A flatness testing system and method based on high-toughness PVC guardrail

Abstract

This invention relates to the field of guardrail flatness detection technology, and particularly to a flatness detection system and method based on high-toughness PVC guardrails. The system includes a lower-level device and a higher-level analysis system. The lower-level device includes a frame with three electric push rods mounted on it. The output ends of the left and right electric push rods are respectively equipped with coating equipment and ink wiping strips. The coating equipment includes an ink sac and an ink rod. The output end of the middle electric push rod is equipped with two 45° inclined scrapers via a telescopic rod. An image acquisition device is mounted on the frame near the ink wiping strips. In this solution, through the physical marking design of the lower-level device, the intelligent analysis and automated classification design of the higher-level system, and the real-time linkage between the two systems, accurate marking, intelligent analysis, and automated classification of surface defects on high-toughness PVC guardrails are achieved, improving detection efficiency and accuracy, and forming a closed-loop automation process from guardrail loading to classification and storage.

Description

Technical Field

[0001] This invention relates to the field of guardrail testing technology, and in particular to a flatness testing system and method based on high-toughness PVC guardrails. Background Technology

[0002] High-toughness PVC guardrails are made of polyvinyl chloride (PVC) as the main raw material, with added toughening agents and other components. They are guardrail products with high flexibility and impact resistance. They are often used for isolation and protection in places such as roadsides, parks, and communities. They have good protective functions and can adapt to certain deformations without breaking easily.

[0003] The flatness test is to ensure that the surface of the guardrail is smooth and free of obvious bumps and defects. This not only affects the appearance quality of the guardrail, but also its service life (for example, protrusions are prone to wear and tear, and depressions are prone to the accumulation of debris leading to corrosion) and the overall stability after installation.

[0004] Traditional inspection systems have many shortcomings: on the one hand, they lack inspection accuracy and are difficult to accurately identify minute defects, especially for products like high-toughness PVC railings that may have slight deformation, which can easily lead to missed detections; on the other hand, they are inefficient, relying heavily on manual visual inspection or semi-automated equipment, which is not only slow but also greatly affected by human factors, with inconsistent judgment standards, and cannot meet the needs of large-scale production.

[0005] Therefore, this invention proposes a flatness detection system and method based on high-toughness PVC guardrails. Through the physical marking design of the lower-level machine equipment, the intelligent analysis and automated classification design of the upper-level machine system, and the real-time linkage between the upper and lower-level machines, the system achieves accurate marking, intelligent analysis, and automated classification of surface concave and convex defects of high-toughness PVC guardrails. This forms a closed-loop automation of the entire process from guardrail loading to classification and storage, improving detection efficiency and accuracy. Summary of the Invention

[0006] The technical problems to be solved were: insufficient precision, difficulty in distinguishing subtle bumps and depressions, and low efficiency.

[0007] To address the shortcomings of existing technologies, this invention provides a flatness detection system and method based on high-toughness PVC guardrails, thereby solving the technical problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A flatness detection system based on high-toughness PVC guardrails includes a lower-level device and a higher-level analysis system:

[0010] The lower-level equipment includes a frame on which three electric push rods are mounted. The output ends of the left and right electric push rods are respectively equipped with coating equipment and ink wiping strips. The coating equipment includes an ink sac and an ink stick. The output end of the middle electric push rod is equipped with two scrapers at a 45° angle via a telescopic rod. An image acquisition device is mounted on the side of the frame near the ink wiping strips. Three conveyor belts are provided on the lower side of the frame, corresponding to the coating, ink wiping, and sorting areas, respectively.

[0011] The ink stick, scraper, and ink wiping strip of the coating equipment are all designed to fit the shape of the guardrail surface on their contact surfaces.

[0012] The host computer analysis system includes an image acquisition unit, an image analysis unit, and a sorting control unit. The image analysis unit includes three levels of processing modules: an image preprocessing module, a defect identification and classification module, and a defect level determination module.

[0013] In one possible implementation, the telescopic rod consists of a sleeve rod, a sliding rod, and a spring. The sliding rod is slidably nested inside the sleeve rod, and the two ends of the spring are respectively connected to the top of the inner wall of the sleeve rod and the top of the sliding rod. The scraper can adaptively conform to the surface of the guardrail through elastic deformation.

[0014] In one possible implementation, the ink sac of the coating device contains water-based environmentally friendly ink, which uses water as a solvent and contains acrylic resin. The ink stick is made of a porous elastic material that can quantitatively draw ink and form a uniform coating under 8N pressure.

[0015] In one possible implementation, the three conveyor belts are segmented structures with independent drives. The distance between the left and middle conveyor belts is ≥5cm. The running speed of the left and middle conveyor belts is synchronously 0.6m / s. The right conveyor belt can be steplessly adjusted in the range of 0.4-0.8m / s by PLC control.

[0016] In one possible implementation, the image acquisition unit of the host computer analysis system establishes real-time communication with the lower-level device through the ModbusTCP industrial protocol, and can synchronously control the conveyor belt speed, electric push rod pressure and the triggering timing of the image acquisition device, and receive the device status data fed back by the lower-level device.

[0017] In one possible implementation, the defect identification and classification module, based on the preprocessed image, extracts features of the ink residue area using three basic algorithms: connected component analysis, feature matching, and geometric parameter calculation, thereby identifying the defect type and determining its location. The specific process is as follows:

[0018] Ink residue area extraction: Based on the color-normalized binarized image (ink is black, guardrail is white), all black areas (i.e. ink residue areas) are identified by connected component analysis algorithm, and the area, perimeter, and bounding rectangle parameters of each area are calculated.

[0019] Defect type determination:

[0020] Raised defects: Based on the characteristic of "ink residue on both sides of the raised area", the ink residue area corresponding to the raised area is in the form of "symmetrical double stripes" (located on both sides of the raised area), and the distance between the two stripes matches the width of the scraper. The determination is made by calculating the symmetry parameters of the residue area (such as the deviation of the symmetry axis < 5%) and the strip length.

[0021] Indentation Defect: Based on the characteristic that "the ink in the indentation has not been scraped off", the ink residue area corresponding to the indentation is in the form of a "single-area closed shape" (consistent with the shape of the indentation). It is determined by detecting whether the residue area is an independent closed area (without symmetrical strip features).

[0022] Defect location: Establish a coordinate system with the upper left corner of the guardrail as the origin. Calculate the specific location of the defect on the guardrail surface using the coordinates of the center point of the circumscribed rectangle of the residual area, and mark it as "protruding area (x1, y1, x2, y2)" or "recessed area (x3, y3, x4, y4)".

[0023] In one possible implementation, the defect level determination module determines the level of the identified defects based on a preset level classification standard and outputs the relevant results, as follows:

[0024] Grading criteria: Three levels are set based on the geometric parameters of the residual area:

[0025] Level 1 Non-destructive: No ink residue, or residue area < 0.5cm² 2 The issue may be due to minor impurities, and the test result is deemed acceptable.

[0026] Level 2 Repairable: Total length of raised residual strips < 5cm and width of each strip < 0.3cm; or residual area of ​​depressions 0.5-2cm. 2 The defect was determined to be minor and repairable.

[0027] Level 3, irreparable: Total length of raised residual strips ≥ 5cm or width of a single strip ≥ 0.3cm; or area of ​​dented residual area ≥ 2cm². 2 It was determined to be a serious defect and irreparable;

[0028] Automatic matching of grade and output of results: The extracted residual area parameters are compared with the grade standard to automatically determine the grade of the guardrail, generate a defect report containing type, location and grade, output in JSON format, and link with the sorting unit to trigger the corresponding sorting action.

[0029] In one possible implementation, the sorting control unit of the host computer analysis system is linked to the PLC controller of the right conveyor belt via a Profinet industrial bus, and can trigger the corresponding actuator according to the 1, 2, or 3 defect levels output by the image analysis unit:

[0030] The Level 1 signal triggers the first group of pneumatic push rods, with a response time of ≤0.5 seconds, which pushes the guide plate to guide the guardrail into the qualified branch area;

[0031] The Level 2 signal triggers the second group of electric swing arms, which swing at a 30° angle to guide the guardrail into the repair area branch.

[0032] The level 3 signal triggers the third waste pushing mechanism, which moves the guardrail to the waste area;

[0033] Meanwhile, this unit can record the quantity and inspection time data of guardrails at each level, and supports data traceability by connecting with the production management system.

[0034] In one possible implementation, a flatness detection method based on high-toughness PVC guardrails is applied to the system described above, the method comprising the following steps:

[0035] S1. Ink Coating and Scraping: During ink coating, the left-side electric push rod drives the coating equipment to move downwards, and the ink bar adheres to the guardrail with 8N pressure, absorbing water-based environmentally friendly ink. The ink is evenly coated on the left conveyor belt running at 0.6m / s. During the scraping stage, the middle electric push rod, through the spring-loaded telescopic rod, causes two 45° inclined scraping strips to adhere to the guardrail with 10-15N pressure. They move with the synchronously running middle conveyor belt, scraping away ink from flat areas, leaving symmetrical double-strip ink marks on raised areas, and leaving closed ink marks on recessed areas.

[0036] S2, Image Acquisition: The guardrail after ink scraping and wiping is transported to the image acquisition area. The 10-megapixel industrial camera triggers shooting 0.3 seconds after ink scraping is completed. It uses a ring light and macro lens to acquire images of the guardrail surface. The light sensor adjusts the exposure time and light source brightness in real time from 5-20ms to ensure image clarity.

[0037] S3. Analysis: After receiving the image, the host computer system first performs noise reduction by Gaussian filtering, corrects the tilt angle by Hough transform and converts it to the HSV color space. Then, it uses connected component analysis to extract parameters of the ink residue area, identifies raised or sunken defects by feature matching, and finally determines the defect level based on the preset threshold and generates an analysis report containing the defect location.

[0038] S4. Ink wiping: The electric push rod on the right side drives the ink wiping strip to adhere to the surface of the guardrail with a pressure of 5N. It moves with the middle conveyor belt to further remove residual ink.

[0039] S5. Sorting: According to the defect level instructions output by the host computer, the right conveyor belt diverts the guardrails to the qualified area, repair area or waste area through the corresponding actuators, with an action response time of ≤0.5 seconds.

[0040] In one possible implementation, the image preprocessing in step S3 also includes grayscale normalization processing, which converts the ink area to black and the guardrail area to white; the defect location marking accuracy is ±0.1cm, and the coordinates are represented in the format of "(x1, y1, x2, y2)"; the JSON report contains guardrail ID, defect type, location and grade information, and is sent to the sorting unit in real time via the industrial bus, triggering the actuator to complete the action within ≤0.5 seconds, and the action force can be adjusted to avoid damaging the guardrail.

[0041] Beneficial effects compared to existing technologies:

[0042] 1. In this solution, the combination of a 45° inclined scraper and a spring-loaded telescopic rod in the lower-level equipment, along with a segmented conveyor belt and water-based environmentally friendly ink, enables precise physical marking of uneven defects on the guardrail surface. The scraper can adaptively conform to the guardrail, forming symmetrical double-strip ink marks on raised areas and closed ink marks on recessed areas. The segmented conveyor belt avoids ink contamination, and the water-based ink is easy to scrape off and has high contrast, ensuring that the defect morphology is clearly identifiable and providing a reliable physical basis for subsequent inspection.

[0043] 2. In this solution, through the design of image preprocessing, multi-algorithm collaborative recognition, and hierarchical sorting logic in the host computer system, intelligent analysis and automated classification of defects are realized; image preprocessing eliminates interference, connected component analysis and feature matching accurately identify defect types and locations, the level judgment criteria are clear, and combined with the sorting execution mechanism controlled by PLC, the guardrail classification can be completed quickly, replacing manual judgment and improving analysis efficiency and accuracy.

[0044] 3. In this solution, the real-time linkage between the host computer and the slave computer based on industrial protocols (such as the synchronization of slave computer equipment timing control with host computer image analysis, and real-time interaction of detection data and sorting instructions) realizes the closed-loop automation of the entire process from guardrail loading to classification and storage; the slave computer accurately executes physical operations, and the host computer efficiently processes data and issues instructions. The collaboration between the two enables seamless connection between detection and sorting, greatly improving production efficiency and reducing manual intervention costs. Attached Figure Description

[0045] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0046] Figure 1 This is a schematic diagram of the overall structure of the lower-level machine device of the present invention;

[0047] Figure 2 This is a top view of the overall structure of the lower-level machine device of the present invention;

[0048] Figure 3 This is a schematic diagram of the coating equipment structure of the present invention;

[0049] Figure 4 This is a schematic diagram of the telescopic rod structure of the present invention;

[0050] Figure 5 This is a schematic diagram of ink retention in the protruding defect area of ​​the present invention;

[0051] Figure 6 This is a schematic diagram of ink retention in the concave defect area of ​​the present invention;

[0052] Figure 7 This is a schematic diagram of the host computer system framework of the present invention;

[0053] Figure 8 This is a flowchart of the method steps of the present invention.

[0054] Legend: 1. Frame; 2. Electric push rod; 3. Coating equipment; 3-1. Ink bladder; 3-2. Ink stick; 4. Ink wiping bar; 5. Scraper bar; 6. Image acquisition equipment; 7. Conveyor belt; 8. Telescopic rod; 8-1. Sleeve rod; 8-2. Slide rod; 8-3. Spring. Detailed Implementation

[0055] Preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, the present invention can also be implemented in various different forms, and therefore the present invention is not limited to the embodiments described below. In addition, for the purpose of more clearly describing the present invention, parts not connected to the invention will be omitted from the drawings.

[0056] The technical solution in this application embodiment is to solve the problems mentioned in the background art, and the overall idea is as follows:

[0057] Example 1:

[0058] Please refer to Figures 1 to 7 As shown in the figure, this embodiment introduces a flatness detection system based on high-toughness PVC guardrails. The system achieves fully automated detection through a combination of "lower-level equipment + upper-level analysis system", as detailed below:

[0059] The lower-level equipment includes three electric push rods 2 mounted on the frame 1. The output ends of the left and right electric push rods 2 are respectively equipped with coating equipment 3 and ink wiping strips 4. The coating equipment 3 includes an ink sac 3-1 and an ink stick 3-2. When the left electric push rod 2 drives the coating equipment 3 to move downward, it will put the ink stick 3-2 into contact with the surface of the guardrail. The ink stick 3-2 draws ink from the ink sac 3-1 and coats it on the surface of the guardrail.

[0060] The electric push rod 2 on the right side drives the ink wiping strip 4 to move down and fit against the surface of the guardrail, and the ink wiping strip 4 removes the residual ink from the surface of the guardrail.

[0061] The output end of the electric push rod 2 in the middle is equipped with two scraper blades 5 via a telescopic rod 8. Both scraper blades 5 are designed with a 45° inclination. The two scraper blades 5 scrape away the ink on the surface of the guardrail. The 45° inclination angle allows the scraper blades 5 to form a more reasonable contact posture with the surface of the guardrail, increasing the effective contact area and making the scraping action more in line with the mechanical principle. When the scraper blades 5 move along the guardrail, the inclination angle can guide the ink to gather in a specific direction and be removed, reducing the probability of residual ink remaining on the surface of the guardrail. Especially for the trace amount of ink that may still exist after the scraper blades 5 are scraped, it can be wiped more thoroughly, avoiding interference with the recognition of the subsequent imaging device 6.

[0062] The telescopic rod 8 includes a sleeve rod 8-1 and a sliding rod 8-2 slidably connected inside the sleeve rod 8-1. A spring 8-3 is provided between the top of the sliding rod 8-2 and the inner wall of the sleeve rod 8-1. The sliding rod 8-2 is pressed down by the sliding rod 8-2, so that the scraper 5 is always in contact with the surface of the guardrail. When scraping the ink, the telescopic rod 8 is driven by the electric push rod 2 in the middle, which in turn drives the two scraper 5 to stick to the surface of the guardrail. When the conveyor belt 7 moves the guardrail, the ink on its surface is scraped off by the scraper 5.

[0063] If the surface of the guardrail has unevenness, such as Figure 5 When the scraper 5 moves across the surface of the guardrail, it will scrape off the ink on a flat surface. If it encounters a raised area, the raised area will push the scraper 5 upward, preventing the ink parallel to both sides of the scraper 5 from being scraped off, thus leaving ink residue. When the scraper 5 moves out of the raised area, it continues to contact the guardrail surface under its own weight and the rebound force of the spring 8-3, continuing to scrape off the ink on the guardrail surface.

[0064] like Figure 6 When the scraper 5 scrapes the ink off the surface of the guardrail, it cannot penetrate into the recessed area and therefore cannot scrape off the ink in the recessed area, while the ink is scraped off in other flat areas.

[0065] Therefore, by identifying whether there is residual ink on the surface of the guardrail, it can be determined whether the guardrail is flat, and by observing the shape of the residual ink, it can be determined whether there are bumps or depressions and the location of abnormal areas.

[0066] In addition, the ink can be water-based environmentally friendly ink. Water-based environmentally friendly ink uses water as a solvent and a small amount of environmentally friendly resin (such as acrylic resin) to provide adhesion. Its adhesion is at a "moderate to weak" level. For flat areas of the guardrail, the scraper 5 is pressed tightly against the surface by the spring 8-3 of the telescopic rod 8, which can easily scrape off the ink (because the ink does not penetrate deeply, it only temporarily adheres to the surface). For raised areas, after the scraper 5 is lifted, the ink on both sides of the raised area can remain in an adhesive state because it is not subjected to the pressure of the scraper 5. For recessed areas, the scraper 5 cannot penetrate deeply, and the ink in the recessed area remains because it is not scraped.

[0067] In addition, when the guardrail is not placed evenly (such as tilted, partially raised, etc.), the slide bar 8-2 can slide flexibly within the sleeve bar 8-1. Combined with the elastic force of the spring 8-3, the scraper 5 is always in close contact with the surface of the guardrail, avoiding gaps between the scraper 5 and the guardrail due to the skewed placement of the guardrail. This ensures the continuity and effectiveness of the ink scraping action. At the same time, the elasticity of the spring 8-3 can buffer the local pressure changes caused by the uneven placement of the guardrail, preventing the scraper 5 from being damaged by the additional force generated by the tilt of the guardrail. It also reduces the impact on the conveyor belt 7 and the overall equipment, ensuring the stability of the detection process.

[0068] In addition, the electric push rod 2 on the right side drives the ink wiping strip 4 to further remove residual ink, and the water-soluble properties of water-based ink make it easy to wipe off with the ink wiping strip 4;

[0069] In addition, the water-based environmentally friendly ink has a flexible coloring system, and high-contrast colors can be customized according to the color of the guardrail to meet the "distinguishing recognition" requirements.

[0070] The identification is performed by an image acquisition device 6 installed on the shelf 1 near the ink wiping strip 4, and the image is then further identified and judged by the host computer system.

[0071] The frame 1 is provided with three conveyor belts 7 on its lower side. The left conveyor belt 7 is located on the lower side of the coating equipment 3, and the middle conveyor belt 7 is located on the lower side of the scraper 5 and the ink wiping strip 4. The left conveyor belt 7 and the middle conveyor belt 7 are separated to prevent excess ink from dripping onto the surface of the conveyor belt 7 when the coating equipment 3 applies ink to the guardrail surface, which would then affect the image acquisition and analysis on the rear side.

[0072] The right-side conveyor belt 7 can be extended into a sorting conveyor to classify and standardize the guardrails in different situations.

[0073] It should be noted that, due to the special shape design of the guardrail, such as the wave pattern, the sides of the ink rod 3-2, scraper 5, and wiping strip 4 of the coating equipment 3 that come into contact with the guardrail are all designed to fit the shape of the guardrail surface. When the guardrail is placed on the conveyor belt 7, the wave pattern of the guardrail needs to be initially aligned with the wave pattern of the ink rod 3-2, scraper 5, and wiping strip 4 of the coating equipment 3. Slight deviations will be offset by the pressure generated when the electric push rod 2 pushes the coating equipment 3, scraper 5, and wiping strip 4 downwards, which will cause the guardrail to move and thus achieve the final alignment.

[0074] The host computer analysis system includes an image acquisition unit, an image analysis unit, and a sorting control unit;

[0075] 1. Image Acquisition Unit

[0076] The lower-level device linkage control is based on a preset detection process and establishes real-time data interaction with the lower-level device through an industrial-grade communication protocol (such as Modbus TCP) to accurately control the timing of each hardware component.

[0077] Conveyor belt 7 control: Segmented speed adjustment is adopted. The speed of the left conveyor belt 7 (coating area) is set to 0.6m / s. The middle conveyor belt 7 (scraping, imaging, and wiping area) is synchronized with the left side. The right sorting conveyor belt 7 switches speeds (0.4-0.8m / s) according to sorting requirements. The position of the guardrail is located by photoelectric sensors to ensure that there is no delay in the connection of actions at each station.

[0078] Electric linear actuator 2 control:

[0079] Left electric push rod 2: After receiving the guardrail positioning signal, it drives the coating equipment 3 to move down, so that the ink rod 3-2 adheres to the guardrail with 8N pressure, ensuring uniform ink coverage;

[0080] The middle electric push rod 2: drives the telescopic rod 8 and the scraper 5 to move down. Through the pressure signal fed back by the spring 8-3 (maintaining 10-15N), it ensures that the 45° inclined scraper 5 is in close contact with the guardrail and moves with the conveyor belt 7 to complete the scraping action.

[0081] Right electric push rod 2: When the guardrail enters the ink wiping area, it drives the ink wiping strip 4 to move down and press against the surface with 5N pressure, and works with the conveyor belt 7 to remove residual ink in the direction;

[0082] Image acquisition:

[0083] Image capture device 6 control: It adopts a 10-megapixel industrial camera (equipped with a macro lens), and triggers shooting 0.3 seconds after the ink scraping is completed (to avoid the shadow of scraping strip 5). The fill light is turned on during shooting to ensure clear imaging of the ink residue area. The image data is uploaded to the local storage server in real time.

[0084] With environmental adaptive adjustment, the imaging device 6 has a built-in light sensor that can automatically adjust the exposure time (5-20ms) and light source brightness according to the reflectivity of the guardrail, avoiding image distortion caused by differences in the gloss of the PVC guardrail surface;

[0085] 2. Image analysis unit: Based on the image of the guardrail surface acquired by image acquisition device 6, the unit uses algorithms to identify areas with residual ink, determine whether there are any uneven defects on the guardrail surface, and clarify the defect type (protrusion / depression), location, and level, providing a precise control basis for the sorting unit.

[0086] 2.1 Image Preprocessing Module

[0087] Image cropping and alignment: edge detection algorithms (such as the Canny operator) are used to identify the edge contour of the guardrail and automatically crop out the background area of ​​conveyor belt 7, retaining the complete guardrail image; the image tilt angle is corrected by Hough transform (because the guardrail image may be tilted due to the slight offset of conveyor belt 7), ensuring that subsequent analysis is based on a horizontal reference.

[0088] Color standardization is performed on the background color of the guardrail (such as green) and the ink color (such as black, blue, and other high-contrast colors). Ambient light interference is eliminated through white balance correction. Then, color space conversion is used (converting the RGB image to HSV space) to extract the hue (H) and saturation (S) features of the ink area, ignoring the influence of brightness (V) fluctuations, to achieve "grayscale standardization" (i.e., uniformly converting the ink area to black and the guardrail area to white to eliminate color difference interference).

[0089] Noise filtering uses Gaussian filtering to remove random noise from the image, and combines morphological operations (such as erosion-dilation) to eliminate burrs at the edges of ink residue, ensuring clear contours of defective areas.

[0090] 2.2 Defect Identification and Classification Module: This module involves three basic algorithms: connected component analysis algorithm, feature matching algorithm, and geometric parameter calculation algorithm.

[0091] The connected component analysis algorithm is used to extract all black ink residue areas from the preprocessed binarized image (ink residue is black, and the flat area of ​​the guardrail is white), and to calculate the geometric parameters such as the area, perimeter, and circumscribed rectangle of each area, providing a data basis for subsequent defect judgment; for example, when there is a protrusion in the guardrail, the algorithm can extract the symmetrical ink residue strips on both sides of the protrusion and obtain its length, width and other parameters.

[0092] The feature matching algorithm performs feature matching on the extracted ink residue areas based on the correspondence between the ink residue morphology and defect type in the file (protrusions correspond to "symmetrical double-band" ink residue, and depressions correspond to "single-area closed-form" ink residue). For example, it determines whether the residue area meets the "symmetrical double-band" feature of a protrusion defect by calculating the symmetry parameters (such as the symmetry axis deviation) of the residue area; and it determines whether the residue area meets the feature of a depression defect by detecting whether the residue area is an independent closed area.

[0093] The geometric parameter calculation algorithm is used to quantify the geometric features of the ink residue area, such as calculating the total length and single width of the raised residue strips, and the area of ​​the recessed residue area. These parameters are important bases for determining the defect level. For example, when there is a recess in the guardrail, the algorithm can calculate the area of ​​the ink residue area in the recess and then compare it with the preset level threshold to determine the defect level.

[0094] Specifically, the process involves: extracting ink residue areas based on a binary image (ink is black and guardrail is white) after color standardization preprocessing; identifying all black areas (i.e. ink residue areas) using a connected component analysis algorithm; and calculating the area, perimeter, bounding rectangle, and other geometric parameters of each area.

[0095] For example, there are two black areas in the image: one is a "parallel double stripe" (approximately 0.2cm wide and 3cm long), and the other is an "irregular closed region" (approximately 1.2cm² in area). Using a connected component analysis algorithm, the geometric parameters of these two regions are extracted: Double stripe region: the total number of pixels is converted to an area of ​​0.6cm², with a perimeter of 6.4cm, and the distance between the two stripes matches the width of the scraping strip 5 (preset 1cm); Closed region: the number of pixels is converted to an area of ​​1.2cm², with a perimeter of 4.5cm, and the edges are curved.

[0096] Defect type determination:

[0097] Raised Defect: Based on the feature of "ink residue on both sides of the raised area" in the document, the ink residue area corresponding to the raised area is in the form of "symmetrical double stripes" (located on both sides of the raised area), and the distance between the two stripes matches the width of the scraper 5 (because when the scraper 5 is lifted, the unscraped areas on both sides are symmetrically distributed); by calculating the symmetry parameters of the residue area (such as the deviation of the symmetry axis < 5%) and the strip length (positively correlated with the length of the raised area), it is determined to be a raised defect;

[0098] For example, the above-mentioned "parallel double stripe" area meets the characteristics of "ink residue on both sides of the protrusion" in the document: when the scraper 5 encounters a 1mm high protrusion on the guardrail surface, the protrusion pushes the scraper 5 upward, causing the ink that was not scraped off on both sides of the protrusion to form symmetrically distributed stripes (similar to "ink on both sides of the protrusion is retained because it is not subjected to the pressure of the scraper 5"); according to the algorithm calculation, the symmetry axis deviation of the double stripe is 3% (<5%), which meets the morphological characteristics of "symmetrical double stripe", so this area is determined to be a protrusion defect;

[0099] Indentation defect: Based on the characteristic that "the ink in the indentation is not scraped off", the ink residue area corresponding to the indentation has a "single-area closed shape" (consistent with the shape of the indentation), and the area of ​​the area is positively correlated with the size of the indentation; by detecting whether the residue area is an independent closed area (without symmetrical strip features), it is determined to be an indentation defect.

[0100] For example, the above-mentioned "irregular closed area" meets the feature of "ink not being scraped off in the depression" in the document: when there is a depression with a diameter of 0.8cm and a depth of 0.5mm on the surface of the guardrail, the scraper 5 cannot penetrate into the depression, resulting in the ink remaining in the depression and forming a closed area with the same shape as the depression (without symmetrical strip features); the algorithm detects that this area is an independent closed shape, so it is judged as a depression defect;

[0101] For defect location, establish a coordinate system with the upper left corner of the guardrail as the origin. Calculate the specific location of the defect on the guardrail surface (e.g., 30cm from the left end on the X-axis and 20cm from the front end on the Y-axis) by using the coordinates of the center point of the circumscribed rectangle of the residual area. Mark it as "protruding area (x1, y1, x2, y2)" or "recessed area (x3, y3, x4, y4)".

[0102] For example, establish a two-dimensional coordinate system (unit: cm) with the top left corner of the guardrail as the origin (0, 0):

[0103] The coordinates of the upper left corner of the rectangle circumscribed by the double stripe area of ​​the above-mentioned protrusion defect are (10.5, 5.2), and the coordinates of the lower right corner are (13.5, 6.2). Therefore, it is marked as "protrusion area (10.5, 5.2, 13.5, 6.2)", indicating that the protrusion is located within the range of 10.5-13.5cm from the left end and 5.2-6.2cm from the front end of the guardrail.

[0104] The coordinates of the center point of the circumscribed rectangle of the closed area of ​​the above-mentioned dent defect are (20.3, 8.7), so it is marked as "dent area (19.9, 8.3, 20.7, 9.1)", indicating that the dent is located within the range of 19.9-20.7cm from the left end and 8.3-9.1cm from the front end of the guardrail;

[0105] 2.3 Defect Level Determination Module

[0106] The grading standard is based on the geometric parameters of the residual area, setting n grades (3 grades in the example):

[0107] Level 1 (Non-destructive): No ink residue, or residue area < 0.5cm² 2 (The interference may be due to minor impurities; the result is deemed acceptable.)

[0108] Level 2 (Repairable): Total length of raised residual strips < 5cm and width of each strip < 0.3cm; or area of ​​dented residual area 0.5-2cm. 2 The defects are classified as minor and repairable. For example, the total length of the two strips is 3cm (<5cm), and the width of each strip is 0.2cm (<0.3cm), which meets the "Level 2 (Repairable)" standard. Therefore, it is classified as a repairable protruding defect; the area of ​​the closed region is 1.2cm². 2 (in the range of 0.5-2cm) 2 The area meets the "Level 2 (Repairable)" standard and is therefore judged as a repairable dent defect.

[0109] Level 3 (Irreversible): Total length of raised residual strips ≥ 5cm or width of a single strip ≥ 0.3cm; or area of ​​dented residual area ≥ 2cm. 2 If the total length of the raised strip of another guardrail is 6cm (≥5cm), it is directly judged as "Level 3 (Unrepairable)";

[0110] (Note: The specific threshold can be adjusted according to the guardrail specifications and industry standards, and the number of n can be expanded.)

[0111] Automatic rating matching: The extracted residual area parameters are compared with the rating standard to automatically determine the rating of the guardrail and generate a defect report (including type, location, and rating).

[0112] The data output and linkage module outputs analysis results in JSON format. Example:

[0113] {“Fence ID”:“PVC20230801001”,

[0114] "Defect Type": "Protrusion"

[0115] "Defect Location": "(15.2, 8.3, 20.5, 10.1)",

[0116] Level: "Level 2"

[0117] "Judgment Result": "Repairable"

[0118] Linked with the sorting unit, the judgment result is sent to the sorting control unit in real time via an industrial bus (such as Profinet), triggering the sorting action of the right conveyor belt 7 (such as Level 1 products entering the qualified area, Level 2 products entering the repair area, and Level 3 products entering the waste area), realizing an automated closed loop of detection and sorting.

[0119] 3. The sorting and control unit, based on the defect level of the guardrail (such as non-destructive, repairable, and irreparable) output by the image analysis unit, realizes the automated classification and conveying of guardrails of different levels through the right conveyor belt 7 and the matching actuator, ensuring a closed loop of the detection-sorting process and improving production efficiency.

[0120] After the guardrail passes under the image acquisition device 6 and the image analysis is completed, the image analysis unit sends the defect level data (such as "Level 1 - No damage" or "Level 2 - Repairable") to the sorting unit PLC in real time.

[0121] The PLC matches the received grade data with the actual guardrails arriving at the sorting section based on the transmission time of the guardrails on the conveyor belt 7 (calculated based on the speed of the conveyor belt 7, which is 0.4-0.8 m / s). (Matching can be assisted by the timestamps of guardrail edge detection.)

[0122] Location-triggered:

[0123] When the guardrail enters the right-side conveyor belt 7 sorting section, the first photoelectric sensor is triggered, the PLC records the guardrail position, and the timing begins.

[0124] When the guardrail reaches the entrance of the sorting section, the secondary positioning sensor detects the signal, and the PLC confirms that the guardrail is in a sortable position and is ready to perform the sorting action.

[0125] Graded sorting execution:

[0126] Based on the defect level, the PLC controls the corresponding sorting mechanism's actions:

[0127] If it is level 1 (non-destructive): The first group of pneumatic push rods in the diversion section extends, pushing the guide plate to tilt towards the level 1 branch conveyor belt 7. The guardrail moves with the conveyor belt 7 into the qualified area branch, and the push rods then return to their original positions.

[0128] If it is level 2 (repairable): the second group of electric swing arms in the diversion section swings, guiding the guardrail to the level 2 repair area branch, and the swing arm resets after 3 seconds;

[0129] If it is level 3 (unrepairable): the third group of actuators in the diversion section will move and push the guardrail into the level 3 waste area branch;

[0130] (Note: The actuator's action time is ≤0.5 seconds to ensure that the guide is completed before the guardrail completely passes through the diversion section. The action force is adjustable to avoid damaging the guardrail.)

[0131] Categorized storage:

[0132] The guardrails entering the branch conveyor belt 7 are blocked by the limit baffle at the end. After 5-10 pieces are accumulated, the branch conveyor belt 7 automatically stops (detected by the end sensor). It restarts after being manually removed, thus realizing batch storage.

[0133] Example 2:

[0134] Please refer to Figure 8 As shown in the figure, this embodiment introduces a flatness detection method based on high-toughness PVC guardrails. The method includes the following steps:

[0135] S1. Applying ink and scraping ink:

[0136] S1.1 Ink Coating: The left electric push rod 2 drives the coating device 3 to move down, so that the ink rod 3-2 adheres to the surface of the guardrail with 8N pressure. The ink rod 3-2 draws water-based environmentally friendly ink (high contrast color can be customized according to the color of the guardrail) from the ink sac 3-1 and coats it evenly on the surface of the guardrail. The left conveyor belt 7 moves the guardrail at a speed of 0.6m / s to complete the ink coating.

[0137] S1.2, Scraping: The central electric push rod 2 drives the telescopic rod 8 and the two 45° inclined scraper blades 5 to move downwards. The springs 8-3 maintain a pressure of 10-15N, so that the scraper blades 5 are in close contact with the guardrail surface. The central conveyor belt 7 runs synchronously with the left side at a speed of 0.6m / s, driving the guardrail to move, and the scraper blades 5 scrape off the ink.

[0138] On the flat areas of the guardrail: the ink has been completely scraped off;

[0139] The raised area: The raised area lifts up the scraper 5, and the ink residue on both sides of the raised area forms symmetrical double stripes; after the scraper 5 moves out of the raised area, it returns to its original position under the force of gravity and the pushing force of the spring 8-3, and continues to scrape ink.

[0140] In the recessed area: Squeegee 5 cannot penetrate the recess deeply, and ink remains inside the recess, forming a closed area.

[0141] S2, Image Capture:

[0142] The image acquisition device 6 (a 10-megapixel industrial camera with a macro lens) is installed on the side close to the ink wiping strip 4. It triggers the shooting 0.3 seconds after the ink is wiped off to avoid the shadow of the wiping strip 5. When shooting, the fill light is turned on, and the built-in light sensor automatically adjusts the exposure time (5-20ms) and the brightness of the light source according to the reflectivity of the guardrail to ensure clear imaging of the ink residue area. The image data is uploaded to the host computer system in real time.

[0143] S3, Analysis:

[0144] S3.1 Image Processing:

[0145] Image cropping and alignment: The Canny operator is used to identify the edges of the guardrail and crop out the background of the conveyor belt 7; the Hough transform is used to correct the image tilt angle to ensure analysis based on a horizontal reference.

[0146] Color standardization: By eliminating ambient light interference through white balance, the RGB image is converted to HSV space, the hue (H) and saturation (S) features of the ink area are extracted, the brightness (V) fluctuation is ignored, the ink area is converted to black and the guardrail area is converted to white, thus achieving gray-white standardization;

[0147] Noise filtering: Gaussian filtering is used to remove random noise, and corrosion-dilation morphological operations are combined to eliminate ink residue edge burrs, making the defect area outline clear;

[0148] S3.2 Defect Identification and Classification:

[0149] Ink residue region extraction: Based on the binarized image, all black residue regions are identified through connected component analysis algorithm, and geometric parameters such as area, perimeter, and bounding rectangle of each region are calculated.

[0150] Defect type determination:

[0151] Protrusion defect: The residual area has a "symmetrical double stripe" shape, the distance between the two stripes matches the width of the scraper 5, and the deviation of the symmetry axis is <5%, which is judged as a protrusion;

[0152] Depression defect: The residual area is an independent closed region without symmetrical strip features, and is judged as a depression;

[0153] Defect location: Establish a coordinate system with the upper left corner of the guardrail as the origin, and mark the defect location by the coordinates of the center point of the circumscribed rectangle of the residual area (e.g., "protruding area (x1, y1, x2, y2)" "depressed area (x3, y3, x4, y4)").

[0154] S3.3 Defect Level Determination:

[0155] Level 1 (Non-destructive): No ink residue, or residue area < 0.5cm² 2 (Determined to be qualified);

[0156] Level 2 (Repairable): Total length of raised residual strips < 5cm and width of each strip < 0.3cm, or area of ​​dented residual strips 0.5-2cm. 2 (Judged as a minor defect);

[0157] Level 3 (Irreversible): The total length of raised residual strips is ≥5cm or the width of a single strip is ≥0.3cm, or the area of ​​dented residual strips is ≥2cm² (judged as a serious defect).

[0158] Generate a JSON report containing the defect type, location, and severity level;

[0159] S4. Ink wiping: The electric push rod 2 on the right side drives the ink wiping strip 4 to adhere to the surface of the guardrail with a pressure of 5N. It moves with the middle conveyor belt 7 to further remove residual ink (using the water-soluble properties of water-based ink).

[0160] S5. Sorting:

[0161] Data matching: The right conveyor belt 7 receives the analyzed guardrail. The PLC calculates the transmission time based on the speed of the conveyor belt 7 (0.4-0.8m / s) and matches the defect level data with the guardrail (combined with the guardrail edge detection timestamp).

[0162] Positioning trigger: When the guardrail enters the right conveyor belt 7 sorting section, the first photoelectric sensor is triggered; when it reaches the diversion section entrance, the secondary positioning sensor confirms the position and prepares for sorting.

[0163] Tiered implementation:

[0164] Level 1 (non-destructive): The first group of pneumatic push rods in the diversion section extends and pushes the guide plate to guide the guardrail into the qualified area branch conveyor belt 7;

[0165] Level 2 (Repairable): The second set of electric swing arms swings to guide the guardrail to the repair area branch;

[0166] Level 3 (Unrepairable): The third actuator moves to push the guardrail into the waste area branch; the actuator action time is ≤0.5 seconds, and the force is adjustable;

[0167] Categorized storage: The end limit baffle of the branch conveyor belt 7 blocks the guardrail. It will automatically stop after accumulating 5-10 pieces. It will restart after being manually removed, thus realizing batch storage.

[0168] Finally, it should be noted that the above embodiments are merely examples for clearly illustrating the present invention and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A flatness detection system based on high-toughness PVC guardrails, comprising lower-level equipment and upper-level analysis system, characterized in that: The lower-level equipment includes a frame (1), on which three electric push rods (2) are installed. The output ends of the left and right electric push rods (2) are respectively equipped with a coating device (3) and an ink wiping strip (4). The coating device (3) includes an ink sac (3-1) and an ink stick (3-2). The output end of the middle electric push rod (2) is equipped with two scrapers (5) inclined at 45° via a telescopic rod (8). An image acquisition device (6) is installed on the side of the frame (1) near the ink wiping strip (4). Three conveyor belts (7) are provided on the lower side of the frame (1), corresponding to the coating, ink scraping and wiping and sorting areas respectively. The ink stick (3-2), scraper (5), and ink wiping strip (4) of the coating equipment (3) are all designed to fit the shape of the guardrail surface on the side that contacts the guardrail. The host computer analysis system includes an image acquisition unit, an image analysis unit, and a sorting control unit. The image analysis unit includes a three-level processing module: an image preprocessing module, a defect identification and classification module, and a defect level determination module. The defect identification and classification module, based on the preprocessed image, extracts features of the ink residue area using three basic algorithms: connected component analysis, feature matching, and geometric parameter calculation, thereby identifying the defect type and determining its location. The specific process is as follows: Ink residue region extraction: Based on the color-normalized binarized image, all black regions are identified through connected component analysis algorithm, and the area, perimeter, and bounding rectangle parameters of each region are calculated. Defect type determination: Raised defects: Based on the characteristic of "ink residue on both sides of the raised area", the ink residue area corresponding to the raised area is in the form of "symmetrical double stripes" and the distance between the two stripes matches the width of the scraper. The determination is made by calculating the symmetry parameters of the residue area and the strip length. Indentation defect: Based on the characteristic that "the ink in the indentation has not been scraped off", the ink residue area corresponding to the indentation is in the form of "single-area closed shape". It is determined by detecting whether the residue area is an independent closed area. Defect location: Establish a coordinate system with the upper left corner of the guardrail as the origin. Calculate the specific location of the defect on the guardrail surface using the coordinates of the center point of the circumscribed rectangle of the residual area, and mark it.

2. The flatness detection system based on high-toughness PVC guardrail as described in claim 1, characterized in that, The telescopic rod (8) consists of a sleeve rod (8-1), a sliding rod (8-2), and a spring (8-3). The sliding rod (8-2) is slidably nested inside the sleeve rod (8-1). The two ends of the spring (8-3) are respectively connected to the top of the inner wall of the sleeve rod (8-1) and the top of the sliding rod (8-2). The scraper (5) can adaptively conform to the surface of the guardrail through elastic deformation.

3. The flatness detection system based on high-toughness PVC guardrail as described in claim 2, characterized in that, The coating device (3) has an ink sac (3-1) containing water-based environmentally friendly ink. The ink uses water as a solvent and contains acrylic resin. The ink stick (3-2) is made of porous elastic material, which can quantitatively draw ink and form a uniform coating under 8N pressure.

4. The flatness detection system based on high-toughness PVC guardrail as described in claim 1, characterized in that, The three conveyor belts (7) are segmented structures with independent drive. The distance between the left conveyor belt and the middle conveyor belt is ≥5cm. The running speed of the left and middle conveyor belts is 0.6m / s. The right conveyor belt can be steplessly adjusted in the range of 0.4-0.8m / s by PLC control.

5. The flatness detection system based on high-toughness PVC guardrail as described in claim 1, characterized in that, The image acquisition unit of the host computer analysis system establishes real-time communication with the lower-level device through the Modbus TCP industrial protocol, and can synchronously control the conveyor belt speed, electric push rod pressure and the triggering timing of the image acquisition device, and receive the device status data fed back by the lower-level device.

6. The flatness detection system based on high-toughness PVC guardrail as described in claim 1, characterized in that, The defect level determination module determines the level of the identified defects based on a preset level classification standard and outputs the relevant results, as follows: Grading criteria: Three levels are set based on the geometric parameters of the residual area: Level 1 Non-destructive: No ink residue, or residue area < 0.5cm² 2 The issue may be due to minor impurities, and the test result is deemed acceptable. Level 2 Repairable: Total length of raised residual strips < 5cm and width of each strip < 0.3cm; or residual area of ​​depressions 0.5-2cm. 2 The defect was determined to be minor and repairable. Level 3, irreparable: Total length of raised residual strips ≥ 5cm or width of a single strip ≥ 0.3cm; or area of ​​dented residual area ≥ 2cm². 2 It was determined to be a serious defect and irreparable; Automatic matching of grade and output of results: The extracted residual area parameters are compared with the grade standard to automatically determine the grade of the guardrail, generate a defect report containing type, location and grade, output in JSON format, and link with the sorting unit to trigger the corresponding sorting action.

7. The flatness detection system based on high-toughness PVC guardrail as described in claim 1, characterized in that, The sorting control unit of the host computer analysis system is linked to the PLC controller of the right conveyor belt via the Profinet industrial bus, and can trigger the corresponding actuators according to the 1, 2, and 3 defect levels output by the image analysis unit: The Level 1 signal triggers the first group of pneumatic push rods, with a response time of ≤0.5 seconds, which pushes the guide plate to guide the guardrail into the qualified branch area; The Level 2 signal triggers the second group of electric swing arms, which swing at a 30° angle to guide the guardrail into the repair area branch. The level 3 signal triggers the third waste pushing mechanism, which moves the guardrail to the waste area; Meanwhile, this unit can record the quantity and inspection time data of guardrails at each level, and supports data traceability by connecting with the production management system.

8. A method for detecting the flatness of a high-toughness PVC guardrail, applicable to the system described in any one of claims 1 to 7, characterized in that, The method includes the following steps: S1. Ink application and scraping: During ink application, the left electric push rod (2) drives the coating equipment to move down, and the ink rod (3-2) adheres to the guardrail with 8N pressure, absorbs water-based environmentally friendly ink, and completes uniform ink application on the left conveyor belt (7) running at 0.6m / s; During the scraping stage, the middle electric push rod (2) passes through the telescopic rod (8) with spring (8-3) to make the two 45° inclined scraping strips (5) adhere to the guardrail with 10-15N pressure, and move with the synchronously running middle conveyor belt (7). The ink is scraped off in flat areas, symmetrical double strip ink marks are left in raised areas, and closed ink marks are left in recessed areas; S2, Image Acquisition: The guardrail after ink coating and scraping is transported to the image acquisition area. The 10-megapixel industrial camera triggers shooting 0.3 seconds after the ink scraping is completed. It uses a ring light and macro lens to acquire images of the guardrail surface. The light sensor adjusts the exposure time and light source brightness in real time from 5-20ms to ensure image clarity. S3. Analysis: After receiving the image, the host computer system first performs noise reduction by Gaussian filtering, corrects the tilt angle by Hough transform and converts it to the HSV color space. Then, it uses connected component analysis to extract parameters of the ink residue area, identifies raised or sunken defects by feature matching, and finally determines the defect level based on the preset threshold and generates an analysis report containing the defect location. S4, Ink wiping: The electric push rod (2) on the right side drives the ink wiping strip (4) to adhere to the guardrail surface with a pressure of 5N, and moves with the middle conveyor belt to further remove residual ink; S5. Sorting: The right conveyor belt (7) according to the defect level instruction output by the host computer, through the corresponding actuator, diverts the guardrails to the qualified area, repair area or waste area respectively, with an action response time of ≤0.5 seconds.

9. The method for detecting the flatness of a high-toughness PVC guardrail as described in claim 8, characterized in that, The image preprocessing in step S3 also includes grayscale normalization, which converts the ink area to black and the guardrail area to white; the defect location marking accuracy is ±0.1cm, and the coordinates are represented in the format "(x1, y1, x2, y2)"; the JSON report contains the guardrail ID, defect type, location and level information, and is sent to the sorting unit in real time via the industrial bus, triggering the actuator to complete the action within ≤0.5 seconds, and the action force can be adjusted to avoid damaging the guardrail.

Images