PVC guardrail surface flatness detection system and method

Abstract

The invention relates to the technical field of material determination, in particular to a PVC guardrail surface flatness detection system and method, and the system comprises a lower computer and an upper computer; the lower computer comprises a conveying platform, a camera obscura, a track, a strip-shaped light source and a camera; the upper computer comprises an image signal acquisition module, an image preprocessing module, an image feature extraction module, an abnormal region analysis module and a flatness judgment module; according to the scheme, the different positions of the guardrail are collected for multiple times, the single-time scanning images collected for multiple times are spliced into the whole guardrail image to be analyzed by the upper computer, the whole guardrail surface can be covered through the multiple-time collection and splicing mode, image information missing caused by a small single-time collection area is avoided, and the image quality is improved. The mode of collecting and splicing the images for multiple times has very high flexibility, and the detection system can adapt to the detection requirements of guardrails of various different specifications by adjusting the collection interval and number of times no matter how the size of the guardrail changes.

Description

Technical Field

[0001] The present invention relates to the technical field of material determination, and in particular to a system and method for detecting the surface flatness of a PVC guardrail. Background Art

[0002] The PVC guardrail surface flatness detection system is mainly used to accurately detect the flatness of the PVC guardrail surface. Traditional detection systems often use a single or limited number of image acquisition methods. Due to the limited field of view of the camera, it is difficult to completely cover the entire PVC guardrail surface. For long guardrails, this acquisition method is prone to missing surface information in some areas, resulting in an incomplete data basis for subsequent analysis and an inability to accurately evaluate the overall flatness of the guardrail.

[0003] Therefore, the present invention proposes a system and method for detecting the flatness of the surface of a PVC guardrail. By performing multiple acquisitions at different positions of the guardrail and splicing the multiple acquired single scan images into an overall guardrail image, a comprehensive and accurate data basis is provided for host computer analysis. The system can completely cover the surface of the guardrail, improve the detection accuracy, and adapt to guardrails of different specifications. At the same time, through reasonable acquisition planning and effective splicing algorithms, the quality of the spliced ​​image is guaranteed, thereby realizing accurate detection and evaluation of the surface flatness of the PVC guardrail. Summary of the invention

[0004] Technical problem solved: Traditional detection systems often use a single or limited number of image acquisition methods, which makes it impossible to accurately evaluate the overall flatness of the guardrail.

[0005] In view of the deficiencies in the prior art, the present invention provides a system and method for detecting the surface flatness of a PVC guardrail, thereby solving the technical problems mentioned in the background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0007] A PVC guardrail surface flatness detection system, the system comprising a lower computer and an upper computer; the lower computer comprises a conveying platform, a dark box, a track, a strip light source, and a camera, and the upper computer comprises an image signal acquisition module, an image preprocessing module, an image feature extraction module, an abnormal area analysis module, and a flatness judgment module;

[0008] The conveying platform is used to convey the PVC guardrail at a uniform speed to the bottom of the dark box. The dark box is made of opaque material to seal the internal environment to reduce external light interference. The track is installed on the upper side of the dark box for installing a strip light source and a camera. The strip light source uses an LED strip light source with adjustable illumination angle and intensity. The camera is used to collect light and shadow images on the surface of the PVC guardrail and transmit them to the host computer.

[0009] In a possible implementation, the image signal acquisition module includes: a data receiving unit, which receives single-shot acquired image data in real time through a communication link with a lower computer; and a cache management unit, which adopts a circular queue method to store the received image data in a memory cache area.

[0010] In a possible implementation, the image preprocessing module includes: an image combination unit, which arranges the single-shot acquired images in the cache area in the order of picture acquisition, and uses a feature matching algorithm to perform geometric transformation and pixel fusion on adjacent images according to a 5% overlap area setting to obtain a complete image; an image processing unit, which converts the combined color image into a grayscale image, uses Gaussian filtering to remove noise, and adjusts the histogram to enhance contrast.

[0011] In a possible implementation, the image feature extraction module includes: an edge detection unit, which uses the Canny edge detection algorithm to perform edge detection on the preprocessed image; and a threshold segmentation unit, which uses the Otsu algorithm to automatically calculate the threshold and divide the image into two parts: foreground and background.

[0012] In a possible implementation, a PVC guardrail surface flatness detection method applied to the above-mentioned PVC guardrail surface flatness detection system comprises the following steps:

[0013] Step 1: Start the conveying platform and convey the PVC guardrail to the bottom of the dark box;

[0014] Step 2: Turn on the bar light source and the camera, and move them synchronously along the track to collect single images at set intervals;

[0015] Step 3, the lower computer transmits the single-collected image to the image signal acquisition module of the upper computer;

[0016] Step 4: The image signal acquisition module of the host computer receives and caches the image data;

[0017] Step 5: The image preprocessing module combines, converts grayscale, removes noise and enhances contrast of the image;

[0018] Step 6: The image feature extraction module extracts the image edge and segments the foreground and background areas;

[0019] Step 7: The abnormal area analysis module marks the foreground area and calculates the feature parameters;

[0020] Step 8: The flatness judgment module judges the flatness of the guardrail surface according to the set abnormal area threshold and quantity threshold, and displays and stores the detection results.

[0021] Beneficial effects compared with the prior art:

[0022] In this solution, multiple acquisitions are performed on different positions of the guardrail, and the single scan images acquired multiple times are stitched into an overall guardrail image for analysis by the host computer. This multiple acquisition and stitching method can cover the entire surface of the guardrail, avoiding the loss of image information due to a small single acquisition area, and ensuring that the host computer obtains complete image data of the guardrail surface, thereby analyzing the overall flatness of the guardrail to prevent potential flatness problems from being missed due to partial non-acquisition. Moreover, the method of multiple acquisition and stitching of images is highly flexible. No matter how the size of the guardrail changes, the acquisition interval and number of times can be adjusted so that the detection system can adapt to the detection needs of guardrails of various specifications. There is no need for large-scale hardware modification or system adjustment for guardrails of different specifications, which improves the versatility and practicality of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention in conjunction with the accompanying drawings.

[0024] Figure 1 It is a schematic diagram of the structure of the host computer of the present invention;

[0025] Figure 2 This is a schematic diagram of the image acquisition process of the host computer of the present invention;

[0026] Figure 3 It is a schematic diagram of image combination of the present invention;

[0027] Figure 4 This is a system framework diagram of a PVC guardrail surface flatness detection system;

[0028] Figure 5 The present invention is a flow chart of the steps of a method for detecting the surface flatness of a PVC guardrail. DETAILED DESCRIPTION

[0029] 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 forms, so the present invention is not limited to the embodiments described below.

[0030] The technical solution in the embodiment of the present application is to solve the problems of the above-mentioned background technology, and the overall idea is as follows:

[0031] Example:

[0032] Please refer to Figure 1-5 As shown, this embodiment introduces a method for detecting the surface flatness of a PVC guardrail, including a lower computer and an upper computer;

[0033] 1. The lower machine includes a conveying platform, on which a dark box is arranged, and a track is installed on the upper side of the dark box, on which a strip light source and a camera move;

[0034] (1) Conveying platform: Select a suitable belt conveyor as the conveying platform. Its speed can be adjusted according to the actual testing requirements. The function of the conveying platform is to uniformly convey the PVC guardrail to the bottom of the dark box to ensure that each PVC guardrail to be tested can pass through the testing area smoothly;

[0035] (2) Dark box: The dark box is made of opaque material. Its internal environment is relatively closed, which can effectively reduce the interference of external light and provide stable lighting conditions for image acquisition;

[0036] (3) Track: A linear track is installed on the upper side of the dark box. The track should have high straightness and stability. The strip light source and camera are installed on the track and move at a uniform speed along the track through a stepper motor or other driving device.

[0037] (4) Strip light source, which uses high-brightness, uniform LED strip light source. The irradiation angle and intensity of the light source need to be adjusted according to the actual situation to ensure a clear light and shadow effect on the surface of the PVC guardrail. When the light shines on the concave and convex parts of the guardrail surface, it will produce different degrees of reflection and shadow, thus forming a clear light and shadow difference;

[0038] (5) Camera: an industrial camera is used as the image acquisition device. The industrial camera has high resolution, high frame rate and good stability. The camera moves synchronously with the bar light source. During the movement, the light and shadow image information on the surface of the PVC guardrail is collected at a certain time interval or distance interval. The collected image is transmitted to the host computer through a data cable for processing;

[0039] Specifically, as the strip light source and the camera move on the track, the camera will collect images of the area illuminated by the light source every time it moves a certain distance. The single collection area is small, and the camera moves with the light source to collect multiple images at different positions of the guardrail. The multiple single scan images are spliced ​​into an overall guardrail image, which then provides a data basis for the upper computer analysis.

[0040] Specifically, the appropriate acquisition interval is determined based on the length and width of the guardrail, as well as the resolution and field of view of the camera. The width of the single camera image acquisition area is 20 cm, and the speed of the strip light source and the camera moving on the track is 0.1 s / cm. In order to ensure that the acquired image can completely cover the surface of the guardrail, a 5% overlap area is set (for subsequent image stitching). Then, the acquisition interval distance R is 19 cm, and the acquisition time interval T should be set to 1.9 s;

[0041] Specifically, the number of times a guardrail is captured is determined by the length of the guardrail. Assuming that the guardrail is 3 meters long, the maximum number of images n that need to be captured for the guardrail is 17;

[0042] 2. The host computer includes an image signal acquisition module, an image preprocessing module, an image feature extraction module, an abnormal area analysis module, and a flatness judgment module;

[0043] 2.1 Image signal acquisition module

[0044] 2.1.1, a data receiving unit, which receives the single-shot image data transmitted by the lower computer in real time through a communication link established with the lower computer;

[0045] 2.1.2, a cache management unit, which is used to store the image data received by the data receiving unit into a memory cache area. The cache area is managed in a circular queue manner, which can effectively adapt to different acquisition speeds and processing speeds;

[0046] 2.2 Image Preprocessing Module

[0047] 2.2.1 Image combination unit

[0048] The camera is driven by the track to move one stroke in the dark box as a cycle. All images collected in each cycle are all single-collected images of a guardrail. Then, according to the order of image collection, the single-collected images in the cache area are arranged by the number in the image file name. The image number is composed of the date of the detection day, the sequence number of the guardrail detected on that day, and the sequence of the single-collected images of the current guardrail detection. For example, the number "2020.01.25_055_001.jpg" can be interpreted as the first image of the 55th guardrail collected on January 25, 2025. Then, the image collection order can be clarified by the number in the file name;

[0049] According to the setting of 5% overlap area, adjacent images are analyzed, and the feature matching algorithm is used to find matching feature points in the overlapping area. For two adjacent images A and B, in the overlapping area, the SIFT algorithm will detect some feature points (such as corner points, edge points, etc.) in image A, and at the same time find matching feature points in the overlapping area of ​​image B. Through these matching feature points, the relative position relationship between image A and image B is calculated, such as how many pixels are translated, how much angle is rotated, etc. Then, according to the calculated relative position relationship, the adjacent images are geometrically transformed so that they are in the same coordinate system. Alignment: Use weighted averaging to fuse the pixel values ​​in the overlapping area. For example, for a certain pixel in the overlapping area, the pixel value in image A is (200, 200, 200), and the pixel value in image B is (180, 180, 180). According to weighted averaging (assuming the weights are 0.6 and 0.4), the fused pixel value is (0.6×200+0.4×180, 0.6×200+0.4×180, 0.6×200+0.4×180) = (192, 192, 192), thereby removing the duplicate parts and obtaining a complete image of the guardrail.

[0050] 2.2.2 Image Processing Unit

[0051] First, the combined color image is converted into a grayscale image to reduce the amount of data while retaining the brightness information, simplify the subsequent processing steps, and improve the processing efficiency. Then, Gaussian filtering is used to remove noise interference in the grayscale image to prevent noise from affecting the accuracy of light and shadow information. Finally, the histogram of the image is adjusted to enhance the image contrast and make the light and shadow details more obvious, which is convenient for subsequent feature extraction.

[0052] 2.3 Image feature extraction module

[0053] 2.3.1. Edge detection unit. In the PVC guardrail image, the uneven surface will cause changes in light reflection and shadow, thus forming edges. The Canny edge detection algorithm performs edge detection on the preprocessed image, which will form obvious edge lines on the uneven surface of the guardrail. These edge lines reflect the shape changes of the guardrail surface.

[0054] 2.3.2, Threshold segmentation unit: Threshold segmentation is to automatically calculate a suitable threshold value based on the gray value distribution of the image using the Otsu algorithm to divide the image into foreground and background. In the PVC guardrail image, the foreground part usually corresponds to the area where the guardrail surface may be uneven, and the background part corresponds to the relatively flat area. For example, for a PVC guardrail image, the threshold value calculated by the Otsu algorithm is 120, then the pixels in the image with gray values ​​greater than 120 are divided into the foreground (areas where there may be unevenness), and the pixels with gray values ​​less than 120 are divided into the background (flat area);

[0055] 2.4. Abnormal area analysis module

[0056] 2.4.1. Region marking unit: For a foreground region segmented from a PVC guardrail image, starting from a certain pixel point, the surrounding pixels connected to it are marked with the same region number through the seed filling algorithm until all the pixels in the region are marked;

[0057] 2.4.2. Feature calculation unit: Calculate the area, perimeter, circularity and other feature parameters of each abnormal area. For an abnormal area, the number of pixels is counted as 500, so its area is 500 square pixels. By calculating the length of the boundary pixels, the perimeter is 80 pixels. Substituting it into the circularity formula, the circularity is (4×3.14×500) / (80×80)≈0.98. The closer the circularity is to 1, the closer the area is to a circle.

[0058] 2.5. Flatness judgment module

[0059] According to the production standards of PVC guardrails and actual inspection experience, the threshold of abnormal area is set to 100 square pixels, and the threshold of abnormal area number is set to 5. If the area of ​​an abnormal area is larger than 100 square pixels, or the number of abnormal areas in the image exceeds 5, it may be considered that the surface of the guardrail is uneven.

[0060] The calculated abnormal area size and number are compared with the set threshold. If the area of ​​the abnormal area is smaller than the area threshold and the number of the abnormal area is smaller than the number threshold, the surface of the guardrail is judged to be flat; otherwise, the surface of the guardrail is judged to be uneven. The flatness detection result is displayed on the user interface in an intuitive manner, such as displaying the words "flat" or "uneven". At the same time, the detection results and related image data are stored in the database for subsequent query and analysis.

[0061] Finally, it should be noted that: Obviously, the above embodiments are only examples for clearly explaining the present invention, and are not intended to limit the implementation methods. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived from this are still within the scope of protection of the present invention.

Claims

1. A PVC guardrail surface flatness detection system, characterized in that: The system includes a lower computer and an upper computer; the lower computer includes a conveying platform, a dark box, a track, a strip light source, and a camera, and the upper computer includes an image signal acquisition module, an image preprocessing module, an image feature extraction module, an abnormal area analysis module, and a flatness judgment module; The conveying platform is used to convey the PVC guardrail at a uniform speed to the bottom of the dark box. The dark box is made of opaque material to seal the internal environment to reduce external light interference. The track is installed on the upper side of the dark box for installing a strip light source and a camera. The strip light source uses an LED strip light source with adjustable illumination angle and intensity. The camera is used to collect light and shadow images on the surface of the PVC guardrail and transmit them to the host computer.

2. A PVC guardrail surface flatness detection system as claimed in claim 1, characterized in that: The image signal acquisition module includes: a data receiving unit, which receives single-time acquired image data in real time through a communication link with a lower computer; and a cache management unit, which adopts a circular queue method to manage and store the received image data in a memory cache area.

3. A PVC guardrail surface flatness detection system as claimed in claim 1, characterized in that: The image preprocessing module includes: an image combination unit, which arranges the single-shot acquisition images in the buffer area according to the order of picture acquisition, and uses a feature matching algorithm to perform geometric transformation and pixel fusion on adjacent images according to the 5% overlap area setting to obtain a complete image; an image processing unit, which converts the combined color image into a grayscale image, uses Gaussian filtering to remove noise, and adjusts the histogram to enhance the contrast.

4. A PVC guardrail surface flatness detection system as claimed in claim 1, characterized in that: The image feature extraction module includes: an edge detection unit, which uses the Canny edge detection algorithm to perform edge detection on the preprocessed image; and a threshold segmentation unit, which uses the Otsu algorithm to automatically calculate the threshold and divide the image into two parts: foreground and background.

5. A PVC guardrail surface flatness detection method applied to a PVC guardrail surface flatness detection system according to any one of claims 1 to 4, characterized in that: The method comprises the following steps: Step 1: Start the conveying platform and convey the PVC guardrail to the bottom of the dark box; Step 2: Turn on the bar light source and the camera, and move them synchronously along the track to collect single images at set intervals; Step 3, the lower computer transmits the single-collected image to the image signal acquisition module of the upper computer; Step 4: The image signal acquisition module of the host computer receives and caches the image data; Step 5: The image preprocessing module combines, converts grayscale, removes noise and enhances contrast of the image; Step 6: The image feature extraction module extracts the image edge and segments the foreground and background areas; Step 7: The abnormal area analysis module marks the foreground area and calculates the feature parameters; Step 8: The flatness judgment module judges the flatness of the guardrail surface according to the set abnormal area threshold and quantity threshold, and displays and stores the detection results.

Images