welding paths of the solder paste spot welding heads, and based on the descending sorting results of the proportion of non-straight paths in the spot welding paths of the solder paste spot welding heads, update the preliminary solder paste allocation sequence table of the solder paste spot welding heads for a second time to obtain the solder paste allocation sequence table of the solder paste spot welding heads;

[0090] Among them, in the solder paste distribution sequence table of the solder paste spot welding heads, the proportion of non-straight paths in the spot welding paths of the solder paste spot welding heads with an earlier solder paste distribution sequence is greater than the proportion of non-straight paths in the spot welding paths of the solder paste spot welding heads with a later solder paste distribution sequence;

[0091] Based on the solder paste distribution sequence table for solder paste spot welding heads, in the solder paste storage area of ​​the target printer, if the real-time solder paste weight data is not greater than the standard solder paste weight data, the target printer will distribute residual solder paste to different solder paste spot welding heads in the solder paste storage area based on the solder paste distribution sequence table for solder paste spot welding heads until the real-time solder paste weight data is greater than the standard solder paste weight data.

[0092] It should be noted that since there may be multiple solder paste spot welding heads on the printing machine, different spot welding heads are responsible for different spot welding positions, so different solder paste spot welding heads use different solder paste usage rates. When the amount of solder paste is small, it is necessary to apply the limited solder paste to more important spot welding positions to achieve smooth spot welding. Therefore, the solder paste spot welding heads are preliminarily sorted according to the solder paste usage rate, and then sorted again according to the spot welding path of the solder paste spot welding heads. Non-linear spot welding paths are more difficult during spot welding, requiring more solder paste to improve the fault tolerance rate, so they are sorted according to the proportion of non-linear paths in the spot welding paths of the solder paste spot welding heads, and finally the solder paste distribution sequence table of the solder paste spot welding heads is obtained. According to the solder paste distribution sequence table of the solder paste spot welding heads, the participating solder pastes are distributed until the real-time weight data of the solder paste is greater than the standard solder paste weight data, proving that all solder paste spot welding heads can complete the spot welding process, and then the distribution is stopped.

[0093] Figure 2 A flow chart of a method for intelligently adding solder paste is shown, comprising the following steps:

[0094] S202: Based on the real-time distribution map of solder paste and the real-time weight data of solder paste, determining the time node for adding solder paste and performing intelligent solder paste adding processing in the solder paste storage area;

[0095] S204: If the solder paste storage area is a Class II solder paste storage area, intelligent solder paste addition processing is performed on the Class II solder paste storage area in combination with the real-time distribution map of the solder paste and the real-time weight data of the solder paste.

[0096] Further, in a preferred embodiment of the present invention, the solder paste adding time node determination and solder paste intelligent adding processing are performed in the solder paste storage area based on the real-time distribution map of the solder paste and the real-time weight data of the solder paste, specifically:

[0097] Obtain a solder paste source, which is a device that provides solder paste to a target printer, and connect the solder paste source to the target printer so that the target printer can add solder paste to a solder paste storage area based on the solder paste source;

[0098] Determining a position of a solder paste spot welding head in a target printer, wherein the solder paste spot welding head is connected to a solder paste storage area, and the solder paste storage area conducts solder paste to the solder paste spot welding head to implement solder paste spot welding processing;

[0099] Preset standard solder paste weight data, obtain the spot welding path of the solder paste spot welding head, and monitor in real time whether the solder paste weight data in the solder paste storage area when the solder paste spot welding head executes the spot welding path is greater than the standard solder paste weight data, and obtain the solder paste adding timeline at the same time, wherein the solder paste adding timeline is equal to the working time of the target printer;

[0100] If yes, then based on the solder paste adding timeline, a class of solder paste adding time points is generated in the target printer, wherein the class of solder paste adding time points is a time point after the solder paste spot welding head completely executes the spot welding path;

[0101] If not, based on the solder paste adding timeline, the critical point at which the solder paste weight data in the solder paste storage area is not greater than the preset value is the corresponding time point, which is calibrated as the second-class solder paste adding time point;

[0102] If the solder paste storage area is a first-class solder paste storage area, then in the first-class solder paste storage area, the amount of solder paste that needs to be added in the sub-area with less solder paste is calculated based on the real-time distribution map of solder paste and the real-time weight data of solder paste, and the amount is marked as the amount of solder paste to be added in the first-class solder paste storage area;

[0103] The real-time weight data of the solder paste is analyzed. If the real-time weight data of the solder paste is greater than the standard solder paste weight data, when the solder paste adding timeline reaches the time point for adding a type of solder paste, a solder paste equal to the amount of solder paste to be added to the type of solder paste storage area is added to the type of solder paste storage area through the solder paste source;

[0104] If the real-time solder paste weight data is not greater than the standard solder paste weight data, when the solder paste adding timeline reaches the second-class solder paste adding time point, solder paste equal to the amount of solder paste to be added to the first-class solder paste storage area is added to the first-class solder paste storage area through the solder paste source;

[0105] If the solder paste storage area is a Class II solder paste storage area, intelligent solder paste addition processing is performed on the Class II solder paste storage area in combination with the real-time solder paste distribution map and the real-time solder paste weight data.

[0106] It should be noted that the solder paste source is a device that provides solder paste to the target printer, and the process provided is fully automatic. During the spot welding process of the solder paste spot welding head, it is necessary to first determine the spot welding path and carry out spot welding in a planned manner. During the spot welding process, if there is an insufficient amount of solder paste, it is easy to cause a decrease in the spot welding effect and even bring safety hazards, so it is necessary to judge the amount of solder paste during the spot welding process. The amount can be judged based on the weight data. If the weight data is low, solder paste needs to be added, and the time point of addition is analyzed based on the weight data. If the solder paste weight data in the solder paste storage area of ​​the solder paste spot welding head is greater than the standard value when executing the spot welding path, it proves that the solder paste can support the completion of the spot welding, and then the solder paste can be added after the spot welding is completed. Otherwise, solder paste needs to be added during the spot welding process, so the time point of adding a type of solder paste and the time point of adding a type of solder paste are obtained. If the solder paste storage area is a Class I solder paste storage area, it proves that the solder paste is of qualified quality and can be added directly without processing the solder paste. During the adding process, the Class I solder paste addition time point and the Class II solder paste addition time point are selected accordingly according to the weight data for solder paste addition.

[0107] Furthermore, in a preferred embodiment of the present invention, if the solder paste storage area is a Class II solder paste storage area, then in combination with the real-time distribution map of the solder paste and the real-time weight data of the solder paste, the Class II solder paste storage area is subjected to intelligent solder paste addition processing, specifically:

[0108] When the solder paste storage area is a Class II solder paste storage area, the amount of solder paste that needs to be added to the sub-area with less solder paste is calibrated as the amount of solder paste to be added to the Class II solder paste storage area;

[0109] In the second-class solder paste storage area, locate the position of the unqualified solder paste, and remove the unqualified solder paste from the position of the unqualified solder paste, so that all the solder paste in the second-class solder paste storage area is qualified solder paste;

[0110] In the solder paste amount sub-area of ​​the second-class solder paste storage area, the real-time solder paste weight data is analyzed. If the real-time solder paste weight data is greater than the standard solder paste weight data, when the solder paste adding timeline reaches the first-class solder paste adding time point, solder paste equal to the amount of solder paste to be added to the second-class solder paste storage area is added to the solder paste amount sub-area of ​​the second-class solder paste storage area through the solder paste source;

[0111] If the real-time solder paste weight data is not greater than the standard solder paste weight data, when the solder paste adding timeline reaches the second-class solder paste adding time point, solder paste equal to the amount of solder paste to be added to the second-class solder paste storage area is added to the solder paste less sub-area of ​​the second-class solder paste storage area through the solder paste source;

[0112] After adding solder paste to the solder paste-less sub-area of ​​the second-class solder paste storage area, performing solder paste paving processing in the second-class solder paste storage area, and obtaining a visual image of the second-class solder paste storage area after adding solder paste;

[0113] The pixel value analysis is performed on the visual image of the Class II solder paste storage area after adding solder paste. When the pixel values ​​of all pixels in the visual image of the Class II solder paste storage area after adding solder paste are the same, the solder paste tiling process is stopped in the Class II solder paste storage area.

[0114] It should be noted that when the solder paste storage area is a Class II solder paste storage area, it proves that there is unqualified solder paste in the solder paste, and this part of the unqualified solder paste needs to be removed first. Only qualified solder paste is left. Then, according to the real-time distribution map of solder paste, determine the sub-area with less solder paste. Then, according to the weight data of solder paste, select the time point for adding Class I solder paste and the time point for adding Class II solder paste to add solder paste, and during the adding process, the amount of solder paste added is equal to the amount of solder paste to be added in the Class II solder paste storage area, that is, the amount of solder paste that needs to be supplemented. Since the storage area is a Class II solder paste storage area, it proves that this area is prone to uneven distribution of solder paste, so solder paste tiling is performed in the Class II solder paste storage area to ensure uniform distribution of each position and realize the normal operation of each solder paste spot welding head during spot welding. When the pixel values ​​of all pixels in the visual image of the Class II solder paste storage area after adding solder paste are the same, it proves that the tiling process is successful.

[0115] like Figure 3 As shown, the second aspect of the present invention further provides a fully automatic visual printer solder paste intelligent adding system based on visual interactive recognition, the solder paste intelligent adding system includes a memory 31 and a processor 32, the memory 31 stores a solder paste intelligent adding method, and when the solder paste intelligent adding method is executed by the processor 32, the following steps are implemented:

[0116] Collect visual images of the solder paste working area in the fully automatic visual printer, perform feature analysis on the collected visual images, determine the distribution of solder paste in the solder paste working area, and classify the solder paste storage area;

[0117] Based on the real-time distribution map of solder paste and the real-time weight data of solder paste, the solder paste adding time node is determined and the solder paste adding process is performed intelligently in the solder paste storage area;

[0118] After solder paste is added to the solder paste storage area, a solder paste distribution sequence table of the solder paste spot welding head is generated based on the spot welding path of the solder paste spot welding head.

[0119] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A fully automatic visual printer solder paste intelligent adding method based on visual interactive recognition, characterized in that: The following steps are involved: Collect visual images of the solder paste working area in the fully automatic visual printer, perform feature analysis on the collected visual images, determine the distribution of solder paste in the solder paste working area, and classify the solder paste storage area; Based on the real-time distribution map of solder paste and the real-time weight data of solder paste, the solder paste adding time node is determined and the solder paste adding process is performed intelligently in the solder paste storage area; After solder paste is added to the solder paste storage area, a solder paste distribution sequence table of the solder paste spot welding head is generated based on the spot welding path of the solder paste spot welding head; The method includes collecting visual images of the solder paste working area in the fully automatic visual printer, and performing feature analysis on the collected visual images to determine the distribution of solder paste in the solder paste working area, and classifying the solder paste storage area, specifically: The fully automatic visual printer that needs to perform solder paste addition analysis is calibrated as the target printer, and the working area for storing solder paste in the target printer is obtained and calibrated as the solder paste working area; In the solder paste storage area, a visual image acquisition device and a visual image analysis device are introduced, and the visual image acquisition device is used to collect the visual image of the current solder paste storage area, and the real-time visual image of the solder paste storage area is calibrated; The Sobel operator is introduced to grayscale all pixels in the real-time visual image of the solder paste storage area. After the grayscale processing, the contour features of the solder paste in the real-time visual image of the solder paste storage area are extracted based on the Sobel operator and calibrated as solder paste contour features. Obtain a color histogram of solder paste contour features, perform color statistics on the real-time visual image of the solder paste based on the color histogram of the solder paste contour features, and generate solder paste color features; A SIFT algorithm is introduced, and a Gaussian convolution kernel is constructed through the SIFT algorithm, and a scale space is constructed, wherein the scale space is obtained by convolving the Gaussian convolution kernel onto the real-time visual image of the solder paste storage area, and is used for feature smoothing of solder paste contour features and solder paste color features; In the scale space, scale-invariant feature points are extracted from the solder paste contour feature and the solder paste color feature, and the scale-invariant feature points on the solder paste contour feature and the solder paste color feature are fused to obtain the solder paste contour smoothing feature and the solder paste color smoothing feature. Perform secondary smoothing of the gradient direction histogram on the solder paste contour smoothing feature and the solder paste color smoothing feature to obtain the secondary smoothing feature of the solder paste contour and the secondary smoothing feature of the solder paste color; The secondary smooth features of solder paste contour and secondary smooth features of solder paste color are fused to obtain real-time fusion features of solder paste. The real-time fusion features of solder paste are analyzed to determine the weight data and distribution of solder paste in the solder paste storage area, and the solder paste storage areas are classified.

2. According to the method for intelligently adding solder paste to a fully automatic visual printer based on visual interactive recognition as described in claim 1, it is characterized in that: The solder paste real-time fusion feature performs feature analysis to determine the weight data and distribution of the solder paste in the solder paste storage area, and classifies the solder paste storage area, specifically: A weight detection sensor is installed in the solder paste storage area, and the weight detection sensor is operated to calculate the real-time weight data of the solder paste in the solder paste storage area, and calibrate it as the real-time weight data of the solder paste; A big data network is introduced, and based on the big data network, standard color features and standard contour features of solder paste are retrieved to generate standard fusion features of solder paste, and a Euclidean distance algorithm is introduced to calculate the Euclidean distance between the standard fusion features of solder paste and the real-time fusion features of solder paste, which are calibrated as a type of Euclidean distance; A Euclidean distance threshold is preset, and if a type of Euclidean distance is maintained within the Euclidean distance threshold, the solder paste in the solder paste storage area is marked as solder paste of qualified quality; if a type of Euclidean distance is not maintained within the Euclidean distance threshold, the solder paste in the solder paste storage area is marked as solder paste of unqualified quality; If only solder paste of qualified quality exists in the solder paste storage area, the solder paste storage area is marked as a Class I solder paste storage area; if unqualified solder paste exists in the solder paste storage area, the solder paste storage area is marked as a Class II solder paste storage area; Calculate the pixel values ​​of different pixels in the solder paste storage area, preset a standard pixel value, mark the pixel points with pixel values ​​less than the standard pixel value as a type of pixel points, and combine all the type of pixel points to obtain a sub-area with less solder paste in the solder paste storage area, and mark it as a sub-area with less solder paste; Calculate the pixel value difference between the pixel value of the pixel in the sub-region with less solder paste and the standard pixel value, calibrate it as the target pixel value difference, and calculate the concentration of solder paste that needs to be added to the sub-region with less solder paste based on the target pixel value difference, calibrate it as the target solder paste concentration; Calculate the volume of the area with less solder paste, and combine it with the target solder paste concentration to generate the amount of solder paste that needs to be added to the area with less solder paste; Based on the solder paste low amount sub-region, a schematic diagram of the real-time distribution of solder paste in the solder paste storage area is constructed and calibrated as a real-time distribution map of solder paste.

3. According to the method for intelligently adding solder paste to a fully automatic visual printer based on visual interactive recognition as described in claim 1, it is characterized in that: Based on the real-time distribution map of solder paste and the real-time weight data of solder paste, the solder paste adding time node is determined and the solder paste intelligent adding process is performed in the solder paste storage area, specifically: Obtain a solder paste source, which is a device that provides solder paste to a target printer, and connect the solder paste source to the target printer so that the target printer can add solder paste to a solder paste storage area based on the solder paste source; Determining a position of a solder paste spot welding head in a target printer, wherein the solder paste spot welding head is connected to a solder paste storage area, and the solder paste storage area conducts solder paste to the solder paste spot welding head to implement solder paste spot welding processing; Preset standard solder paste weight data, obtain the spot welding path of the solder paste spot welding head, and monitor in real time whether the solder paste weight data in the solder paste storage area when the solder paste spot welding head executes the spot welding path is greater than the standard solder paste weight data, and obtain the solder paste adding timeline at the same time, wherein the solder paste adding timeline is equal to the working time of the target printer; If yes, then based on the solder paste adding timeline, a class of solder paste adding time points is generated in the target printer, wherein the class of solder paste adding time points is a time point after the solder paste spot welding head completely executes the spot welding path; If not, based on the solder paste adding timeline, the critical point at which the solder paste weight data in the solder paste storage area is not greater than the preset value is the corresponding time point, which is calibrated as the second-class solder paste adding time point; If the solder paste storage area is a first-class solder paste storage area, then in the first-class solder paste storage area, the amount of solder paste that needs to be added in the sub-area with less solder paste is calculated based on the real-time distribution map of solder paste and the real-time weight data of solder paste, and the amount is marked as the amount of solder paste to be added in the first-class solder paste storage area; The real-time weight data of the solder paste is analyzed. If the real-time weight data of the solder paste is greater than the standard solder paste weight data, when the solder paste adding timeline reaches the time point for adding a type of solder paste, a solder paste equal to the amount of solder paste to be added to the type of solder paste storage area is added to the type of solder paste storage area through the solder paste source; If the real-time solder paste weight data is not greater than the standard solder paste weight data, when the solder paste adding timeline reaches the second-class solder paste adding time point, solder paste equal to the amount of solder paste to be added to the first-class solder paste storage area is added to the first-class solder paste storage area through the solder paste source; If the solder paste storage area is a Class II solder paste storage area, intelligent solder paste addition processing is performed on the Class II solder paste storage area in combination with the real-time solder paste distribution map and the real-time solder paste weight data.

4. According to the method for intelligently adding solder paste to a fully automatic visual printer based on visual interactive recognition as described in claim 3, it is characterized in that: If the solder paste storage area is a Class II solder paste storage area, then the Class II solder paste storage area is subjected to intelligent solder paste addition processing in combination with the real-time solder paste distribution map and the real-time solder paste weight data, specifically: When the solder paste storage area is a Class II solder paste storage area, the amount of solder paste that needs to be added to the sub-area with less solder paste is calibrated as the amount of solder paste to be added to the Class II solder paste storage area; In the second-class solder paste storage area, locate the position of the unqualified solder paste, and remove the unqualified solder paste at the position of the unqualified solder paste, so that all the solder paste in the second-class solder paste storage area is qualified solder paste; In the solder paste amount sub-area of ​​the second-class solder paste storage area, the real-time solder paste weight data is analyzed. If the real-time solder paste weight data is greater than the standard solder paste weight data, when the solder paste adding timeline reaches the first-class solder paste adding time point, solder paste equal to the amount of solder paste to be added to the second-class solder paste storage area is added to the solder paste amount sub-area of ​​the second-class solder paste storage area through the solder paste source; If the real-time solder paste weight data is not greater than the standard solder paste weight data, when the solder paste adding timeline reaches the second-class solder paste adding time point, solder paste equal to the amount of solder paste to be added to the second-class solder paste storage area is added to the solder paste less sub-area of ​​the second-class solder paste storage area through the solder paste source; After adding solder paste to the solder paste-less sub-area of ​​the second-class solder paste storage area, performing solder paste paving processing in the second-class solder paste storage area, and obtaining a visual image of the second-class solder paste storage area after adding solder paste; The pixel value analysis is performed on the visual image of the Class II solder paste storage area after adding solder paste. When the pixel values ​​of all pixels in the visual image of the Class II solder paste storage area after adding solder paste are the same, the solder paste tiling process is stopped in the Class II solder paste storage area.

5. According to the method for intelligently adding solder paste to a fully automatic visual printer based on visual interactive recognition as described in claim 1, it is characterized in that: After adding solder paste to the solder paste storage area, a solder paste distribution sequence table of the solder paste spot welding head is generated based on the spot welding path of the solder paste spot welding head, specifically: Calculate the solder paste usage rate of all solder paste spot welding heads on the target printer, and preliminarily sort the solder paste distribution order of the solder paste spot welding heads based on the descending principle and the solder paste usage rate, and generate a preliminary solder paste distribution order table for the solder paste spot welding heads; Combining the spot welding paths of all solder paste spot welding heads to obtain a total spot welding path, analyzing the total spot welding path to divide the total spot welding path into a non-linear spot welding path and a linear spot welding path; Analyze the spot welding paths of all solder paste spot welding heads, calculate the proportion of non-straight paths in the spot welding paths of the solder paste spot welding heads, and based on the descending sorting results of the proportion of non-straight paths in the spot welding paths of the solder paste spot welding heads, update the preliminary solder paste allocation sequence table of the solder paste spot welding heads for a second time to obtain the solder paste allocation sequence table of the solder paste spot welding heads; Among them, in the solder paste distribution sequence table of the solder paste spot welding heads, the proportion of non-straight paths in the spot welding paths of the solder paste spot welding heads with an earlier solder paste distribution sequence is greater than the proportion of non-straight paths in the spot welding paths of the solder paste spot welding heads with a later solder paste distribution sequence; Based on the solder paste distribution sequence table for solder paste spot welding heads, in the solder paste storage area of ​​the target printer, if the real-time solder paste weight data is not greater than the standard solder paste weight data, the target printer will distribute residual solder paste to different solder paste spot welding heads in the solder paste storage area based on the solder paste distribution sequence table for solder paste spot welding heads until the real-time solder paste weight data is greater than the standard solder paste weight data.

6. The fully automatic visual printer solder paste intelligent adding system based on visual interactive recognition is characterized by: The solder paste intelligent adding system includes a memory and a processor, wherein the memory stores a solder paste intelligent adding method program. When the solder paste intelligent adding method program is executed by the processor, the steps of the solder paste intelligent adding method of a fully automatic visual printing machine based on visual interactive recognition as described in any one of claims 1-5 are implemented.

Images