A high-precision packing counting method and system for microchannel aluminum flat tubes

Abstract

The present invention discloses a high-precision packing counting method and system for microchannel aluminum flat tubes, which relates to the field of batch production and packaging of microchannel aluminum flat tubes, wherein the method comprises: presetting the inclination angle of the collection area to be no more than 10 degrees, installing a straight laser and making the laser line evenly cover the detection area; automatically matching the best recognition area acquisition image P1 according to the spatial brightness distribution, and collecting the detection area image P2 at the same time; performing image segmentation according to morphology and calculating the number of aluminum flat tubes N1, and calculating whether there is a connected area C and a missed detection area M at the same time; if so, calculating the high probability width L of the aluminum flat tube, performing secondary segmentation on the areas C and M based on the width L, and calculating the respective numbers N2 and N3; verifying whether there is inclination and height difference in the detection area based on the image P2; if so, manually intervening in the detection area to re-measure; and obtaining the number of microchannel aluminum flat tubes N(N1+N2+N3) in the detection area. The present invention can realize accurate packing counting of microchannel aluminum flat tubes.

Description

Technical Field

[0001] The present invention relates to the batch production and packaging of microchannel aluminum flat tubes in the field of industrial manufacturing of air conditioners and refrigeration equipment, aims to solve the problems of large packing counting errors and low efficiency in the production process, and relates to a high-precision packing counting method and system for microchannel aluminum flat tubes based on machine vision recognition technology. Background Art

[0002] Microchannel aluminum flat tubes are a key component used in heat exchange equipment. They are generally 180mm to 400mm long, 15mm to 25mm wide, 1mm to 2mm thick, and weigh 30g to 90g. They have the advantages of light weight, high heat transfer efficiency, and corrosion resistance, and have become an important material in the field of air conditioning and refrigeration equipment. Due to the different requirements of downstream manufacturers of air conditioning and refrigeration equipment for the size, length, and weight of microchannel aluminum flat tubes, microchannel aluminum flat tubes need to be packed according to the requirements of downstream manufacturers when shipped, and the number of each box ranges from 1,500 to 5,000. The traditional packing counting method can no longer meet production needs in terms of efficiency and accuracy, and often leads to a decrease in production efficiency during the production process. Incorrect packing counting leads to frequent rework, wasting time and resources; insufficient or excessive counting causes delivery problems, affecting the company's reputation; inaccurate counting leads to excessive use or waste of materials.

[0003] Machine vision technology has been widely used in industrial production in recent years. It collects images through industrial cameras and combines image processing algorithms to realize automatic detection and identification of target objects. Compared with traditional methods, the advantages of machine vision in counting applications include: non-contact detection, avoiding mechanical wear, and being suitable for high-speed production environments; high precision, using image algorithms to identify the number, position and status of aluminum flat tubes, with significantly improved accuracy; multi-scene adaptability, capable of handling complex scenes, such as overlapping, tilted or partially obscured aluminum flat tubes; intelligence, further improving recognition efficiency and reliability through deep learning algorithms and verification methods. The present invention combines machine vision and structured light verification technology to achieve accurate box counting of microchannel aluminum flat tubes on the production line, with an accuracy of up to 99.9999%. Summary of the invention

[0004] With the rapid development of new energy vehicles and new refrigeration equipment, microchannel aluminum flat tubes are widely used in new energy vehicle air conditioners, refrigeration equipment and other heat exchangers. They are characterized by light weight, excellent thermal conductivity and strong corrosion resistance. Since microchannel aluminum flat tubes usually need to be packed in a certain number and transported to downstream manufacturers for subsequent production, accurate packing count is crucial. The present invention combines machine vision and structured light verification technology to achieve accurate packing count of microchannel aluminum flat tubes.

[0005] In the first aspect, the present application provides a high-precision box counting method for microchannel aluminum flat tubes, which adopts the following technical solution:

[0006] Step S1: Preset the microchannel aluminum flat tube collection area with an inclination angle of no more than 10 degrees, install a 660nm straight laser and make the laser line evenly cover the detection area;

[0007] Step S2: automatically matching the best recognition area to capture the image P1 according to the spatial brightness distribution, and simultaneously capturing the detection area image P2 irradiated by the laser;

[0008] Step S3: Perform Blob analysis on the image and perform corrosion and expansion processing on the identified area;

[0009] Step S4: Perform image segmentation based on morphology and calculate the number of aluminum flat tubes N1, and calculate whether there are connected areas C and missed areas M; if there are connected areas C or missed areas M, proceed to step S5; if there are no connected areas C or missed areas M, proceed to step S6;

[0010] Step S5: Calculate the high probability width L of the microchannel aluminum flat tube, perform secondary segmentation on the connected area C and the missed detection area M based on the aluminum flat tube width L, and calculate the number N2 and N3 of aluminum flat tubes in the connected area and the missed detection area;

[0011] Step S6: Based on the image P2, check whether there is any tilt or height difference of the aluminum tube in the detection area; if so, manually intervene in the detection area and re-measure; if not, proceed to step S7;

[0012] Step S7: Obtain the number N of microchannel aluminum flat tubes in the detection area, that is, N=N1+N2+N3.

[0013] By adopting the above technical solution, the method is composed of multiple modules, each module is responsible for realizing a specific function.

[0014] In step S1, the acquisition workspace module includes a microchannel aluminum flat tube detection workbench and a detection area. The workbench includes a high-resolution industrial camera, a high-brightness bar light source, a 660nm straight laser, a touch-integrated industrial computer, and a wireless micro keyboard. The light source emitted by the high-brightness bar light source and the laser line emitted by the 660nm straight laser evenly cover the detection area, providing a clear reference for subsequent image acquisition; the inclination angle of the detection area is not greater than 10 degrees to ensure that the aluminum flat tubes are neatly placed.

[0015] In step S2, the visual acquisition module automatically matches and selects the best recognition area based on the spatial brightness distribution, and uses a high-resolution industrial camera to collect the detection area image P1. At the same time, the verification area image P2 under the irradiation of a single-word laser is collected for subsequent verification to ensure the integrity and accuracy of the count.

[0016] Specifically, the system captures a preliminary image with a high-resolution camera and analyzes its spatial brightness distribution. This analysis involves evaluating the brightness values ​​of different areas in the image to identify potential aluminum flat tube locations. The system automatically matches and selects the best identification area based on the preset brightness threshold and pattern recognition algorithm. The selection of this area is designed to maximize the visibility of the aluminum flat tube while minimizing background noise and other interference factors.

[0017] After selecting the optimal recognition area, the system controls the camera to acquire a high-resolution image P1 with appropriate exposure and gain settings, ensuring that the image P1 has sufficient details and contrast for accurate image processing and object recognition.

[0018] At the same time, the system also collects the inspection area image P2 under the irradiation of a straight laser. The 660nm straight laser evenly covers the inspection area and provides a bright reference line, which helps to enhance the outline of the aluminum flat tube and highlight its geometric features.

[0019] In step S3, the image preprocessing module performs Blob analysis on the collected image P1 and performs morphological erosion and dilation operations on the identified areas.

[0020] First, the system preprocesses the image, including denoising and contrast enhancement, to ensure that the image quality is suitable for Blob analysis. Then, through Blob analysis, the system identifies connected regions in the image, which correspond to potential locations of aluminum flat tubes. Blob analysis helps the system determine candidate regions for aluminum flat tubes by detecting groups of connected pixels in the image.

[0021] After identifying these blobs, the system performs morphological erosion and dilation operations on each blob. The erosion operation makes the blob smaller by removing pixels on the blob boundary, thereby removing small noise and disconnecting adjacent blobs. This operation helps eliminate fine structures and noise in the image and ensures the accuracy of subsequent counting.

[0022] The system then performs a dilation operation to make the Blob larger by adding pixels near the Blob boundary and filling the holes and small gaps inside the Blob. The dilation operation not only helps to restore some Blob parts lost during the corrosion process, but also connects adjacent Blobs to ensure the integrity and connectivity of the aluminum flat tube.

[0023] Through this series of image processing steps, the system can effectively identify and separate aluminum flat tubes, providing clear and accurate image data for the subsequent counting steps.

[0024] In step S4, the morphological segmentation and counting module, in the subsequent stage of image processing, the system uses morphological methods to further segment the image after Blob analysis and morphological transformation (erosion and expansion). The purpose of this step is to accurately separate adjacent aluminum flat tube areas to ensure that each aluminum flat tube can be individually identified and counted.

[0025] First, the system applies morphological operations, opening and closing operations, to refine the structure in the image. The opening operation removes small objects and disconnects adjacent objects by first corroding and then dilating, which is particularly important for separating closely packed aluminum flat tubes. The closing operation fills small holes inside the object and connects adjacent objects by first dilating and then corroding, ensuring the integrity and connectivity of the aluminum flat tubes.

[0026] After these morphological operations, the system uses the connected domain analysis algorithm to identify and count the independent aluminum flat tubes in the image. Each connected domain represents an independent aluminum flat tube, and the system determines the number N1 of aluminum flat tubes by counting these connected domains.

[0027] At the same time, the system evaluates whether there are connected regions C and missed regions M in the image. Connected regions C refer to the aluminum flat tube regions that remain connected after morphological segmentation, which usually means that some aluminum flat tubes are not correctly separated due to light reflection, irregular arrangement, or adhesion. Missed regions M refer to the regions in the image that are not identified as aluminum flat tubes due to occlusion, vacancies, or image quality issues of the aluminum flat tubes.

[0028] In step S5, after the connected area and missed area processing module identifies and counts the initial number of aluminum flat tubes N1, the system further processes the connected area C and the missed area M to ensure the completeness and accuracy of the count. This step involves calculating the high probability width L of the microchannel aluminum flat tube, and based on this width, the connected area C and the missed area M are divided twice, so as to accurately calculate the number of aluminum flat tubes N2 and N3 in these areas.

[0029] First, the system calculates the most probable width L of the microchannel aluminum flat tube by analyzing the size of the identified aluminum flat tube, which is the average width of aluminum flat tubes with similar widths in the identification area. This width L represents the typical width of the aluminum flat tube in the image and provides a standard for the width of the aluminum flat tube for subsequent secondary segmentation.

[0030] Next, for the area C identified as connected in the initial segmentation process, the system performs secondary segmentation using the calculated width L. The system divides these areas into smaller sub-areas by applying a threshold segmentation algorithm based on the width L, each of which corresponds to a separate aluminum flat tube, thereby accurately counting the number of aluminum flat tubes N2 in these areas.

[0031] Similarly, for the missed detection areas M, which were not correctly identified as aluminum flat tubes during the initial segmentation process, the system also uses a secondary segmentation method based on the width L to process them. By analyzing the image features of these areas and combining the standard of width L, the system can identify aluminum flat tubes that may be missed and count the number of aluminum flat tubes N3 in these areas.

[0032] During the entire secondary segmentation process, the system continuously verifies the rationality of each segmentation result to ensure that each identified aluminum flat tube meets the preset size and shape standards. This not only improves the counting accuracy, but also reduces the counting errors caused by light reflection, irregular arrangement and missed detection of aluminum flat tubes.

[0033] In step S6, after the data verification module completes the initial counting and secondary segmentation, the system uses the laser-assisted image P2 to further verify the detection area to identify whether the aluminum flat tube has tilt or height difference abnormalities. This step is crucial to ensure the accuracy of the counting results, because tilted and height-difference aluminum flat tubes are the biggest factors affecting the counting accuracy.

[0034] First, the system compares and analyzes image P2 with image P1. Image P2 was captured under laser irradiation, which provides a bright reference line that helps highlight the contour and position of the aluminum flat tube. By analyzing the reflection and shadow of the laser line on the aluminum flat tube, the system can detect the tilt angle of the aluminum flat tube and the spacing and height between adjacent aluminum flat tubes.

[0035] Specifically, the system uses an edge detection algorithm to identify feature points at the intersection of the laser line and the aluminum flat tube. These feature points reveal the boundaries and arrangement of the aluminum flat tube. The system further calculates the distance and angle between adjacent feature points to assess the degree of inclination of the aluminum flat tube. If the inclination angle of the aluminum flat tube is detected to exceed the preset threshold, the system will mark the area for further review.

[0036] In addition, the system identifies whether there are height differences between the aluminum flat tubes by analyzing the shadows and reflections in image P2. Aluminum flat tubes with height differences show abnormal reflection patterns under laser irradiation, and these features are significantly different from the normal features of a single aluminum flat tube. By identifying these abnormal features, the system is able to mark the abnormal areas for secondary segmentation or manual review.

[0037] Once the system identifies aluminum flat tubes that are tilted or have height differences, the system will issue a warning prompting the operator to manually intervene in the inspection area and re-measure. The operator manually organizes the aluminum flat tubes in the collection area and re-executes the visual collection module through system control.

[0038] In step S7, the counting output module combines the above steps to output the final counting result N of the aluminum flat tube, that is, N = N1+ N2 + N3, to ensure the high accuracy and comprehensiveness of the counting result.

[0039] In the second aspect, the present application provides a microchannel aluminum flat tube high-precision box counting system, which adopts the following technical solution:

[0040] Inspection workbench: The inspection workbench is connected to the aluminum flat tube production line through a horizontal conveyor belt to collect the aluminum flat tubes to be inspected. The workbench consists of a visual bracket, a high-resolution industrial camera, a high-brightness bar light source, a 660nm straight laser, a touch-integrated industrial computer, a wireless micro keyboard, and an inspection area.

[0041] Acquisition and debugging module: The acquisition and debugging module includes industrial camera loading, camera shutdown, camera parameter setting, acquisition image display, light source parameter setting, and one-word laser parameter setting functions to achieve debugging and status monitoring of detection related equipment.

[0042] Image acquisition and analysis module: including microchannel aluminum flat tube type selection, image preprocessing method setting, image analysis algorithm deployment and version management, connectivity and missed detection image processing.

[0043] Verification warning and log analysis module: This module includes sound and light warning, log analysis, verification settings, exception handling and other functions. When the verification fails or an abnormal situation occurs, the sound and light warning device will activate the alarm and generate a log.

[0044] User interaction module: This module includes preset box value settings, cumulative total, current quantity, cumulative times, difference, system status, and the control of the detection process includes start, count, recount, clear, rollback, and close functions.

[0045] In summary, the present application includes at least one of the following beneficial technical effects:

[0046] 1. A high-precision packing and counting method for microchannel aluminum flat tubes, characterized by comprising an acquisition workspace module, a visual acquisition module, an image preprocessing module, a morphological segmentation and counting module, a connected area and missed area processing module, a data verification module, and a counting output module.

[0047] 2. A high-precision packing and counting system for microchannel aluminum flat tubes, including a detection workbench, an acquisition and debugging module, an image acquisition and analysis module, a verification warning and log analysis module, and a user interaction module.

[0048] 3. The high-precision packing counting method for microchannel aluminum flat tubes provided in this application can effectively improve the counting detection accuracy of microchannel aluminum flat tubes. The detection accuracy of the device and method adopted exceeds 99.9999%, and the performance meets the requirements of large-scale production. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a flow chart of a high-precision packing and counting method for microchannel aluminum flat tubes in an embodiment of the present application.

[0050] Figure 2 It is a structural block diagram of a high-precision packing and counting method for microchannel aluminum flat tubes in an embodiment of the present application.

[0051] Figure 3 These are images captured by the visual acquisition module of the embodiment of the present application, wherein (a) is the detection area image P1 captured under the illumination of a high-brightness strip light source, and (b) is the verification area image P2 captured under the illumination of a 660nm straight laser.

[0052] Figure 4 It is the recognition effect diagram of the embodiment of the present application, wherein (a) is the effect diagram recognized by the morphological segmentation and counting module using the connected domain analysis algorithm, (b) is the secondary segmentation effect diagram of the connected area C by the connected area and missed area processing module, and (c) is the secondary segmentation effect diagram of the missed area M by the connected area and missed area processing module.

[0053] Figure 5 It is a structural block diagram of a high-precision packing counting system for microchannel aluminum flat tubes in an embodiment of the present application. DETAILED DESCRIPTION

[0054] The following is combined with Figure 1-5 This application is described in further detail.

[0055] The present application embodiment discloses a high-precision box counting method for microchannel aluminum flat tubes, such as Figure 1 As shown, the method includes:

[0056] Step S1: Preset the microchannel aluminum flat tube collection area with an inclination angle of no more than 10 degrees, install a 660nm straight laser and make the laser line evenly cover the detection area;

[0057] Step S2: automatically matching the best recognition area to capture the image P1 according to the spatial brightness distribution, and simultaneously capturing the detection area image P2 irradiated by the laser;

[0058] Step S3: Perform Blob analysis on the image and perform corrosion and expansion processing on the identified area;

[0059] Step S4: Perform image segmentation based on morphology and calculate the number of aluminum flat tubes N1, and calculate whether there are connected areas C and missed areas M; if there are connected areas C or missed areas M, proceed to step S5; if there are no connected areas C or missed areas M, proceed to step S6;

[0060] Step S5: Calculate the high probability width L of the microchannel aluminum flat tube, perform secondary segmentation on the connected area C and the missed detection area M based on the aluminum flat tube width L, and calculate the number N2 and N3 of aluminum flat tubes in the connected area and the missed detection area;

[0061] Step S6: Based on the image P2, check whether there is any tilt or height difference of the aluminum tube in the detection area; if so, manually intervene in the detection area and re-measure; if not, proceed to step S7;

[0062] Step S7: Obtain the number N of microchannel aluminum flat tubes in the detection area, that is, N=N1+N2+N3.

[0063] like Figure 2 As shown, the method includes the following modules:

[0064] Collection workspace module: includes a microchannel aluminum flat tube detection workbench and a detection area. The workbench includes a high-resolution industrial camera, a high-brightness strip light source, a 660nm straight laser, a touch-integrated industrial computer, and a wireless micro keyboard. The light source emitted by the high-brightness strip light source and the laser line emitted by the 660nm straight laser evenly cover the detection area, providing a clear reference for subsequent image acquisition; the inclination angle of the detection area is not greater than 10 degrees to ensure that the aluminum flat tubes are neatly arranged; the high-resolution industrial camera adopts Hikvision's 20-megapixel industrial array camera, and is connected to the touch-integrated industrial computer via Gigabit Ethernet. The industrial camera, strip light source, and straight laser are set above the detection area using a visual bracket, so that the industrial camera's field of view completely covers the detection area.

[0065] Visual acquisition module: According to the spatial brightness distribution, it automatically matches and selects the best recognition area, and uses a high-resolution industrial camera to collect the detection area image P1. At the same time, it collects the verification area image P2 under the irradiation of a single-word laser for subsequent verification to ensure the integrity and accuracy of the counting. Figure 3 (a) is the detection area image P1, and (b) is the verification area image P2.

[0066] Specifically, the system captures a preliminary image through a high-resolution camera and analyzes its spatial brightness distribution. This analysis involves evaluating the brightness values ​​of different areas in the image to identify potential locations of aluminum flat tubes. The system automatically matches and selects the best identification area based on the preset brightness threshold and pattern recognition algorithm. The selection of this area is intended to maximize the visibility of the aluminum flat tube while minimizing background noise and other interference factors. In this embodiment, based on the analysis of the imaging effect of the industrial camera, the system presets a brightness threshold of 240, and binarizes the captured image to divide the pixels in the image into two categories: pixels with brightness higher than the threshold are aluminum flat tubes, and pixels with brightness lower than the threshold are backgrounds, thereby segmenting each aluminum flat tube. At the same time, the system matches the extracted pixel features with the pre-collected aluminum flat tube shape feature patterns, and identifies those that match consistently as aluminum flat tubes.

[0067] After selecting the optimal recognition area, the system controls the camera to acquire a high-resolution image P1 with appropriate exposure and gain settings, ensuring that the image P1 has sufficient details and contrast for accurate image processing and object recognition.

[0068] At the same time, the system also collects the inspection area image P2 under the irradiation of a straight laser. The 660nm straight laser evenly covers the inspection area and provides a bright reference line, which helps to enhance the outline of the aluminum flat tube and highlight its geometric features.

[0069] Image preprocessing module: Perform Blob analysis on the collected image P1 and perform morphological erosion and dilation operations on the identified areas.

[0070] First, the system preprocesses the image, including denoising and contrast enhancement, to ensure that the image quality is suitable for Blob analysis. Then, through Blob analysis, the system identifies connected regions in the image, which correspond to potential locations of aluminum flat tubes. Blob analysis helps the system determine candidate regions for aluminum flat tubes by detecting groups of connected pixels in the image.

[0071] After identifying these blobs, the system performs morphological erosion and dilation operations on each blob. The erosion operation makes the blob smaller by removing pixels on the blob boundary, thereby removing small noise and disconnecting adjacent blobs. This operation helps eliminate fine structures and noise in the image and ensures the accuracy of subsequent counting.

[0072] The system then performs a dilation operation to make the Blob larger by adding pixels near the Blob boundary and filling the holes and small gaps inside the Blob. The dilation operation not only helps to restore some Blob parts lost during the corrosion process, but also connects adjacent Blobs to ensure the integrity and connectivity of the aluminum flat tube.

[0073] Through this series of image processing steps, the system can effectively identify and separate aluminum flat tubes, providing clear and accurate image data for the subsequent counting steps.

[0074] Morphological segmentation and counting module: In the subsequent stage of image processing, the system uses morphological methods to further segment the image after Blob analysis and morphological transformation (erosion and expansion). The purpose of this step is to accurately separate adjacent aluminum flat tube areas to ensure that each aluminum flat tube can be individually identified and counted.

[0075] First, the system applies morphological operations, opening and closing operations, to refine the structure in the image. The opening operation removes small objects and disconnects adjacent objects by first corroding and then dilating, which is particularly important for separating closely packed aluminum flat tubes. The closing operation fills small holes inside the object and connects adjacent objects by first dilating and then corroding, ensuring the integrity and connectivity of the aluminum flat tubes.

[0076] After these morphological operations, the system uses the connected domain analysis algorithm to identify and count the independent aluminum flat tubes in the image. Each connected domain represents an independent aluminum flat tube, and the system determines the number of aluminum flat tubes N1 by counting these connected domains, such as Figure 4 As shown in (a), each red-marked area is an independent aluminum flat tube identified by the system, and the identified aluminum flat tubes are numbered for easy identification and statistics.

[0077] At the same time, the system evaluates whether there are connected areas C and missed areas M in the image. Connected area C refers to the area of ​​aluminum flat tubes that remains connected after morphological segmentation, which usually means that some aluminum flat tubes have not been correctly separated due to light reflection, irregular arrangement, and adhesion. Missed area M refers to the area in the image that is not identified as an aluminum flat tube, which is caused by occlusion, vacancy or image quality problems of the aluminum flat tube. In this embodiment, after morphological operations, the system calculates the characteristic parameters of each connected domain. If the area or shape characteristics of a connected domain obviously exceed the normal range of a single aluminum flat tube, the system marks it as a connected area C. For example, if the area of ​​a connected domain is greater than twice the average area of ​​a single aluminum flat tube, or its shape characteristics do not conform to the typical characteristics of a single aluminum flat tube, then this area may contain multiple adhered aluminum flat tubes. At the same time, the center distance between two adjacent connected domains is calculated. The distance is calculated based on the center coordinates of the identified independent aluminum flat tubes, and the center distances between all adjacent connected domains are counted, and their average value is calculated and recorded as the average distance. When the center distance between two adjacent connected domains obviously exceeds the average distance, the system marks the middle area between the two connected domains as the missed detection area M. For example, if the average distance between the centers of the identified aluminum flat tubes is D, and the center distance between two connected domains is greater than 2D, then the area may contain missed aluminum flat tubes.

[0078] Connected area and missed area processing module: After identifying and counting the initial number of aluminum flat tubes N1, the system further processes the connected area C and missed area M to ensure the completeness and accuracy of the count. This step involves calculating the high probability width L of the microchannel aluminum flat tubes and performing secondary segmentation of the connected area C and missed area M based on this width, thereby accurately calculating the number of aluminum flat tubes N2 and N3 in these areas. Figure 4 In (b), the red rectangular area numbered 12 in the middle is the secondary segmentation effect of the connected area C, and in (c), the red rectangular area numbered 14 in the middle is the secondary segmentation effect of the missed area M.

[0079] First, the system calculates the most probable width L of the microchannel aluminum flat tube by analyzing the size of the identified aluminum flat tube, which is the average width of aluminum flat tubes with similar widths in the identification area. This width L represents the typical width of the aluminum flat tube in the image and provides a standard for the width of the aluminum flat tube for subsequent secondary segmentation.

[0080] Next, for the area C identified as connected in the initial segmentation process, the system performs secondary segmentation using the calculated width L. The system divides these areas into smaller sub-areas by applying a threshold segmentation algorithm based on the width L, each of which corresponds to a separate aluminum flat tube, thereby accurately counting the number of aluminum flat tubes N2 in these areas.

[0081] Similarly, for the missed detection areas M, which were not correctly identified as aluminum flat tubes during the initial segmentation process, the system also uses a secondary segmentation method based on the width L to process them. By analyzing the image features of these areas and combining the standard of width L, the system can identify aluminum flat tubes that may be missed and count the number of aluminum flat tubes N3 in these areas.

[0082] During the entire secondary segmentation process, in addition to calculating the average width, the system also calculated the width standard deviation, the average interval, the interval standard deviation, and the confidence interval. The system continuously verifies the rationality of each segmentation result through these analysis data to ensure that each identified aluminum flat tube meets the preset size and shape standards. This not only improves the accuracy of counting, but also reduces the counting errors caused by light reflection, irregular arrangement, and missed detection of aluminum flat tubes. For example, when detecting 252 1.8mm thick aluminum flat tubes, the system analyzed the following data: average width: 10.35, width standard deviation: 0.59, average interval: 8.18, interval standard deviation: 1.01, confidence interval: [7.13, 11.69].

[0083] In the data verification module, after completing the initial counting and secondary segmentation, the system uses the laser-assisted image P2 to further verify the detection area to identify whether the aluminum flat tube has tilt or height difference abnormalities. This step is crucial to ensure the accuracy of the counting results, because tilted and height-difference aluminum flat tubes are the biggest factors affecting the counting accuracy.

[0084] First, the system compares and analyzes image P2 with image P1. Image P2 was captured under laser irradiation, which provides a bright reference line that helps highlight the contour and position of the aluminum flat tube. By analyzing the reflection and shadow of the laser line on the aluminum flat tube, the system can detect the tilt angle of the aluminum flat tube and the spacing and height between adjacent aluminum flat tubes.

[0085] Specifically, the system uses an edge detection algorithm to identify feature points at the intersection of the laser line and the aluminum flat tube. These feature points reveal the boundaries and arrangement of the aluminum flat tube. The system further calculates the distance and angle between adjacent feature points to assess the degree of inclination of the aluminum flat tube. If the inclination angle of the aluminum flat tube is detected to exceed the preset threshold, the system will mark the area for further review.

[0086] In addition, the system identifies whether there are height differences between the aluminum flat tubes by analyzing the shadows and reflections in image P2. Aluminum flat tubes with height differences show abnormal reflection patterns under laser irradiation, and these features are significantly different from the normal features of a single aluminum flat tube. By identifying these abnormal features, the system is able to mark the abnormal areas for secondary segmentation or manual review.

[0087] Once the system identifies aluminum flat tubes that are tilted or have height differences, the system will issue a warning prompting the operator to manually intervene in the inspection area and re-measure. The operator manually organizes the aluminum flat tubes in the collection area and re-executes the visual collection module through system control.

[0088] The counting output module combines the above steps to output the final counting result N of the aluminum flat tube, that is, N = N1 + N2 + N3, ensuring the high accuracy and comprehensiveness of the counting result.

[0089] Based on the above method, the embodiment of the present application also discloses a high-precision packing counting system for microchannel aluminum flat tubes.

[0090] Combination Figure 5 The system includes the following modules:

[0091] Inspection workbench: The inspection workbench is connected to the aluminum flat tube production line through a horizontal conveyor belt to collect the aluminum flat tubes to be inspected. The workbench consists of a visual bracket, a high-resolution industrial camera, a high-brightness bar light source, a 660nm straight laser, a touch-integrated industrial computer, a wireless micro keyboard, and an inspection area.

[0092] Acquisition debugging module: The acquisition debugging module includes industrial camera loading, camera shutdown, camera parameter setting, acquired image display, light source parameter setting, and one-word laser parameter setting functions to achieve debugging and status monitoring of detection-related equipment. This embodiment is developed using C# language based on Hikvision camera SDK.

[0093] Image acquisition and analysis module: including microchannel aluminum flat tube type selection, image preprocessing method setting, image analysis algorithm deployment and version management, connectivity and missed detection image processing. This embodiment classifies the microchannel aluminum flat tubes produced by the actual production line according to their thickness and edge shape, and develops corresponding analysis algorithms and connectivity and missed detection image processing algorithms for each category.

[0094] Verification warning and log analysis module: This module includes sound and light warning, log analysis, verification settings, exception handling and other functions. When the verification fails or an abnormal situation occurs, the sound and light warning device will activate the alarm and generate a log.

[0095] User interaction module: This module includes preset box value settings, cumulative totals, current quantities, cumulative times, differences, and system status. The control of the detection process includes start, counting, recounting, clearing, rollback, and shutdown functions. In this embodiment, the user interaction module adopts a simplified interface design, and the system can be controlled by touching an integrated industrial computer or by a wireless micro keyboard.

[0096] The embodiments of this specific implementation method are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Therefore, all equivalent changes made based on the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A high-precision packing and counting method for microchannel aluminum flat tubes, characterized in that: include: Step S1: Preset the microchannel aluminum flat tube collection area with an inclination angle of no more than 10 degrees, install a 660nm straight laser and make the laser line evenly cover the detection area; Step S2: automatically matching the best recognition area to capture the image P1 according to the spatial brightness distribution, and simultaneously capturing the detection area image P2 irradiated by the laser; Step S3: Perform Blob analysis on the image and perform corrosion and expansion processing on the identified area; Step S4: Perform image segmentation based on morphology and calculate the number of aluminum flat tubes N1, and calculate whether there are connected areas C and missed areas M. If there are connected areas C or missed areas M, proceed to step S5; if there are no connected areas C or missed areas M, proceed to step S6; Step S5: Calculate the high probability width L of the microchannel aluminum flat tube, perform secondary segmentation on the connected area C and the missed detection area M based on the width L of the aluminum flat tube, and calculate the number N2 and N3 of aluminum flat tubes in the connected area and the missed detection area; Step S6: Based on the image P2, check whether there is any tilt or height difference of the aluminum tube in the detection area; if so, manually intervene in the detection area and re-measure; if not, proceed to step S7; Step S7: Obtain the number N of microchannel aluminum flat tubes in the detection area, that is, N=N1+N2+N3.

2. A high-precision packing and counting method for microchannel aluminum flat tubes according to claim 1, characterized in that: include: Step S2, the system automatically matches and selects the best recognition area based on the spatial brightness distribution, uses a high-resolution industrial camera to collect the detection area image P1, and at the same time, collects the verification area image P2 under the irradiation of a single-word laser; Specifically, the system captures preliminary images through a high-resolution camera and analyzes their spatial brightness distribution. This analysis involves evaluating the brightness values ​​of different areas in the image to identify potential aluminum flat tube locations. The system automatically matches and selects the best identification area based on the preset brightness threshold and pattern recognition algorithm. After selecting the best recognition area, the system controls the camera to capture the high-resolution image P1 with appropriate exposure and gain settings. At the same time, the system also captures the verification area image P2 under laser irradiation.

3. A high-precision packing and counting method for microchannel aluminum flat tubes according to claim 1, characterized in that: include: Step S4, in the subsequent stage of image processing, the system uses a morphological method to further segment the image after Blob analysis and morphological transformation processing, and uses a connected domain analysis algorithm to identify and count independent aluminum flat tubes in the image. Each connected domain represents an independent aluminum flat tube. The system determines the number N1 of aluminum flat tubes by counting these connected domains. At the same time, the system evaluates whether there are connected regions C and missed regions M in the image.

4. A high-precision packing and counting method for microchannel aluminum flat tubes according to claim 1, characterized in that: include: Step S5, first, the system calculates the most probable width L of the microchannel aluminum flat tube by analyzing the size of the identified aluminum flat tube. The width is the average width of aluminum flat tubes with similar widths in the identified area. This width L represents the typical width of the aluminum flat tube in the image and provides a standard for the width of the aluminum flat tube for subsequent secondary segmentation. Next, for the area C identified as connected in the initial segmentation process, the system uses the calculated width L for secondary segmentation. The system divides these areas into smaller sub-areas by applying a threshold segmentation algorithm based on the width L. Each sub-area corresponds to a separate aluminum flat tube, thereby accurately counting the number of aluminum flat tubes N2 in these areas; Similarly, for the missed areas M, these areas were not correctly identified as aluminum flat tubes during the initial segmentation process. The system also uses a secondary segmentation method based on the width L for processing. By analyzing the image features of these areas and combining the standard of the width L, the system can identify the aluminum flat tubes that may be missed and count the number N3 of aluminum flat tubes in these areas.

5. The high-precision packing and counting method for microchannel aluminum flat tubes according to claim 1 is characterized in that: include: Step S6, after completing the initial counting and secondary segmentation, the system uses the laser-assisted image P2 to further verify the detection area to identify whether the aluminum flat tube has any abnormal tilt or height difference; First, the system compares and analyzes image P2 with image P1. By analyzing the reflection and shadow of the laser line on the aluminum flat tube, the system can detect the inclination angle of the aluminum flat tube and the spacing and height between adjacent aluminum flat tubes. Specifically, the system uses an edge detection algorithm to identify feature points at the junction of the laser line and the aluminum flat tube. These feature points reveal the boundaries and arrangement of the aluminum flat tube. The system further calculates the distance and angle between adjacent feature points to assess the degree of inclination of the aluminum flat tube. If the inclination angle of the aluminum flat tube is detected to exceed the preset threshold, the system will mark the area for further review. In addition, the system identifies whether there are height differences between the aluminum flat tubes by analyzing the shadows and reflections in the image P2. The aluminum flat tubes with height differences show abnormal reflection patterns under laser irradiation. These features are significantly different from the normal features of a single aluminum flat tube. By identifying these abnormal features, the system can mark the abnormal area for secondary segmentation or manual review. Once the system identifies aluminum flat tubes that are tilted or have height differences, the system will issue a warning prompting the operator to manually intervene in the inspection area and re-measure. The operator manually organizes the aluminum flat tubes in the collection area and re-executes the visual collection module through system control.

6. A high-precision packing and counting system for microchannel aluminum flat tubes, applied to a high-precision packing and counting method for microchannel aluminum flat tubes as claimed in claim 1, characterized in that: include: Inspection workbench: The inspection workbench is connected to the aluminum flat tube production line through a transverse conveyor belt to collect the aluminum flat tubes to be inspected; The workbench consists of a visual support, a high-resolution industrial camera, a high-brightness bar light source, a 660nm straight laser, a touch-integrated industrial computer, a wireless micro keyboard, and a detection area; Acquisition debugging module: The acquisition debugging module includes industrial camera loading, camera shutdown, camera parameter setting, acquisition image display, light source parameter setting, and one-word laser parameter setting functions to achieve debugging and status monitoring of detection-related equipment; Image acquisition and analysis module: including microchannel aluminum flat tube type selection, image preprocessing method setting, image analysis algorithm deployment and version management, connectivity and missed detection image processing; Verification warning and log analysis module: This module includes sound and light warning, log analysis, verification settings, and exception handling functions. When the verification fails or an abnormal situation occurs, the sound and light warning device will start the alarm and generate a log; User interaction module: This module includes preset box value settings, cumulative total, current quantity, cumulative times, difference, system status, and the control of the detection process includes start, count, recount, clear, rollback, and close functions.

Images