Defect positioning method and system based on laser and vision online detection

Abstract

The present application relates to the technical field of optical testing defects, in particular to a defect positioning method and system based on laser and visual online detection. The method comprises the following steps: monitoring the concave-convex defect problem of the enameled wire surface by laser, and monitoring the color defect problem of the enameled wire surface by vision; meanwhile, the image data of the concave-convex defect position is called from the real-time image data, the image data is identified by using the shadow depth analysis technology, the matching difference value of the laser concave-convex data and the visual concave-convex data is calculated, when the matching difference value is greater than the difference threshold value, the error signal is output, and the concave-convex data of the concave-convex defect position is verified again by using the image data fed back by vision. The laser monitoring and the visual monitoring are mutually verified, the accuracy of defect detection is improved, and when any one of the laser monitoring or the visual monitoring outputs abnormal data, the worker can be timely warned, the equipment damage is avoided, and the monitoring effect of the enameled wire performance is affected.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of optical testing defects, in particular, to a defect positioning method and system based on laser and visual online detection. BACKGROUND

[0002] The enameled wire will cause occasional quality problems during production due to the mold, paint, and foreign matter, such as air bubble particles, surface impurities, etc. These defects will cause the electrical insulation performance of the enameled wire to decrease significantly, far below the product design requirements. Therefore, in order to ensure 100% safety and reliability of the product, online monitoring technology must be used for full inspection of the product.

[0003] In recent years, with the development of online monitoring technology, laser online monitoring technology and visual online monitoring technology have been widely applied. For example, Chinese patent application No. CN202410984093.X discloses a cable detection system combining laser and vision, which includes a processing module, a laser generating device, a cable installation rotating mechanism, a laser receiving device, and a visual detection device. The cable installation rotating mechanism is used to place and drive the cable to rotate. The laser generating device is located obliquely above the cable installation rotating mechanism. The visual detection device is located directly above the cable installation rotating mechanism. The laser receiving device is located obliquely above the cable installation rotating mechanism. The purpose of the present application is to provide a cable detection system combining laser and vision, which screens out suspected defect areas by laser, and then analyzes and judges the suspected defect areas by vision, thereby greatly improving the detection efficiency of cable defects.

[0004] However, these technologies still have some shortcomings. The laser monitoring system has high measurement accuracy, but is greatly affected by wire fluctuations, is prone to false positives, and cannot distinguish the size of particles, and it is also difficult to verify the accuracy of particle positioning. The visual monitoring system is good at identifying color difference defects, but has many false positives for oil stains, cannot find defects with high height, and is prone to miss small particles. In view of this, we propose a system that includes laser + 2D camera + insulation detection, detects defects through the above-mentioned "three-in-one" device, marks the defect position using a small character inkjet printer when a defect is detected, and then adds a visual system to judge whether the inkjet printer marks normally, forming a detection closed-loop system. SUMMARY

[0005] The present application relates to the technical field of optical testing defects, in particular, to a defect positioning method and system based on laser and visual online detection.

[0006] To solve the above technical problems, one of the purposes of the present application is to provide a defect positioning method based on laser and visual online detection, which includes the following steps:

[0007] S1, continuously scanning the same position on the surface of the enameled wire by the laser and the visual monitoring device, the visual monitoring device capturing the reflection points of the laser on the surface of the enameled wire to feed back the concave-convex defects on the surface of the enameled wire, and recording the laser concave-convex data, the laser concave-convex data including the concave-convex defect height information and the concave-convex defect position, when the concave-convex defect height information does not match the height threshold range, outputting the concave-convex defect signal and giving a warning;

[0008] S2, shooting the real-time image data of the surface of the enameled wire by the visual monitoring device, and identifying the color data of the surface of the enameled wire by using the image analysis technology, when the color data does not match the color threshold range, outputting the color defect signal and giving a warning, the color defect including the color difference defect and the color spot defect;

[0009] S3, after receiving the concave-convex defect signal, calling out the image data of the concave-convex defect position from the real-time image data, identifying the image data by using the shadow depth analysis technology, feeding back the visual concave-convex data in the image data, and calculating the matching difference value of the laser concave-convex data and the visual concave-convex data, when the matching difference value is greater than the difference threshold, outputting the error signal.

[0010] Preferably, the warning in S1 is electrically connected with the first indicator light, for driving the first indicator light to emit light and give a warning when the concave-convex defect signal is perceived, and the warning in S2 is electrically connected with the second indicator light, for driving the second indicator light to emit light and give a warning when the color defect signal is perceived.

[0011] Preferably, the step of capturing the reflection points of the laser on the surface of the enameled wire to feed back the concave-convex defects on the surface of the enameled wire in S1 includes the following steps:

[0012] the laser emits a beam of laser to the surface of the enameled wire, the visual monitoring device captures the reflection points of the laser on the surface of the enameled wire, and obtains the distance from the laser to the visual monitoring device , the incident angle of the laser , the included angle between the reflection points of the laser captured by the visual monitoring device and the baseline , the displacement of the laser reflection points on the image of the visual monitoring device , then the height of the surface of the enameled wire:

[0013] ;

[0014] the laser and the visual monitoring device continuously scan the same position on the enameled wire, average a plurality of height data , if , output the normal signal, if , output the concave-convex defect signal.

[0015] Preferably, the position information of the concave-convex defects is determined by recording the position of the laser reflection point in S1, and the ink-jet printer is driven to mark the position of the concave-convex defects when the output concave-convex defect signal is received.

[0016] Preferably, the color data of the enameled wire surface is identified by using image analysis technology in S2, including:

[0017] When identifying the color difference defects, the real-time image data is converted from the RGB color space to the HSV color space;

[0018] The standard color of the enameled wire is determined, the color difference of each pixel is calculated by using the Euclidean distance, and the color difference is compared with the color difference threshold by using the numerical comparison algorithm, if the color difference ≤ color difference threshold, the normal signal is output, if the color difference > color difference threshold, the color difference defect is output;

[0019] When identifying the color spot defects, the different regions in the real-time image data are segmented by using the edge detection algorithm, the color features of each region are extracted, and the color variance threshold is set, if the color variance of a certain region exceeds the threshold, the color spot defect is determined.

[0020] Preferably, the image data is identified by using the shadow depth analysis technology in S3, and the visual concave-convex data in the image data is fed back, including the following steps:

[0021] The edges in the image data are detected by using the edge detection algorithm, and the gradient information in the image is calculated, which is used for detecting the shadow edges, and a plurality of pixel points are output;

[0022] The surface depth h2 of each pixel point is estimated by using the illumination model, and the expression is:

[0023]

[0024] Wherein, is the brightness of a certain point in the image, is the ambient light intensity, is the diffuse reflection intensity, is the included angle between the light source direction and the surface normal vector;

[0025] The surface depth h2 is quantified as a specific numerical output.

[0026] Preferably, the matching difference value of the laser concave-convex data and the visual concave-convex data is calculated in S3, including the following steps:

[0027] The average height data of the laser monitoring is accepted and the surface depth h2 of the visual monitoring is normalized to calculate the difference value of h1 and h2;

[0028] Set the difference threshold, if the difference > difference threshold, output error signal, if the difference = difference threshold, output normal signal, if the difference < difference threshold, also output normal signal, and the mean method is calculated And h2 average value as the S1 in the concave and convex defect height information.

[0029] Preferably, the S3 also includes a mark verification algorithm, the mark verification algorithm is used for collecting inkjet printer mark features, and the image data features of the concave and convex defects are extracted by using image analysis technology, the mark features are matched and compared with the image data features, if matching, the inkjet printer normal signal is output, if not matching, the inkjet printer abnormal signal is output.

[0030] The second purpose of the application is to provide a defect positioning system based on laser and visual online detection, comprising the defect positioning method based on laser and visual online detection in any one of the above, comprising a laser monitoring module, a visual monitoring module and a data verification module.

[0031] The laser monitoring module is used for monitoring the surface of the enameled wire through laser, capturing the reflection point feedback of the laser on the surface of the enameled wire to capture the concave and convex defects of the surface of the enameled wire, and recording the laser concave and convex data, when the concave and convex defect height information does not match the height threshold range, outputting the concave and convex defect signal and warning.

[0032] The visual monitoring module is used for capturing real-time image data of the surface of the enameled wire through a visual monitoring device, and identifying color data of the surface of the enameled wire by using image analysis technology, when the color data does not match the color threshold range, outputting a color defect signal and warning, and the color defect includes color difference defect and color spot defect.

[0033] The data verification module is used for receiving the concave and convex defect signal, calling out the image data of the concave and convex defect position from the real-time image data, identifying the image data by using shadow depth analysis technology, feeding back the visual concave and convex data in the image data, and calculating the matching difference value of the laser concave and convex data and the visual concave and convex data, when the matching difference value is greater than the difference threshold, outputting an error signal.

[0034] Compared with the prior art, the application has the following beneficial effects:

[0035] The present application not only monitors the concave-convex defect problem of the enameled wire surface through laser, but also monitors the color defect problem of the enameled wire surface through vision, at the same time, the image data of the concave-convex defect position is called from the real-time image data, the image data is identified by using the shadow depth analysis technology, the visual concave-convex data in the image data is fed back, the matching difference value of the laser concave-convex data and the visual concave-convex data is calculated, when the matching difference value is greater than the difference threshold value, the error signal is output, and the concave-convex data of the concave-convex defect position is verified again through the image data fed back by vision, which is beneficial to the mutual verification of laser monitoring and vision monitoring, improves the accuracy of defect detection, and when any one of laser monitoring or vision monitoring outputs abnormal data, the worker can be timely warned to avoid the influence of the monitoring effect of the enameled wire performance when the equipment is damaged. BRIEF DESCRIPTION OF DRAWINGS

[0036] Fig. 1 It is the overall flow chart of example 1;

[0037] Fig. 2 It is the principle diagram of marking verification of example 1. DETAILED DESCRIPTION

[0038] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0039] Example 1, as shown in Figs. 1-2 One of the purposes of the present application is to provide a defect positioning method based on laser and vision online detection, which includes the following steps:

[0040] S1, the same position of the enameled wire surface is continuously scanned by a laser and a vision monitoring device, the vision monitoring device captures the reflection point of the laser on the enameled wire surface to feed back the concave-convex defect of the enameled wire surface, and records the laser concave-convex data, the laser concave-convex data includes the concave-convex defect height information and the concave-convex defect position, when the concave-convex defect height information does not match the height threshold range, the concave-convex defect signal is output and a warning is given, which is beneficial to detecting the concave-convex defect of the enameled wire surface through laser monitoring means, and when the concave-convex defect signal appears, the worker can be reminded to find the concave-convex defect problem in time;

[0041] S2, the real-time image data of the enameled wire surface is photographed by the visual monitoring device, color data of the enameled wire surface is recognized by using image analysis technology, when the color data does not match the color threshold range, a color defect signal is output and a warning is given, the color defect includes color difference defect and color spot defect, which is beneficial to feed back the color data of the enameled wire surface through vision, and facilitate timely discovery of color defects, so that the staff can understand the color defect problem;

[0042] S3, after receiving the concave-convex defect signal, the image data of the concave-convex defect position is called out from the real-time image data, the image data is recognized by using shadow depth analysis technology, the visual concave-convex data in the image data is fed back, and the matching difference value of the laser concave-convex data and the visual concave-convex data is calculated, when the matching difference value is greater than the difference threshold value, an error signal is output, the concave-convex data of the concave-convex defect position is verified again through the image data fed back by vision, which is beneficial to mutual verification of laser monitoring and visual monitoring, improves the accuracy of defect detection, and when any one of laser monitoring or visual monitoring outputs abnormal data, the staff can be timely alerted, avoiding equipment damage and affecting the monitoring effect of enameled wire performance.

[0043] On the basis of the above, the detailed working scheme is disclosed:

[0044] Among them, in order to enable the staff to distinguish between concave-convex defects and color defects when receiving the warning, the warning in S1 is electrically connected with the first indicator light, which is used to drive the first indicator light to emit light and warn when the concave-convex defect signal is sensed, the warning in S2 is electrically connected with the second indicator light, which is used to drive the second indicator light to emit light and warn when the color defect signal is sensed, so that the staff can judge whether there is a concave-convex defect according to the light emission of the first indicator light, and whether there is a color defect according to the light emission of the second indicator light, which facilitates timely determination of the type of defect.

[0045] Further, the S1 captures the reflection point of laser on the enameled wire surface to feed back the concave-convex defect of the enameled wire surface, which includes the following steps:

[0046] The laser emits a beam of laser to the enameled wire surface, the visual monitoring device captures the reflection point of laser on the enameled wire surface, and obtains the distance from the laser to the visual monitoring device , the incident angle of the laser , the included angle between the laser reflection point captured by the visual monitoring device and the baseline , the displacement of the laser reflection point on the image of the visual monitoring device , then the height of the enameled wire surface:

[0047] ;

[0048] The laser and the visual monitoring device continuously scan the same position of the enameled wire, average multiple height data If , output normal signal, if , output convex defect signal, by calculating the average value of multiple height data at the same position, it is beneficial to improve the accuracy of height data, and further ensure the accuracy of convex defect determination.

[0049] Secondly, the position information of the convex defect is determined by recording the position of the laser reflection point in S1, when the output convex defect signal is received, the inkjet printer marks the position of the convex defect, converts the convex defect signal into a control signal that can be recognized by the inkjet printer, and converts the position information into the coordinate system of the inkjet printer, the control signal includes the moving instruction and the inkjet instruction of the inkjet printer, the generated control signal is sent to the control system of the inkjet printer, which is realized through serial communication (such as RS-232, USB) or network communication (such as TCP / IP), the inkjet printer accurately controls the moving position through the equipped stepping motor or servo motor, realizes moving to the specified coordinate after receiving the control signal, and executes the inkjet operation, the inkjet printer usually uses inkjet or laser engraving technology to mark on the enameled wire.

[0050] Moreover, the color data of the enameled wire surface is recognized by using image analysis technology in S2, including:

[0051] When identifying color difference defects: convert the real-time image data from RGB color space to HSV color space to better separate color and brightness information;

[0052] Determine the standard color of the enameled wire, calculate the color difference between each pixel and the standard color by using the Euclidean distance, and compare the color difference with the color difference threshold by using the numerical comparison algorithm, if the color difference ≤ color difference threshold, output normal signal, if the color difference > color difference threshold, output color difference defect, which is beneficial for enameled wire surface color difference defect detection;

[0053] When identifying color spot defects: use edge detection algorithm (such as Canny edge detection) to segment different regions in real-time image data, extract color features of each region such as average color, color variance, etc., and set a color variance threshold, if the color variance of a region exceeds the threshold, it is judged as color spot defect, which is beneficial for enameled wire surface color spot defect detection.

[0054] Further, the image data is recognized by using shadow depth analysis technology in S3, and the visual convex data in the image data is fed back, including the following steps:

[0055] Detect the edges in the image data by using edge detection algorithm, and calculate the gradient information in the image for detecting shadow edges, output multiple pixel points;

[0056] Estimate the surface depth h2 of each pixel point by using the illumination model, the expression is:

[0057]

[0058] wherein, is the brightness of a certain point in the image, is the ambient light intensity, is the diffuse reflection intensity, is the included angle between the light source direction and the surface normal vector, the commonly used illumination model is Lambertian reflection model, wherein, is the surface normal vector, is the light source direction, the surface normal vector is deduced through the illumination model, and then the depth is estimated;

[0059] By quantifying the surface depth h2 into a specific numerical output, it is beneficial to monitor the concave-convex depth of the enameled wire surface through vision, accurately detect and quantify the concave-convex condition of the surface, and provide strong support for the enameled wire surface quality detection.

[0060] It is worth noting that the calculation of the laser concave-convex data and the visual concave-convex data matching difference value in S3 includes the following steps:

[0061] accepting the average height data of laser monitoring and the surface depth h2 of visual monitoring, normalizing calculation and the difference value of h2;

[0062] Setting a difference threshold, if the difference value > the difference threshold, an error signal is output, indicating that the error of the two detection methods is large, which may be caused by the deviation of the damage of a certain device, therefore, it is beneficial for the staff to find the running status of the device in time, if the difference value = the difference threshold, a normal signal is output, and if the difference value < the difference threshold, a normal signal is also output, and the average value of h1 and h2 is calculated as the concave-convex defect height information in S1, which is beneficial to improve the accuracy of the concave-convex defect height information and further improve the detection of the concave-convex defect.

[0063] In order to ensure that the inkjet printer can accurately mark at the concave-convex defect position, the S3 further comprises a marking verification algorithm, the marking verification algorithm is used for collecting marking features of the inkjet printer, and image analysis technology is used to extract image data features of the concave-convex defect, the marking features are matched and compared with the image data features, if the matching is correct, a normal signal of the inkjet printer is output, and if the matching is incorrect, an abnormal signal of the inkjet printer is output.

[0064] The second purpose of the present application is to provide a defect positioning system based on laser and vision online detection, which comprises the defect positioning method based on laser and vision online detection in any one of the above, and comprises a laser monitoring module, a visual monitoring module and a data verification module.​

[0065] The laser monitoring module is used for monitoring the surface of the enameled wire by laser, capturing the reflection points of the laser on the surface of the enameled wire to feedback the concave-convex defects of the surface of the enameled wire, and recording the laser concave-convex data, and outputting a concave-convex defect signal and giving a warning when the height information of the concave-convex defects does not match the height threshold range.

[0066] The visual monitoring module is used for taking real-time image data of the surface of the enameled wire by a visual monitoring device, and identifying color data of the surface of the enameled wire by using an image analysis technology, and outputting a color defect signal and giving a warning when the color data does not match the color threshold range, and the color defects include color difference defects and color spot defects.

[0067] The data verification module is used for receiving the concave-convex defect signal, calling out image data of the position of the concave-convex defects from the real-time image data, identifying the image data by using a shadow depth analysis technology, feeding back visual concave-convex data in the image data, and calculating a matching difference value of the laser concave-convex data and the visual concave-convex data, and outputting an error signal when the matching difference value is greater than a difference threshold.

[0068] The basic principles, main features and advantages of the present application are shown and described above. It should be understood by those skilled in the art that the present application is not limited by the above-mentioned embodiments, and the above-mentioned embodiments and descriptions in the specification are only preferred examples of the present application and are not intended to limit the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A defect localization method based on laser and visual online detection, characterized in that, Includes the following steps: S1. The same position on the surface of the enameled wire is continuously scanned by a laser and a visual monitoring device. The visual monitoring device captures the reflection point of the laser on the surface of the enameled wire to report the unevenness of the surface of the enameled wire and records the laser unevenness data. The laser unevenness data includes the height information of the unevenness and the position of the unevenness. Multiple height data are collected to calculate the average value of the height information of the unevenness. When the height information of the unevenness does not match the height threshold range, the unevenness signal is output and an early warning is issued. S2. Real-time image data of the surface of the enameled wire is captured by a visual monitoring device, and the color data of the surface of the enameled wire is identified by image analysis technology. When the color data does not match the color threshold range, a color defect signal is output and an early warning is issued. Color defects include color difference defects and color spot defects. S3. After receiving the concave-convex defect signal, retrieve the image data of the concave-convex defect location from the real-time image data, use shadow depth analysis technology to identify the image data, output the visual concave-convex data in the image data to visually monitor the surface depth, and calculate the matching difference value between the laser concave-convex data and the visual concave-convex data. When the matching difference value is greater than the difference threshold, output an error signal, which helps the staff to discover the equipment operation status in time. If the difference value equals the difference threshold, output a normal signal. If the difference value is less than the difference threshold, also output a normal signal. Use the mean method to calculate the average value of the concave-convex defect height information and the average value of the visually monitored surface depth as the concave-convex defect height information in S1. The S3 step employs shadow depth analysis technology to identify image data and provides feedback on visual convexity / concave data within the image data. This includes the following steps: detecting edges in the image data using an edge detection algorithm and calculating gradient information in the image for detecting shadow edges, outputting multiple pixels; estimating the surface depth h2 of each pixel using a lighting model, expressed as: in, It refers to the brightness of a point in the image. It is ambient light intensity. It is the intensity of diffuse reflection. It is the angle between the light source direction and the surface normal vector; by quantizing the surface depth h2 into a specific numerical output; The calculation of the difference between laser concavity / convexity data and visual concavity / convexity data in S3 includes the following steps: receiving average height data from laser monitoring. The surface depth h2 of visual monitoring is normalized and calculated. The difference between h1 and h2 is calculated. A difference threshold is set. If the difference is greater than the difference threshold, an error signal is output. If the difference equals the difference threshold, a normal signal is output. If the difference is less than the difference threshold, a normal signal is also output. The mean value method is used to calculate the error. The average value of h2 is used as the height information of the concave and convex defects in S1.

2. The defect localization method based on laser and visual online detection according to claim 1, characterized in that: The warning signal in S1 is electrically connected to the first indicator light and is used to drive the first indicator light to illuminate as a warning signal when a concave or convex defect signal is detected. The warning signal in S2 is electrically connected to the second indicator light and is used to drive the second indicator light to illuminate as a warning signal when a color defect signal is detected.

3. The defect localization method based on laser and visual online detection according to claim 2, characterized in that: The step S1, which involves capturing the reflection points of the laser on the surface of the enameled wire to provide feedback on surface irregularities, includes the following steps: A laser emits a laser beam onto the surface of an enameled wire, and a visual monitoring device captures the reflection point of the laser on the surface of the enameled wire. Laser and visual monitoring equipment continuously scan the same location of the enameled wire, collecting multiple height data and averaging them. ,like If, then a normal signal is output; if Then the output signal for unevenness / convexity defects will be generated.

4. The defect localization method based on laser and visual online detection according to claim 3, characterized in that: In step S1, the location information of the uneven defect is determined by recording the position of the laser reflection point. When the uneven defect signal is received, the inkjet printer is driven to mark the position of the uneven defect.

5. The defect localization method based on laser and vision online detection according to claim 4, characterized in that: The S2 step uses image analysis technology to identify the color data of the enameled wire surface, including: When identifying color difference defects: convert real-time image data from RGB color space to HSV color space; The standard color of the enameled wire is determined. The color difference between each pixel and the standard color is calculated using Euclidean distance. A numerical comparison algorithm is used to compare the color difference with the color difference threshold. If the color difference is less than or equal to the color difference threshold, a normal signal is output. If the color difference is greater than the color difference threshold, a color difference defect is output. When identifying color spot defects: use an edge detection algorithm to segment different regions in real-time image data, extract the color features of each region, and set a color variance threshold. If the color variance of a certain region exceeds the threshold, it is judged as a color spot defect.

6. The defect localization method based on laser and vision online detection according to claim 5, characterized in that: The S3 also includes a mark verification algorithm, which is used to collect the mark features of the inkjet printer and use image analysis technology to extract the image data features of the concave and convex defects. The mark features are matched and compared with the image data features. If they match, the inkjet printer outputs a normal signal; if they do not match, the inkjet printer outputs an abnormal signal.

7. A defect localization system for implementing online laser and visual detection, comprising the defect localization method for online laser and visual detection as described in any one of claims 1-6, characterized in that: It includes a laser monitoring module, a visual monitoring module, and a data verification module; The laser monitoring module is used to monitor the surface of the enameled wire by laser, capture the reflection points of the laser on the surface of the enameled wire to provide feedback on the unevenness of the surface of the enameled wire, and record the laser unevenness data. When the height information of the unevenness defect does not match the height threshold range, it outputs the unevenness defect signal and issues an early warning. The visual monitoring module is used to capture real-time image data of the surface of the enameled wire through the visual monitoring device, and to identify the color data of the surface of the enameled wire using image analysis technology. When the color data does not match the color threshold range, it outputs a color defect signal and issues an early warning. Color defects include color difference defects and color spot defects. The data verification module is used to receive the concave and convex defect signal, retrieve the image data of the concave and convex defect location from the real-time image data, identify the image data using shadow depth analysis technology, feed back the visual concave and convex data in the image data, and calculate the matching difference value between the laser concave and convex data and the visual concave and convex data. When the matching difference value is greater than the difference threshold, an error signal is output.

Images