Visual identification device for grouting engineering pressure panel

Abstract

The application relates to a visual recognition device for a grouting engineering pressure dial, which comprises a dial recognition algorithm, a data transmission algorithm and a hardware system; the dial recognition algorithm is a deep learning-based object detection algorithm, and the detection algorithm is a yolov5 algorithm; the data transmission algorithm is a TCP client created on a Raspberry Pi by using a Socket programming library in C++, which is used for communication with an upper computer supporting TCP communication to realize mutual communication of reading results, and the upper computer is a TCP server during work. The application can efficiently process and analyze dial images captured by a camera. The device can quickly and accurately recognize the position and state of a dial pointer under various industrial environments, which provides important support for subsequent production or maintenance work; the device can realize data interaction and linkage control among a display screen, a camera, a PLC and other equipment, and further improves the comprehensive performance of the device.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent construction equipment for water conservancy and hydropower engineering grouting, image processing and object recognition, in particular to a visual recognition device for grouting engineering pressure gauge. BACKGROUND

[0002] In cement grouting engineering, monitoring the grouting pressure gauge value is crucial to ensure construction quality. The installation location of the monitoring equipment may not be very flexible, and may not meet the actual construction needs. Traditional monitoring methods may require manual observation and recording. Manual checking of the gauge has some difficulties and limitations. First, human errors or negligence may result in inaccurate readings, affecting construction quality and safety. Second, for large-scale engineering sites or situations requiring 24-hour monitoring, human resources costs are high and not sustainable. In addition, long-term continuous work can cause personnel fatigue, reducing the accuracy and efficiency of monitoring. SUMMARY

[0003] The present application aims to provide a visual recognition device for grouting engineering pressure gauges, solving the problems of low recognition efficiency, misreading and high cost caused by traditional manual observation and reading, to promote the development of automation and intelligentization in industrial production and equipment maintenance, and improve production efficiency and reduce costs.

[0004] To achieve the above purpose, the technical solution adopted by the present application is: a visual recognition device for grouting engineering pressure gauges, including a gauge recognition algorithm, a data transmission algorithm, and a hardware system.

[0005] The gauge recognition algorithm is a deep learning-based object detection algorithm, and the detection algorithm is yolov5 algorithm.

[0006] The data transmission algorithm uses the Socket programming library in C++ to create a TCP client on the Raspberry Pi to communicate with the host computer that supports TCP communication, thereby realizing the intercommunication of reading results. The host computer is the TCP server when working.

[0007] The hardware system includes a Raspberry Pi, a display screen, a camera, and a device shell.

[0008] The Raspberry Pi is a small single-board computer used to load yolov5 algorithm, data transmission algorithm and connect hardware.

[0009] The display screen is used to display the camera-captured image and the dial pointer position and state recognized by the yolov5 algorithm in real time, and focuses on displaying the image and the recognition result, thereby providing intuitive and clear information feedback for the user;

[0010] The camera is used to capture the dial image, so that the captured image can present detailed and clear features.

[0011] The device shell integrates a raspberry pi slot, a camera slot, a display screen slot, a dial slot, a heat dissipation hole and a handle, thereby providing protection and support for the composition of the entire device hardware part.

[0012] In the above technical solution, the raspberry pi slot is used to fix the raspberry pi single board computer; the camera slot is used to install the camera module; the display screen slot is used to install the display screen module; the dial slot is used to install the dial of the equipment to be maintained; the heat dissipation hole is used to dissipate the heat generated inside the raspberry pi single board computer; and the handle is convenient for the maintenance personnel to carry and operate the device.

[0013] In the above technical solution, the raspberry pi single board computer is equipped with a yolov5 algorithm, supports the running of an image processing and recognition algorithm, and contains a plurality of input and output interfaces for connecting with the camera, the display screen and external equipment; and has an Ethernet data communication function, and can transmit the recognition result to an external PLC device.

[0014] In the above technical solution, the device shell integrates the raspberry pi, the camera, the display screen and the dial together.

[0015] In the above technical solution, the device is connected to the external equipment through Ethernet, and can share and save the dial historical data.

[0016] In the above technical solution, the yolov5 algorithm mainly includes three parts: a backbone network Backbone, a feature fusion layer Neck and an output end Head, wherein the backbone network Backbone contains a CSP and an EfficientNet algorithm with good feature extraction capability to extract image features, the feature fusion layer Neck extracts features through an SPPF feature extraction mechanism, and fuses semantic features and position information at different levels to improve the detection effect of the network on objects of different scales, and the output end Head is responsible for mapping the extracted feature map to a prediction frame and outputting the final detection result.

[0017] The visual recognition device for the grouting engineering pressure dial can efficiently process and analyze the dial image captured by the camera. Its excellent performance ensures that the device can quickly and accurately identify the position and state of the dial pointer in various industrial environments, which provides important support for subsequent production or maintenance work; at the same time, the raspberry pi has rich interfaces and can be connected with other devices, which enables the device to realize data interaction and linkage control between the display screen, camera and PLC and other devices, further improving the comprehensive performance of the device. The specific advantages are as follows:

[0018] (1) The position and state of the dial pointer can be accurately identified, avoiding the error of manual identification and ensuring the accuracy of monitoring.

[0019] (2) It can be easily carried to different work sites or fixedly installed for use, providing users with more flexible and convenient operation experience.

[0020] (3) Without complex installation and debugging process, users can quickly deploy the device and immediately put it into use, saving a lot of installation time and cost.

[0021] (4) It is suitable for different environments and working scenes, such as indoor, outdoor, mobile vehicles, etc., and has strong adaptability. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 The network structure diagram of the dial recognition algorithm yolov5

[0023] Figure 2 The structure diagram of the data transmission algorithm

[0024] Figure 3 The isometric structural view of a visual recognition device for a grouting engineering pressure dial.

[0025] Figure 4 The isometric structural view of a visual recognition device for a grouting engineering pressure dial.

[0026] Figure 5 The isometric structural view of a visual recognition device for a grouting engineering pressure dial.

[0027] Figure 6 The A-A sectional view of a visual recognition device for a grouting engineering pressure dial.

[0028] Figure 7 The front view structural diagram of a visual recognition device for a grouting engineering pressure dial.

[0029] Figure 8It is a right view structure diagram of a visual identification device for a grouting engineering pressure dial.

[0030] In the figure: 1 - display screen, 2 - camera, 3 - Raspberry Pi, 4 - device shell, 4.1 - handle, 4.2 - display screen slot, 4.3 - camera slot, 4.4 - Raspberry Pi slot, 4.5 - heat dissipation hole, 4.6 - dial slot. DETAILED DESCRIPTION

[0031] The embodiments of the present application will be described in detail below with reference to the accompanying drawings, but they do not constitute limitations on the present application, and are only exemplary, and at the same time, the advantages of the present application will become clearer and easier to understand through the description.

[0032] As can be seen from the drawings, the visual identification device for a grouting engineering pressure dial comprises a dial recognition algorithm, a data transmission algorithm, and a hardware system.

[0033] The dial recognition algorithm is an object detection algorithm based on deep learning, and the detection algorithm is a yolov5 algorithm.

[0034] The data transmission algorithm is a Socket programming library in C++ used to create a TCP client on Raspberry Pi to communicate with the host computer supporting TCP communication to realize the intercommunication of reading results, and the host computer is a TCP server when working.

[0035] The hardware system comprises a Raspberry Pi, a display screen, a camera, and a device shell.

[0036] The Raspberry Pi is a small single-board computer used to carry the yolov5 algorithm, the data transmission algorithm, and the connection hardware.

[0037] The display screen is used to display the image captured by the camera and the dial pointer position and state recognized by the yolov5 algorithm in real time, and is focused on displaying the image and the recognition result to provide intuitive and clear information feedback for the user.

[0038] The camera is used to capture the dial image to ensure that the captured image can present detailed and clear features.

[0039] The device shell integrates the Raspberry Pi slot, the camera slot, the display screen slot, the dial slot, the heat dissipation hole, and the handle to provide protection and support for the composition of the entire device hardware part.

[0040] The Raspberry Pi slot is used to fix the Raspberry Pi single-board computer; the camera slot is used to install the camera module; the display screen slot is used to install the display screen module; the dial slot is used to install the dial of the equipment to be maintained; the heat dissipation hole is used to dissipate the heat generated inside the Raspberry Pi single-board computer; and the handle is convenient for the maintenance personnel to carry and operate the device.

[0041] The Raspberry Pi single-board computer is equipped with a Yolov5 algorithm, supports the running of image processing and recognition algorithms, and includes multiple input and output interfaces for connecting with a camera, a display screen, and external devices; has an Ethernet data communication function and can transmit recognition results to external PLC devices.

[0042] The device housing integrates the Raspberry Pi, the camera, the display screen, and the dial together. Figure 3 、 Figure 4 、 Figure 5 、 Figure 6 、 Figure 7 、 Figure 8

[0043] The device is connected to external devices through Ethernet and can share and save dial historical data.

[0044] The yolov5 algorithm mainly includes three parts: a backbone network Backbone, a feature fusion layer Neck, and an output end Head, wherein the backbone network Backbone includes CSP and EfficientNet algorithms with good feature extraction capability and is used to extract image features such as dial pictures; the feature fusion layer (Neck) extracts features through the SPPF feature extraction mechanism and fuses semantic features and position information at different levels to improve the detection effect of the network on objects of different scales; the output end Head is responsible for mapping the extracted feature map to a prediction frame and outputting the final detection result in a convolutional layer Convolutional Layer; as shown in Figure 1

[0045] Embodiment 1

[0046] Fire extinguishers are usually installed in hidden or high places, and manual checking requires climbing or using ladders and other operations, which poses a safety risk. Secondly, fire extinguishers may be placed in harsh environments such as high temperature, low temperature, dust, etc., which will affect the effect and accuracy of manual checking. Therefore, the visual recognition device for the pressure dial of the grouting engineering has the principle and structure as shown in Figure 1 、 2 , 3, 4, 5, 6, 7, 8, which includes a display screen 1, a camera 2, a Raspberry Pi 3, a device housing 4, a handle 4.1, a display screen slot 4.2, a camera slot 4.3, a Raspberry Pi slot 4.4, a heat dissipation hole 4.5, and a dial slot 4.6.

[0047] The display screen 1 is used to display the image captured by the camera and the dial pointer position and state recognized by the yolov5 algorithm in real time. It focuses on displaying the image and the recognition result to provide intuitive and clear information feedback for the user;

[0048] ​​The camera 2 is mainly used to capture the dial image, ensuring that the captured image can present detailed and clear features. High-resolution images can provide more information, which helps the subsequent image processing algorithm to accurately identify and locate the dial pointer;

[0049] The Raspberry Pi 3 is a small single-board computer with powerful computing power and rich interfaces, which is the core control unit of the application. Through the application of yolov5 algorithm, the device can efficiently process and analyze the images captured by the camera. Its excellent performance ensures that the device can quickly and accurately identify the position and state of the dial pointer in various industrial environments, which provides important data support for subsequent production or maintenance work. Raspberry Pi also has rich interfaces, which can be connected and communicated with other devices, making the device able to realize data interaction and linkage control with display screen, camera and PLC and other devices, further improving the comprehensive performance of the device;

[0050] The device housing 4 integrates key components such as handle 4.1, display screen slot 4.2, camera slot 4.3, Raspberry Pi slot 4.4, heat dissipation hole 4.5, dial slot 4.6, etc., providing protection and support for the entire device, while also facilitating carrying and operation;

[0051] The handle 4.1 facilitates the maintenance personnel to carry and operate the device. The handle is comfortable and easy to hold, which can effectively reduce the burden of carrying the device and improve the convenience and comfort of operation.

[0052] The display screen slot 4.2 is used to install the display screen module, so that the display screen can be safely fixed on the device, providing clear image and identification result display for maintenance personnel.

[0053] The camera slot 4.3 is used to install the camera module, so that the camera can be accurately positioned and fixed on the device, ensuring that the camera can accurately capture the image of the dial and maintain a relatively stable working state;

[0054] The Raspberry Pi slot 4.4 is used to fix the Raspberry Pi single-board computer, so that the Raspberry Pi can be stably installed inside the device, ensuring that it will not fall off due to vibration or movement, and ensuring the stability and reliability of the device;

[0055] The heat dissipation hole 4.5 is used to dissipate the heat generated inside the Raspberry Pi, preventing the device from being damaged due to overheating, and ensuring its stable and reliable operation;

[0056] The dial slot 4.6 is used to place the dial of the device to be maintained on the device, so that the dial can be accurately placed on the device and maintain a relatively stable position, providing a clear view for the image captured by the camera;

[0057] The pointer type dial recognition device of the application works as follows: the operator connects the power cord of Raspberry Pi 3 to the power supply, and connects the Raspberry Pi and the PLC (or other host computer supporting TCP communication) system through Ethernet. After the Raspberry Pi 3 is powered on, it also powers the entire device system, i.e. the display screen 1 and the camera 2, at this time the LED light is on and the camera starts working. Set the PLC network address to ensure that the PLC and the Raspberry Pi 3 realize data communication. Position the pointer type dial in the dial slot 4.6, the PLC sends a request signal, and after the Raspberry Pi 3 receives the signal, it controls the camera 2 to take a picture of the dial. After the image is quickly and accurately processed by the Raspberry Pi, the display screen 1 displays the camera picture and the recognition result, and the recognition result is transmitted back to the PLC. After use, the operator disconnects the power supply and Ethernet connection to complete the use of the device. As shown in Figure 2 .

[0058] Example 2

[0059] In a certain cement grouting project, due to the low degree of intelligence of the grouting equipment, a mechanical pressure dial is used to measure the pressure value of the grouting equipment, which cannot realize automatic data uploading. At the same time, in the grouting project, the installation position of the monitoring equipment is tricky, and the form of manually confirming the pressure dial reading is complex, which cannot meet the actual construction requirements. Therefore, the application provides a visual recognition device for the pressure dial of the grouting project, which is fixedly installed opposite to the pressure dial of the grouting equipment to monitor the reading of the dial. When the pressure value exceeds the safe pressure range, an alarm will be triggered. The principle and structure thereof are shown in Figure 1 , 2 , 3, 4, 5, 6, 7, 8, which comprises a display screen 1, a camera 2, a Raspberry Pi 3, a device shell 4, a display screen slot 4.2, a camera slot 4.3, a Raspberry Pi slot 4.4, a heat dissipation hole 4.5, and a dial slot 4.6. The device has the following characteristics:

[0060] The display screen 1 is used to display the image captured by the camera and the dial pointer position and state recognized by the yolov5 algorithm in real time. It focuses on displaying the image and the recognition result, and provides intuitive and clear information feedback for the user;

[0061] The camera 2 is mainly used to capture the dial image, and ensure that the captured image can present detailed and clear features. High-resolution images can provide more information, which is helpful for the subsequent image processing algorithm to accurately recognize and locate the dial pointer;

[0062] The Raspberry Pi 3 is a small single-board computer with powerful computing power and rich interfaces, which is the core control unit of the application. Through the application of the yolov5 algorithm, the device can efficiently process and analyze the images captured by the camera. Its excellent performance ensures that the device can quickly and accurately identify the position and state of the dial pointer in various industrial environments, providing important data support for subsequent production or maintenance work. Raspberry Pi also has rich interfaces that can connect and communicate with other devices, enabling the device to interact with display screens, cameras, PLCs, and other devices, further enhancing the overall performance of the device.

[0063] The device housing 4 integrates key components such as the handle 4.1, display screen slot 4.2, camera slot 4.3, Raspberry Pi slot 4.4, heat dissipation hole 4.5, and dial slot 4.6, providing protection and support for the entire device, while also facilitating portability and operation.

[0064] The handle 4.1 facilitates the carrying and operation of the device by maintenance personnel. The handle is comfortable and easy to hold, effectively reducing the burden of carrying the device and improving the convenience and comfort of operation.

[0065] The display screen slot 4.2 is used to install the display screen module, allowing the display screen to be securely fixed on the device and providing clear image and recognition result display for maintenance personnel.

[0066] The camera slot 4.3 is used to install the camera module, allowing the camera to be accurately positioned and fixed on the device, ensuring that the camera can accurately capture images of the dial and maintain a relatively stable working state.

[0067] The Raspberry Pi slot 4.4 is used to secure the Raspberry Pi single-board computer, allowing the Raspberry Pi to be stably installed inside the device, ensuring that it does not fall off due to vibration or movement, and ensuring the stability and reliability of the device.

[0068] The heat dissipation hole 4.5 is used to dissipate the heat generated inside the Raspberry Pi, preventing the device from being damaged due to overheating, and ensuring its stable and reliable operation.

[0069] The dial slot 4.6 is used to place the dial of the device to be maintained on the device, allowing the dial to be accurately placed on the device and maintaining a relatively stable position, providing a clear view for the images captured by the camera.

[0070] The pointer type dial identification device of the application works as follows: the operator connects the power cord of Raspberry Pi 3 to the power supply, and connects the Raspberry Pi and the PLC (or other host computer supporting TCP communication) system through Ethernet. After the power supply of Raspberry Pi 3 is turned on, the entire device system, i.e. the display screen 1 and the camera 2, is powered, at this time the LED light is on and the camera starts to work. The network address of the PLC is set to ensure that the PLC and Raspberry Pi 3 realize data communication. The pointer type dial is positioned in the dial slot 4.6, the PLC sends a request signal, Raspberry Pi 3 receives the signal and controls the camera 2 to take pictures of the dial, the image is processed quickly and accurately by Raspberry Pi, the display screen 1 displays the camera shooting picture and the identification result, and the identification result is transmitted back to the PLC. After use, the operator disconnects the power supply and Ethernet connection to complete the use of the device, as shown in Figure 2

[0071] The others not specifically described are prior art.​

Claims

1. A visual recognition device for grouting engineering pressure dial, characterized in that: It comprises dial recognition algorithm, data transmission algorithm and hardware system; The dial recognition algorithm is an object detection algorithm based on deep learning, and the detection algorithm is yolov5 algorithm; The data transmission algorithm is to create a TCP client on Raspberry Pi using the Socket programming library in C++, which is used to communicate with the host computer supporting TCP communication to realize the intercommunication of reading results, and the host computer is the TCP server during work; The hardware system comprises Raspberry Pi, display screen, camera and device shell; The Raspberry Pi is a small single-board computer used to carry yolov5 algorithm, data transmission algorithm and connected hardware; The display screen is used to display the image captured by the camera and the dial pointer position and state recognized by the yolov5 algorithm in real time, and is focused on displaying the image and the recognition result to provide intuitive and clear information feedback for the user; The camera is used to capture the dial image to ensure that the captured image can present detailed and clear features; The device shell integrates Raspberry Pi slot, camera slot, display screen slot, dial slot, heat dissipation hole and handle to provide protection and support for the composition of the entire device hardware part; The Raspberry Pi slot is used to fix the Raspberry Pi single-board computer; the camera slot is used to install the camera module; and the display screen slot is used to install the display screen module; The dial slot is used to install the dial of the equipment to be maintained; the heat dissipation hole is used to dissipate the heat generated inside the Raspberry Pi single-board computer; and the handle is convenient for the maintenance personnel to carry and operate the device; The Raspberry Pi single-board computer carries yolov5 algorithm, supports the running of image processing and recognition algorithm, and contains multiple input and output interfaces for connection with the camera, display screen and external equipment; and has Ethernet data communication function to transmit the recognition result to external PLC equipment; The device is connected to external equipment through Ethernet, and can share and save dial historical data; The yolov5 algorithm comprises three parts: backbone network Backbone, feature fusion layer Neck and output end Head, wherein the backbone network Backbone contains CSP and EfficientNet algorithms with good feature extraction capability to extract image features, the feature fusion layer Neck extracts features through the feature extraction mechanism of SPPF, and fuses semantic features and position information of different levels to improve the detection effect of the network on objects of different scales, and the output end Head is responsible for mapping the extracted feature map to the prediction frame and outputting the final detection result.

2. A visual identification device for grouting engineering pressure gauges according to claim 1, characterized in that The device shell integrates Raspberry Pi, camera, display screen and dial together.

Images