Automatic identification method and device for indicating value of float flowmeter
Through the analysis of industrial cameras and target detection network model, combined with the lifting mechanism to track the top view float, the automatic display value recognition of the float flowmeter is realized, solving the problems of large and small verification difficulties in the existing technology, and improving the measurement accuracy and efficiency.
Patent Information
- Application Number
- CN202411230658.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-03
- Publication Date
- 2024-12-06
AI Technical Summary
During the verification process, the existing float flowmeter is susceptible to pressure, temperature and other conditions, and the measurement is difficult. There are large errors and low efficiency in manual interpretation, especially the small flow verification is inaccurate, and traditional automatic recognition algorithms are difficult to deal with uneven lighting. Complex background and other issues.
An industrial camera is used to collect images, and the digital identification and bounding box coordinates of the float flowmeter surface are obtained through the target detection network model analysis. The scale coordinates are calculated using the proportional method, and the lifting mechanism is controlled to track the top view float, so as to determine the flow range and calculate the flow display value.
It improves the accuracy and efficiency of the float flowmeter verification, and can automatically identify the display values of float flowmeters of different specifications, which are easy to operate, and have high accuracy and robustness.
Smart Images

Figure CN119091425A_ABST
Abstract
Description
[0002] The present invention studies a method and device for automatically identifying the indication of a float flowmeter. The method collects images through an industrial camera, imports them into a target detection network model for analysis, obtains the digital identification of the flowmeter surface and the bounding box coordinates of the float, calculates the scale line coordinates by offsetting the proportional method, and then determines the movement of the lifting mechanism according to the center coordinates of the float to track the horizontal float. Finally, the flow range corresponding to the float position is determined, and the current flow indication is calculated using the proportional method. The present invention can realize the indication identification of float flowmeters of different specifications, is easy to operate, and improves efficiency and accuracy. Technical Field
[0003] The present invention relates to the field of deep learning technology, and in particular to a method and device for identifying a float flowmeter indication based on a target detection method. Background Art
[0004] The float flowmeter is an instrument that measures flow based on the throttling principle. It keeps the differential pressure above and below the float constant by changing the flow area of the fluid. It has the advantages of simple structure, intuitive scale value, easy use and maintenance, small and constant pressure loss, etc. It is widely used in the field of industrial flow measurement, and needs to be regularly traced (calibrated or calibrated) to ensure the accuracy of its indication. However, during the calibration process, the gas medium is easily affected by conditions such as pressure and temperature, the measurement is difficult, and the work efficiency is not high; in addition, it also involves the problem of inaccurate calibration of small flow. When manually interpreting the float flowmeter, the human eye and the tangency of the float and the numerical scale line are used as correct operations. There are problems such as large interpretation errors and low efficiency. It is easily affected by the operator's skill level and subjective judgment, resulting in unstable reading results. Therefore, a more efficient and accurate automatic recognition technology is needed to improve the calibration accuracy and efficiency of the float flowmeter. The present invention incorporates deep learning technology to solve this problem, and uses industrial cameras to replace human eyes to observe and interpret, which is of great significance.
[0005] Traditional automatic meter reading recognition algorithms are based on computer image processing algorithms, which cannot solve problems such as uneven lighting, large lighting changes, complex backgrounds, image blur, meter tilt, scale changes, and cumbersome image preprocessing. In recent years, deep learning technology has gradually shown great potential in the field of computer vision. Deep learning methods build multi-layer neural networks and use nonlinear mapping relationships to learn and represent features of input data, thereby achieving recognition and understanding of complex patterns. Deep learning methods have powerful feature learning and abstraction capabilities, can effectively handle complex visual tasks, and have good robustness.
[0006] In the process of using an industrial camera to read the float flowmeter, it is important to align the line of sight with the tangent of the float, and the float will move during the flow change process. The present invention uses a lifting mechanism to track the horizontal float, which can further improve the accuracy of the reading.
[0007] The measuring ranges of most traditional instruments are uniformly distributed as a whole, while the measuring range uniformity of the float flowmeter is segmented, that is, the ranges between each two adjacent scale lines are uniformly distributed, and there is no correlation between the flow ranges. This problem leads to the need to segment the flowmeter during the identification process. The present invention has formulated a solution to this problem. Summary of the invention
[0008] The object of the present invention is to provide a method and device, system and equipment for identifying the indication of a float flowmeter, which can realize automatic tracking and identification of the indication of the float flowmeter, is easy to operate, and improves efficiency and accuracy.
[0009] In order to achieve the above purpose, the following technical solutions are proposed:
[0010] A first aspect of an embodiment of the present invention discloses a method for identifying an indication of a float flowmeter, comprising:
[0011] The industrial camera is connected to the host computer for communication, the image of the float flowmeter is collected and uploaded, and the image is initially processed such as image scaling, binarization, and image filtering;
[0012] Importing the image after the preliminary processing into the target detection network model to obtain the digital identification on the surface of the float flowmeter and the bounding box information of the float;
[0013] Analyzing the coordinates of the vertices extracted from the digital identification boundary box, and obtaining the coordinate values of the scale lines by offset calculation;
[0014] Analyzing the center coordinates of the float boundary frame extraction, and controlling the lifting mechanism to track the horizontal float;
[0015] According to the coordinates of the center of the float and the coordinates of each scale line, the flow range of the float position is determined, and then the flow indication is calculated using the proportional method.
[0016] In one embodiment, after the image is imported into the target detection network model, the image is segmented into SxS grids, and for each grid we predict B bounding boxes (including the confidence probability that each bounding box is an object and the probability of each bounding box area in multiple categories);
[0017] The predicted SxSxB bounding boxes are then removed based on a threshold value, and finally redundant bounding boxes (i.e., bounding boxes where the same object is detected multiple times) are removed based on the NMS algorithm;
[0018] The float flowmeter surface digital identification and float bounding box information include their position information (x, y, w, h) and classification information (c), where x and y represent the upper left corner coordinates of the bounding box, w and h represent the width and height of the bounding box, and c represents the category to which it belongs.
[0019] In one embodiment, the target detection network model imported into the image includes:
[0020] The indication recognition of the float flowmeters of various specifications is achieved by collecting a large amount of image data of float flowmeters of various specifications and obtaining a stable large network model through training. The model can complete the indication recognition of the flowmeters of various specifications.
[0021] In one embodiment, the offset calculation of extracting vertex coordinates of the digital identification bounding box includes:
[0022] The range uniformity of the float flowmeter is segmented, that is, the range between each two adjacent scale lines is evenly distributed, so the coordinates need to be located at the boundary of each flow range;
[0023] According to the coordinate position information extracted from the digital identification boundary box, the height ratio of the digital identification frame is calculated and offset downward to locate the scale boundary, and the algorithm is repeatedly executed to divide the total range of the float flowmeter into multiple uniform small ranges.
[0024] In one embodiment, the method for tracking a head-up float by an industrial camera comprises:
[0025] Extracting the center coordinates of the float boundary box, and comparing the center coordinate value with the center coordinate of the industrial camera screen to determine whether to issue a control instruction;
[0026] In one embodiment, the method for calculating the ratio of the flow indication includes:
[0027] The center coordinates of the float boundary box are extracted and compared with the various dividing boundaries located by digital identification to determine the small range of the float. The proportional method is used to calculate the difference between the float coordinates and the lower limit coordinates, and the difference between the upper limit coordinates and the lower limit coordinates. The current indication of the flow meter is obtained by multiplying the ratio of the two with the product of the small range and then adding the flow value corresponding to the lower limit.
[0028] A second aspect of an embodiment of the present invention discloses a tracking head-up device for a float flowmeter, comprising:
[0029] The center coordinate value of the float is subtracted from the center coordinate value of the industrial camera screen, and the difference is judged to control the start and stop of the lifting mechanism;
[0030] The lifting mechanism is composed of a sliding guide rail and an industrial camera;
[0031] If it is greater than the upper warning value, the host computer algorithm determines and sends a command to the single-axis controller to control the motor to rotate forward, and the slide rail carrying the industrial camera rises;
[0032] If it is less than the lower limit warning value, the host computer algorithm will issue a command to the single-axis controller to control the motor to reverse, and the slide rail carrying the industrial camera will descend;
[0033] If it is between the upper warning limit and the lower warning limit, the motor stops and it is judged as a horizontal float;
[0034] The algorithm is continuously cycled to control the guide rail equipped with the industrial camera to move up and down, thereby achieving tracking and level viewing of the float.
[0035] The method, device, system and equipment for identifying the indication of a float flowmeter provided by the present invention perform target detection and analysis on the image collected by the industrial camera, obtain the coordinate information of each target boundary box, divide a plurality of uniform range areas, control the lifting mechanism to carry a camera to track the horizontal float, and calculate the current indication of the flowmeter by proportional calculation. The present invention uses a large network model of various specifications, has strong adaptability, can realize the indication identification of float flowmeters of different specifications, is easy to operate and has high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] The following drawings show specific examples of the technical solutions of the present invention, and are used to explain the content, principle and effect of the present invention.
[0037] Unless otherwise specified or defined, the same reference numerals in different drawings represent the same or similar technical features, and the same or similar technical features may also be represented by different reference numerals.
[0038] Figure 1 It is a flow chart of the method for identifying the indication of a float flowmeter disclosed in an embodiment of the present invention;
[0039] Figure 2 It is a schematic diagram of bounding box information extracted after target detection by the method for identifying the indication of a float flowmeter disclosed in an embodiment of the present invention;
[0040] Figure 3 It is a schematic diagram of the indication recognition method of a float flowmeter disclosed in an embodiment of the present invention and the offset positioning of a digital identification boundary box;
[0041] Figure 4 It is a schematic diagram of the method for identifying the indication of a float flowmeter disclosed in an embodiment of the present invention and calculating the current flow value by a proportional method;
[0042] Figure 5The invention is a schematic diagram of a method for identifying the indication of a float flowmeter and a device for tracking a horizontal float disclosed in an embodiment of the invention.
[0043] Description of reference numerals:
[0044] 501. float flowmeter; 502. industrial camera; 503. sliding guide rail; 504. motor; 505. motor drive; 506. single-axis controller; 507. host computer. Specific implementation plan
[0045] In order to facilitate the understanding of the present invention, specific embodiments of the present invention will be described in more detail below with reference to the accompanying drawings.
[0046] Unless otherwise specified or defined, all technical and scientific terms used herein have the same meaning as those generally understood by those skilled in the art. In combination with the technical solution of the present invention in a realistic scenario, all technical and scientific terms used herein may also have meanings corresponding to the technical solution for implementing the present invention. The "first, second..." used herein is merely used to distinguish the names and does not represent a specific quantity or order. The term "and / or" used herein includes any and all combinations of one or more related listed items.
[0047] Unless otherwise specified or defined, the “said” and “the” used in this document refer to the technical features or contents mentioned or described before the corresponding position, and the technical features or technical contents may be the same as or similar to the technical features or technical contents mentioned therein.
[0048] Undoubtedly, technical contents or technical features that are contrary to the purpose of the present invention or are obviously contradictory should be excluded.
[0049] like Figure 1 As shown, an embodiment of the present invention discloses a method for identifying the indication of a float flowmeter, comprising:
[0050] 1. Extract the floating flowmeter imaging picture collected by the industrial camera. First, perform preliminary processing such as image scaling, binarization, and image filtering on the image. Then, import the processed image into the target detection network model for analysis to obtain the bounding box information of each target. Figure 2 shown.
[0051] 2. Extract the bounding box coordinates of the digital mark on the flow meter surface, and obtain the scale line coordinates corresponding to each flow value through proportional offset calculation.
[0052] Specifically, the position information (x, y, w, h) of the digital marker is extracted, where x and y represent the coordinates of the upper left corner of the bounding box, and w and h represent the width and height of the bounding box. Figure 3As shown in the embodiment, the position information y6 and y4 of the digital symbols "6" and "4" in the figure, and the heights h6 and h4 of the bounding boxes are extracted, and the coordinates Y6 and Y4 of the scale lines corresponding to "6" and "4" are located by formula ratio calculation. Figure 3 The geometric relationship shown in the figure can be obtained by the following calculation formulas (1) and (2):
[0053] Y6=y6-1.5*h6 (1)
[0054] Y4=y4-1.5*h4 (2)
[0055] 3. Extract the center coordinates of the float boundary box and compare the center coordinates with the center coordinates of the industrial camera image to determine whether to issue a control instruction.
[0056] Specifically, the center coordinate value of the float is subtracted from the center coordinate value of the industrial camera screen, and the difference is judged to control the start and stop of the lifting mechanism;
[0057] If it is greater than the upper warning value, the host computer algorithm determines and sends a command to the single-axis controller to control the motor to rotate forward, and the slide rail carrying the industrial camera rises;
[0058] If it is less than the lower limit warning value, the host computer algorithm will issue a command to the single-axis controller to control the motor to reverse, and the slide rail carrying the industrial camera will descend;
[0059] If it is between the upper warning limit and the lower warning limit, the motor stops and it is judged as a horizontal float;
[0060] The algorithm is continuously cycled to control the guide rail equipped with the industrial camera to move up and down, thereby achieving tracking and level viewing of the float.
[0061] 4. Extract the center coordinates of the float boundary box, compare them with the boundaries located by digital markings, determine the small range of the float, and calculate the current flow indication by proportional method.
[0062] Specifically, the difference between the float coordinate and the lower limit coordinate (Yf-Y4), and the difference between the upper limit coordinate and the lower limit coordinate (Y6-Y4) are calculated, and the current flow indication (Fval) is obtained through proportional calculation. Figure 4 The geometric relationship shown in the figure can be obtained by the following calculation formula (3):
[0063]
[0064] The purpose of the above embodiments is to exemplarily reproduce and deduce the technical solution of the present invention, and to fully describe the technical solution, purpose and effect of the present invention. Its purpose is to enable the public to have a more thorough and comprehensive understanding of the disclosed content of the present invention, and it does not limit the protection scope of the present invention.
[0065] The above embodiments are not exhaustive enumerations of the present invention, and there may be multiple other implementations not listed. Any replacement and improvement made without violating the concept of the present invention shall fall within the protection scope of the present invention.
Claims
1. A method for automatically identifying the indication of a float flowmeter, characterized in that: include: Perform image processing on the images captured by the industrial camera and import them into the target detection network model to obtain the digital logo on the surface of the float flowmeter and the bounding box of the float; Analyzing the coordinates of the vertices extracted from the digital identification boundary box, and obtaining the coordinate values of the scale lines by offset calculation; Analyzing the center coordinates of the float boundary box extraction, and controlling the guide rail to move and track the horizontal float; According to the coordinates of the center of the float and the coordinates of each scale line, the flow range of the float position is determined, and then the flow indication is calculated using the proportional method.
2. The method for automatically identifying the indication of a float flowmeter according to claim 1, characterized in that: The range uniformity of the float flowmeter is segmented, and the distance between two adjacent scale lines is evenly distributed. The coordinates extracted from the digital identification boundary box are located to the corresponding scale line coordinates according to the proportional method.
3. The method and device for automatically identifying the indication of a float flowmeter according to claim 1, characterized in that: The lifting device for tracking the horizontal float uses a sliding guide rail and an industrial camera combination, the camera is fixed on the guide rail to face the float flowmeter, the host computer judges and sends instructions to the single-axis controller, the controller sends PWM waves to the motor to rotate forward and reverse, and the guide rail moves up and down to achieve the effect of the picture following the movement; Extracting the center coordinates of the float boundary frame, performing difference calculation and comparison between the float coordinates and the center coordinates of the screen, and controlling the guide rail to move the camera at different speeds to track the horizontal float; If there is no float bounding box information in the camera extraction picture, the algorithm will control the guide rail to reverse to the bottom limiter and continue to rise. After the float bounding box is captured in the picture, it will start tracking and looking straight ahead.
4. The method and device for automatically identifying the indication of a float flowmeter according to claim 3, characterized in that: If there is no float bounding box information in the camera extraction picture, the algorithm will control the guide rail to reverse to the bottom stopper and continue to rise, and start tracking after the float bounding box is captured in the picture; Two high and low limiters are added to the sliding guide rail and camera combination, and the heights correspond to the top and bottom of the float flow meter, thereby improving the efficiency of finding the float boundary box.
5. The method for automatically identifying the indication of a float flowmeter according to claim 1, characterized in that: The vertex coordinates of the digital identification boundary box are extracted, the scale line coordinates are offset and positioned, the float center coordinates are compared with the scale line coordinates, the flow range of the current float position is determined, and the coordinate difference is proportionally calculated to obtain the flow indication.
6. The method for automatically identifying the indication of a float flowmeter according to claim 1, characterized in that: The image information collected by industrial cameras needs to be processed, including image scaling, binarization, image filtering, etc., which can denoise, smooth and enhance the image, and improve the positioning accuracy of the subsequent target detection bounding box.
7. The method for automatically identifying the indication of a float flowmeter according to claim 1, characterized in that: The target detection network model imported into the image uses the YOLOV8 target detection model. Compared with traditional machine vision technology, deep learning technology is less affected by environmental changes and has stronger robustness.
8. The method for automatically identifying the indication of a float flowmeter according to claim 7, characterized in that: For the target detection network model used for bounding box positioning, a plurality of float flowmeter image data sets of different specifications are added to the training, and the large network model finally obtained can adapt to the bounding box recognition of float flowmeters of different specifications.
