Large component appearance size checking system based on machine vision

Abstract

The utility model discloses a large-scale component appearance dimension checking system based on machine vision, which comprises a mounting platform, a fixing frame, a plurality of industrial cameras, a central control box, a server and a conveying device, the fixing frame is arranged vertical to the ground and is connected with the mounting platform, the industrial cameras are uniformly arranged on the mounting platform, and the central control box is connected with the server. The central control box, the server and the conveying device are all arranged below the installation platform, the industrial cameras are connected with the server through the central control box, and the conveying device is arranged on the ground and used for conveying a measured component to a measurement center area below the installation platform. The device is simple in structure, is combined with the server and the central control box, is suitable for the external dimension detection of a large-scale component, is provided with the conveying device below the installation platform, improves the automation degree of the detection process, is also provided with the edge calculation equipment, improves the data processing capability of the server, and improves the detection speed.

Description

Technical Field

[0001] This utility model relates to the field of appearance inspection technology, and more specifically, to a machine vision-based system for checking the appearance dimensions of large components. Background Technology

[0002] Large components are prefabricated parts that are large in size and heavy in the fields of construction, bridges, and infrastructure. The manufacturing precision of large components in prefabricated buildings needs to be very high. With the rapid development of the industry, due to the shortage of skilled workers and the imperfect supporting industrial chain, it is difficult to guarantee the quality control of large components, especially the accuracy of the appearance dimensions of large components.

[0003] Currently, the appearance and dimensional inspection of large components typically uses laser scanning combined with a 3D camera as the actuator. This method requires the configuration of a moving mechanism, such as a multi-axis robotic arm, to assist the moving inspection actuator in the inspection of the components. This mode involves many inspection actions and takes a long time, resulting in a slow overall inspection speed. Furthermore, laser scanning and multi-axis robotic arm equipment are expensive.

[0004] The patent number is 202122084491.0, and the name is a utility model patent for an appearance inspection system. It is used to detect whether there are appearance defects such as damage, scratches or dirt on the front, side or back of the product to be inspected by setting up a loading and conveying component, a handling component, a front shooting component, a back shooting component, a left shooting component, a right shooting component, a first support component and a second support component. The inspection system has a complex structure and is concentrated in one part, resulting in high equipment cost. It cannot be used to inspect the appearance dimensions of large components.

[0005] The invention patent with patent number 201910294405.3, entitled "A System and Method for Detecting the External Dimensions of Precast Components", involves setting up a mold platform with at least four image acquisition devices arranged at the four corners of a rectangle above the platform. A sensing device connected to each image acquisition device is set on the side of the mold platform below the image acquisition devices. When a precast component is detected, the image acquisition devices are triggered to acquire images of the precast component. This method gathers multiple structures together, but it is not suitable for detecting the external dimensions of large components. Utility Model Content

[0006] To address the problems of existing component appearance dimension inspection equipment being complex in structure, high in cost, and unsuitable for the size inspection of large components, this utility model provides a machine vision-based large component appearance dimension inspection system. By setting up a large mounting platform, multiple industrial cameras are installed on the platform, and combined with a server and central control box, it is suitable for the appearance dimension inspection of large components. At the same time, the overall cost of the device is low, which is conducive to its widespread adoption. A conveyor device is provided below the mounting platform to improve the automation of the inspection process. Furthermore, the mounting platform is a multi-axis micro-motion platform, and also includes edge computing devices and multiple computing nodes to improve the server's data processing capabilities and thus increase the inspection speed.

[0007] To achieve the above objectives, according to a first aspect of this utility model, a machine vision-based system for inspecting the appearance and dimensions of large components is provided, comprising an installation platform, a mounting frame, industrial cameras, a central control box, a server, and a conveying device. The mounting frame is perpendicular to the ground and connected to the installation platform. Multiple industrial cameras are evenly distributed on the installation platform. The central control box, server, and conveying device are all located below the installation platform. The industrial cameras are connected to the server through the central control box. The conveying device is located on the ground and is used to transport the component to be tested to a measurement center area below the installation platform.

[0008] Preferably, the mounting platform is a rectangular frame with a hollow center, and three to eight industrial cameras are installed on each edge of the frame.

[0009] Preferably, the industrial camera model is MER2-2000-6GC.

[0010] Preferably, the mounting platform is also equipped with tilt sensors combined with cameras or digital level integrated cameras at the corners of its four edges.

[0011] Preferably, the length of the installation platform ranges from 12m to 16m, and the width ranges from 8m to 12m.

[0012] Preferably, there are six fixing frames symmetrically arranged on both sides of the mounting platform.

[0013] Preferably, the central control box is equipped with a power supply and a wireless communication network, wherein the wireless communication network is 5G or Wi-Fi 6.

[0014] Preferably, the conveying device is a retractable conveyor belt or a guide rail system.

[0015] Preferably, the mounting platform is a multi-axis micro-motion platform.

[0016] Preferably, an edge computing device or multiple computing nodes are provided between the central control box and the server.

[0017] In summary, compared with the prior art, the above-described technical solution conceived by this utility model can achieve the following beneficial effects:

[0018] (1) The machine vision-based large component appearance size inspection system of this utility model is suitable for the appearance size inspection of large components by setting up a large installation platform, installing multiple industrial cameras on the installation platform, and combining a server and a central control box. At the same time, the overall cost of the device is low, which is conducive to its widespread use.

[0019] (2) The machine vision-based large component appearance size inspection system of this utility model improves the automation of the inspection process by providing a conveying device below the installation platform. At the same time, the installation platform is a multi-axis micro-motion platform, and is also equipped with edge computing devices and multiple computing nodes to improve the data processing capability of the server and thus improve the inspection speed. Attached Figure Description

[0020] Figure 1 This utility model presents a three-dimensional machine vision-based system for inspecting the appearance and dimensions of large components. Figure 1 .

[0021] Figure 2 This utility model presents a three-dimensional machine vision-based system for inspecting the appearance and dimensions of large components. Figure 2 .

[0022] Figure 3 This is a side view of the machine vision-based large component appearance dimension inspection system of this utility model.

[0023] Figure 4 This is a top view of the machine vision-based large component appearance and size inspection system of this utility model.

[0024] Figure 5 This is a schematic diagram of a machine vision-based system for inspecting the appearance and dimensions of large components.

[0025] In the diagram: 1. Mounting platform; 2. Mounting frame; 3. Industrial camera; 4. Central control box; 5. Server; 6. Component under test; 7. Conveying device. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this utility model clearer, the present utility model will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining the present utility model and are not intended to limit the present utility model. Furthermore, the technical features involved in the various embodiments of the present utility model described below can be combined with each other as long as they do not conflict with each other.

[0027] Depend on Figures 1 to 4 As shown, the machine vision-based large component appearance dimension inspection system includes a mounting platform 1, a fixing frame 2, industrial cameras 3, a central control box 4, a server 5, and a conveying device 7. The mounting platform 1 is constructed of steel profiles. The fixing frame 2 consists of six symmetrically arranged perpendicular to the ground and is fixedly connected to both sides of the mounting platform 1. The mounting platform 1 is a rectangular frame with a hollow center. Three to eight industrial cameras 3 are installed on each edge of the frame. Specifically, eight industrial cameras 3 are evenly distributed on the longer side of the frame, and three industrial cameras 3 are evenly distributed on the wider side. The length of the mounting platform 1 ranges from 12m to 16m, and the width ranges from 8m to 12m, which is convenient for inspecting large components. It can measure component dimensions from 4.0×2.0×3.0 meters to 8.0×3.0×4.0 meters. The mounting platform 1 also has multiple welding frames for mounting brackets, which are used to connect with the upper steel beams to ensure the stability of the device.

[0028] The industrial camera 3 is model MER2-2000-6GC. The mounting platform 1 is also equipped with tilt sensors combined with cameras or digital level integrated cameras at the four edges and corners to detect image data of the component under test 6 in various directions.

[0029] The central control box 4, server 5, and conveyor 7 are all located below the installation platform 1. The industrial camera 3 is connected to the server 5 through the central control box 4. The central control box 4 is equipped with a power supply and a wireless communication network, which is either 5G or Wi-Fi 6, to realize real-time transmission and remote control of detection data.

[0030] An edge computing device or multiple computing nodes are installed between the central control box 4 and the server 5. The edge computing device includes embedded computing devices, industrial computers, and smart gateways. It performs preliminary processing on the image data collected by the industrial camera 3, such as noise reduction, compression, and feature extraction, to reduce the amount of data transmitted to the central control box 4 and the server 5. At the same time, it performs real-time analysis on the edge computing device and automatically adjusts the detection parameters based on the detection results, such as changing the camera's exposure time and focal length, to improve the accuracy and efficiency of detection. It can also cache some data on the edge computing device to reduce dependence on the central server and improve the system's fault tolerance and reliability.

[0031] The computing nodes include physical servers, workstations, and embedded devices. Image processing tasks can be distributed to multiple computing nodes, each responsible for processing a portion of the image. Finally, the results are aggregated in the central control box 4 or server 5. At the same time, the parallel computing capabilities of multiple computing nodes are utilized to accelerate the execution of complex algorithms, such as the training and inference of deep learning models, thereby improving the accuracy of detection data.

[0032] The conveying device 7 is located on the ground and is used to transport the component 6 to be measured to the measurement center area below the installation platform 1. The conveying device 7 is a telescopic conveyor belt or a guide rail system. The telescopic conveyor belt can be adjusted in length according to the actual length of the installation platform 1 and the size of the component 6 to be measured, and is suitable for components 6 of different sizes. The guide rail system includes a fixed rail and a trolley that can move on the rail. The trolley carries the component 6 to be measured along a predetermined path to the measurement area, ensuring the stability of the transport of the component 6 to be measured, while reducing the damage to the component 6 to be measured due to transport problems, thereby affecting the accuracy of the measurement data and improving the automation level of the entire inspection system.

[0033] Mounting platform 1 can be a multi-axis micro-motion platform, enabling high-precision scanning and inspection of component surfaces. Precise movement along the X, Y, and Z axes ensures the inspection probe can cover the entire inspection area.

[0034] Depend on Figure 5 As shown, the working process of the machine vision-based large component appearance dimension inspection system is as follows:

[0035] Step 1, Identification and Calibration: Calibrate the industrial camera 3 using a standard reference or an object of known size, and calibrate the digital level integrated camera or the tilt sensor combined with the camera.

[0036] Step 2, Data Acquisition: The component 6 to be measured is transported to the measurement center area below the installation platform 1 via the conveying device 7. The server 5 controls the industrial camera 3 and the tilt sensor through the central control box 4 to collect image data in combination with the camera and transmit it back to the server 5 via the wireless communication network.

[0037] Step 4, Data Processing: The image processing software in server 5 processes the image data. At the same time, the edge computing device or multiple computing nodes cooperate with the image processing software to process the data, including data preprocessing, data alignment, feature extraction and size measurement. Feature extraction includes edge detection and feature point detection. Algorithms such as Canny and Sobel are used to detect edges in the image and extract component contours. Algorithms such as Harris corner detection and SIFT are used to extract feature points in the image.

[0038] Dimensional measurement includes two-dimensional dimensional measurement and three-dimensional dimensional measurement. Two-dimensional dimensional measurement: based on extracted edges and feature points, it measures the two-dimensional dimensions of a component, such as length, width, and height.

[0039] 3D dimensional measurement: Combining 3D scanning data, the three-dimensional dimensions of the component are measured, such as volume and surface area.

[0040] Step 3: Data verification: This includes data verification and result comparison. The measurement results are compared with the design drawings or standard dimensions to verify whether the dimensions meet the requirements. Statistical methods or machine learning algorithms are used to detect and remove outliers.

[0041] The current detection results are compared with historical data to detect any abnormal changes. By combining multimodal data such as images, 3D scans, and ultrasound, a comprehensive analysis is performed to improve the accuracy and reliability of the detection.

[0042] Finally, the verification results are displayed on a monitor connected to server 5, and staff judge whether the dimensions of the tested component 6 are qualified based on the verification results.

[0043] This utility model discloses a machine vision-based large component appearance dimension inspection system. By setting up a large installation platform, multiple industrial cameras are installed on the installation platform, and combined with a server and a central control box, it is suitable for the appearance dimension inspection of large components. At the same time, the overall cost of the device is low, which is conducive to its widespread use. A conveying device is provided below the installation platform to improve the automation of the inspection process. The installation platform is a multi-axis micro-motion platform, and it is also equipped with edge computing devices and multiple computing nodes to improve the data processing capabilities of the server and thus improve the inspection speed.

[0044] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present utility model and is not intended to limit the present utility model. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present utility model should be included within the protection scope of the present utility model.

Claims

1. A machine vision-based system for inspecting the appearance and dimensions of large components, characterized in that, The system includes an installation platform (1), a mounting bracket (2), an industrial camera (3), a central control box (4), a server (5), and a conveying device (7). The mounting bracket (2) is set vertically to the ground and connected to the installation platform (1). There are multiple industrial cameras (3) evenly arranged on the installation platform (1). The central control box (4), the server (5), and the conveying device (7) are all located below the installation platform (1). The industrial camera (3) is connected to the server (5) through the central control box (4). The conveying device (7) is located on the ground and is used to transport the component (6) to be measured to the measurement center area below the installation platform (1).

2. The machine vision-based large component appearance dimension inspection system according to claim 1, characterized in that, The installation platform (1) is a rectangular frame with a hollow center, and three to eight industrial cameras (3) are installed on the edges of the frame.

3. The machine vision-based large component appearance and dimension inspection system according to claim 1, characterized in that, The industrial camera (3) is model MER2-2000-6GC.

4. The machine vision-based large component appearance and dimension inspection system according to claim 2, characterized in that, The mounting platform (1) is also equipped with tilt sensors combined with cameras or digital level integrated cameras at the four edges and corners.

5. The machine vision-based large component appearance and dimension inspection system according to claim 1, characterized in that, The length of the installation platform (1) ranges from 12m to 16m, and the width ranges from 8m to 12m.

6. The machine vision-based large component appearance and dimension inspection system according to claim 1, characterized in that, There are six fixing frames (2) symmetrically arranged on both sides of the mounting platform (1).

7. The machine vision-based large component appearance and dimension inspection system according to claim 1, characterized in that, The central control box (4) is equipped with a power supply and a wireless communication network, wherein the wireless communication network is 5G or WiFi 6.

8. The machine vision-based large component appearance and dimension inspection system according to claim 1, characterized in that, The conveying device (7) is a retractable conveyor belt or guide rail system.

9. The machine vision-based large component appearance dimension inspection system according to claim 1, characterized in that, The installation platform (1) is a multi-axis micro-motion platform.

10. The machine vision-based large component appearance and dimension inspection system according to claim 1, characterized in that, An edge computing device or multiple computing nodes are provided between the central control box (4) and the server (5).

Images