Instrument information identification method and device based on image identification

Abstract

The invention relates to the technical field of industry, in particular to an instrument information identification method and device based on image identification. The method comprises the following steps: firstly, shooting an information display surface of instrument equipment to be identified into an information block diagram through shooting equipment; then, intercepting the information block diagram into a target block diagram with a fixed size; thirdly, performing instrument information identification based on the training image set, and outputting identification results such as a target block diagram, instrument digital reading information and a working state; and finally, correcting and identifying the target block diagram according to the combined model, and outputting accurate instrument information. The device comprises a first information acquisition unit, a second information acquisition unit, a combined model training unit and an information correction unit, and is used for acquiring, training and correcting instrument information and ensuring accuracy and reliability. The method and the device can be widely applied to instrument equipment identification and information acquisition in an industrial field, and the identification accuracy and the working efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the field of engineering computing technology, and in particular to an instrument information recognition method and device based on image recognition. Background Art

[0002] With the continuous improvement of industrial automation, the accurate identification and collection of instrumentation information on industrial sites is becoming increasingly important. Traditional instrumentation information recognition methods rely on manual observation and recording, which suffers from low information collection efficiency, prone to errors, and high costs. Therefore, an automated instrumentation information recognition method based on image recognition is needed to improve recognition accuracy and efficiency.

[0003] Object recognition is a key research and application area in artificial intelligence. With the continuous development of object recognition technology, it has been widely used in applications such as face recognition and object recognition. As object recognition technology continues to advance in various application areas, higher requirements are being placed on the target recognition performance achieved by this technology in certain application scenarios in many application areas.

[0004] There are already some instrument information recognition technologies based on image recognition on the market, but there are some problems: first, the recognition effect of instrument information frame diagrams of different sizes is poor; second, it is impossible to recognize multiple instruments at the same time; third, the accuracy and stability of the instrument information recognition results are required to be high;

[0005] For example, in a patented general switch state recognition method based on image processing, the SSIM algorithm can only recognize two states of switch instruments, that is, between the 0 and 1 states of the switch instrument through threshold control, and cannot recognize instruments with more complex data states (such as the numbers displayed in digital instruments);

[0006] Therefore, it is urgent to propose a new information recognition method based on image recognition for industrial field instrument equipment to solve the problems existing in the existing technology and improve the accuracy and reliability of instrument information collection. Summary of the Invention

[0007] In view of this, the present invention provides an instrument information recognition method and device based on image recognition. This solution can improve the accuracy, efficiency and stability of instrument information recognition, and provide strong support for industrial automation.

[0008] The technical solution adopted in the present invention is:

[0009] In a first aspect, the present invention provides a method for identifying instrument information based on image recognition, characterized in that the method comprises:

[0010] S1: Obtain the information block diagram of the instrument device to be identified;

[0011] S2: Cutting the information frame into a target frame of a preset size;

[0012] S3: Recognize the content of the target frame image based on the training image set and output the recognition result;

[0013] The recognition result includes the target block diagram, instrument digital reading information and instrument working status;

[0014] The preset size range of the target frame diagram is within the range of 224 mm*224 mm.

[0015] Preferably, the S1 includes:

[0016] S11: placing a photographing device on the information display surface of the instrument device to be identified;

[0017] S12: Adjusting the direction of the lens of the photographing device to be parallel to the information display surface of the instrument device to be identified;

[0018] S13: photographing the information display surface of the instrument device to be identified by the photographing device, and outputting an information frame diagram.

[0019] Preferably, before S2, the step further includes:

[0020] S021: establishing a plane rectangular coordinate system on the information block diagram with the lower left corner of the information block diagram as the origin;

[0021] S022: Identify the location of key information in the information block diagram;

[0022] S023: Enlarging the target frame diagram within a preset size range based on the location of the key information;

[0023] S024: intercepting the target block diagram.

[0024] Preferably, before S3, the step further includes:

[0025] S031: Collect image datasets of all instrumentation equipment in industrial sites in any state and perform inference training;

[0026] S032: Obtaining a combined model through the inference training;

[0027] S033: deploying the combined model to a server;

[0028] S034: connecting the server to a camera at the industrial site, and receiving real-time image data from the camera via the server;

[0029] Wherein, the combined model can simultaneously identify multiple instrument devices;

[0030] The server includes a local server and / or a cloud server.

[0031] Preferably, the S3 further includes:

[0032] S31: Use FasterRCNN to identify the target block diagram of the instrument equipment to be detected;

[0033] S32: Correcting the image information of the target frame image through the combined model;

[0034] S33: Recognizing instrument information on the corrected target block diagram through the combined model;

[0035] S34: Output recognition results;

[0036] The recognition result also includes the target frame Figure 4 The coordinates of the corner points on the plane rectangular coordinate system.

[0037] In a second aspect, an instrument information recognition device is provided, characterized by comprising:

[0038] The first information collection unit is used to collect image data sets of any state of all instrument equipment on the industrial site and make them into a set;

[0039] The second information collection unit is used to collect the display information and working status of all instruments and equipment on the site;

[0040] a combined model training unit, which performs recognition training on the information in the set generated by the first information acquisition unit, and compares the recognition training result with the final correct information corresponding to the recognition information;

[0041] The information correction unit performs information identification on the instrument equipment information collected by the second information collection unit through the combined model training unit, and outputs correct information to ensure the accuracy of the information output of the instrument equipment collected by the second information collection unit.

[0042] In summary, the beneficial effects of the present invention are as follows:

[0043] 1. Improve recognition accuracy: Using training image sets and combined models for recognition can accurately identify the information of instrument equipment and output correct digital readings and working status, avoiding the influence of human factors on the recognition results and improving recognition accuracy.

[0044] 2. Improve identification efficiency: The method is simple and clear, and information collection and identification are carried out quickly through an automated process, saving manpower and time costs, improving identification efficiency, and is suitable for information collection of a large number of instruments and equipment in industrial sites.

[0045] 3. Simultaneous identification of multiple instruments and equipment: Using combined models and technologies such as FasterRCNN, it can simultaneously identify multiple different instrument and equipment targets, not just a single instrument and equipment target, improving the flexibility and applicability of identification.

[0046] 4. Information correction function: The instrument information is corrected through the information correction unit to ensure the accuracy of the information and improve the accuracy and reliability of the information output; after rigorous image data training in the early stage, the model image recognition can ensure high accuracy.

[0047] 5. Wide range of industrial applications: Suitable for identification and collection of instrument equipment information at industrial sites, it can help enterprises achieve intelligent management and optimize production processes, with good application prospects and economic benefits; it can identify any state of instrument equipment (not limited to the two states of on and off).

[0048] In summary, the method and device of this patent have broad application prospects in the industrial field, can improve the accuracy, efficiency and stability of instrument information recognition, and provide strong support for industrial automation. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work, and these are all within the scope of protection of the present invention.

[0050] Figure 1 A schematic diagram of the overall working process of an instrument information identification method in Examples 1 and 2 of the present invention;

[0051] Figure 2 This is a flowchart of obtaining information about the instrument device to be identified in Examples 1 and 2 of the present invention;

[0052] Figure 3 This is a schematic diagram of the process of cutting the information block diagram into a target block diagram of a preset size in Embodiments 1 and 2 of the present invention;

[0053] Figure 4 This is a flow chart illustrating the process of recognizing the target block diagram content based on the training image set and outputting the recognition result in Embodiments 1 and 2 of the present invention;

[0054] Figure 5A schematic diagram of a process for identifying the target block diagram content based on a training image set and outputting the recognition result in embodiments 1 and 2 of the present invention;

[0055] Figure 6 This is a structural block diagram of a device based on an instrument information identification method in Examples 1 and 2 of the present invention; DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. It should be noted that, in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. In the description of the present invention, it should be understood that the orientation or position relationship indicated by the terms "center", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like is based on the orientation or position relationship shown in the drawings, and is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the present invention. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further limitations, elements defined by the phrase "comprising..." do not preclude the presence of additional identical elements in the process, method, article, or apparatus comprising the elements. The embodiments of the present invention and the features thereof may be combined with each other if there is no conflict, and all are within the scope of protection of the present invention.

[0057] Example 1

[0058] See Figures 1 to 6 Embodiment 1 of the present invention discloses a method for identifying instrument information based on image recognition, characterized in that the method includes:

[0059] S1: Obtain the information block diagram of the instrument device to be identified;

[0060] Since there are many types of instrument devices in actual use, each instrument device has a variety of display settings. First, the display surface of the instrument device to be identified must be locked. After confirming the display surface of the instrument device to be identified, it is photographed.

[0061] S2: Cutting the information frame into a target frame of a preset size;

[0062] The image of the display surface of the instrument device to be identified is captured. Since the entire captured image has a large shooting range and there are too many useless images, in order to more easily locate useful information, the captured image is captured, leaving only the useful information. The maximum size frame of the image to be captured is determined based on a large number of instrument devices in actual application. The size frame can completely cover the largest display surface of the instrument device in actual application, ensuring that all useful information can be captured.

[0063] S3: Recognize the content of the target frame image based on the training image set and output the recognition result;

[0064] Compare and identify the contents of the target block diagram with the image set trained in the combined training model, and output the contents;

[0065] The recognition results include the original target block diagram, digital readings on the instrument equipment, such as temperature, pressure, etc., as well as the working status of the instrument equipment, the switch status of the instrument equipment, whether the instrument equipment is operating normally, etc.

[0066] The preset size range of the target frame diagram is within the range of 224 mm*224 mm.

[0067] Preferably, the S1 includes:

[0068] S11: placing a photographing device on the information display surface of the instrument device to be identified;

[0069] S12: Adjusting the direction of the lens of the photographing device to be parallel to the information display surface of the instrument device to be identified;

[0070] S13: photographing the information display surface of the instrument device to be identified by the photographing device, and outputting an information frame diagram.

[0071] Preferably, before S2, the step further includes:

[0072] S021: Establishing a plane rectangular coordinate system on the information frame diagram with the lower left corner of the information frame diagram as the origin. A picture taken by a camera is generally rectangular. The length and width of the rectangle are used as coordinate axes, and the lower left corner of the picture is used as the origin of the coordinate system to construct the plane rectangular coordinate system. The unit of the plane rectangular coordinate system is millimeters.

[0073] S022: Identify the location of key information in the information block diagram;

[0074] Locate the information part in the captured image and intercept the relevant information in an overlay manner;

[0075] S023: Enlarging the target frame diagram within a preset size range based on the location of the key information;

[0076] Position the information part in the captured image and set a base point, which is automatically determined by the system. The center point or the lower left corner of the information part may be rotated as the base point for expansion to cover the information between the gates;

[0077] S024: intercepting the target block diagram.

[0078] Preferably, before S3, the step further includes:

[0079] S031: Collect image datasets of all instrumentation equipment in industrial sites in any state and perform inference training;

[0080] Since there are many instruments and equipment in actual applications, we first collect image data of instruments and equipment in actual applications or in any state, and try to identify the data. For unclear parts, we perform reasoning training to simulate them into definite data. Then, we compare the simulated data with the actual confirmed data to determine its accuracy, and revise the subsequent judgment until we can finally correctly identify the model of various instrument and equipment data.

[0081] S032: Obtaining a combined model through the inference training;

[0082] By performing the above-mentioned inference and recognition training on the image data collected in actual applications or in any state of existing instrument equipment, a combined model is obtained;

[0083] S033: deploying the combined model to a server;

[0084] The combined model is loaded into the server for use in identifying pictures actually taken on site;

[0085] S034: connecting the server to a camera at the industrial site, and receiving real-time image data from the camera via the server;

[0086] Wherein, the combined model can simultaneously identify multiple instrument devices;

[0087] The server includes a local server and / or a cloud server.

[0088] Preferably, the S3 further includes:

[0089] S31: Use FasterRCNN to identify the target block diagram of the instrument equipment to be detected;

[0090] Faster R-CNN is a unified object detection network that combines the RPN and Fast R-CNN modules. The first module is used to generate regions of interest, and the second module is used to determine the object type and perform bounding box regression. Therefore, it is also called a two-stage object detection network. The specific steps of the network operation are as follows:

[0091] 1. Input the image into the pre-trained feature extraction network to obtain a feature map.

[0092] 2. The RPN network generates a series of proposed regions that may contain the objects to be detected based on the feature map.

[0093] 3. According to the proposed area in the second step, the corresponding features are intercepted in the feature map, and finally the type of object is obtained and the anchor frame is corrected.

[0094] S32: Correcting the image information of the target frame image through the combined model;

[0095] Compare the proposed data with the actual confirmed data to determine its accuracy and calibrate the subsequent judgments. Finally, the model of various instrument equipment data can be correctly identified to obtain accurate data.

[0096] S33: Recognizing instrument information on the corrected target block diagram through the combined model;

[0097] S34: Output recognition results;

[0098] The recognition result also includes the target frame Figure 4 The coordinates of the corner points in the plane rectangular coordinate system; the detection frame and vertex coordinates are used to pass back to the system to analyze the lens focus point and return it to the camera

[0099] Example 3

[0100] See Figures 1 to 6 , provides an instrument information recognition device, characterized by comprising:

[0101] The first information collection unit is used to collect image data sets of any state of all instrument equipment on the industrial site and make them into a set;

[0102] The second information collection unit is used to collect the display information and working status of all instruments and equipment on the site;

[0103] a combined model training unit, which performs recognition training on the information in the set generated by the first information acquisition unit, and compares the recognition training result with the final correct information corresponding to the recognition information;

[0104] The information correction unit performs information identification on the instrument equipment information collected by the second information collection unit through the combined model training unit, and outputs correct information to ensure the accuracy of the information output of the instrument equipment collected by the second information collection unit.

[0105] Example 3

[0106] See Figures 1 to 6 , provides an instrument information identification method and device. In summary, the beneficial effects of the present invention are as follows:

[0107] 1. Improve recognition accuracy: Using training image sets and combined models for recognition can accurately identify the information of instrument equipment and output correct digital readings and working status, avoiding the influence of human factors on the recognition results and improving recognition accuracy.

[0108] 2. Improve identification efficiency: The method is simple and clear, and information collection and identification are carried out quickly through an automated process, saving manpower and time costs, improving identification efficiency, and is suitable for information collection of a large number of instruments and equipment in industrial sites.

[0109] 3. Simultaneous identification of multiple instruments and equipment: Using combined models and technologies such as FasterRCNN, it can simultaneously identify multiple different instrument and equipment targets, not just a single instrument and equipment target, improving the flexibility and applicability of identification.

[0110] 4. Information correction function: The instrument information is corrected through the information correction unit to ensure the accuracy of the information and improve the accuracy and reliability of the information output; after rigorous image data training in the early stage, the model image recognition can ensure high accuracy.

[0111] 5. Wide range of industrial applications: Suitable for identification and collection of instrument equipment information at industrial sites, it can help enterprises achieve intelligent management and optimize production processes, with good application prospects and economic benefits; it can identify any state of instrument equipment (not limited to the two states of on and off).

[0112] In summary, the method and device of this patent have broad application prospects in the industrial field, can improve the accuracy, efficiency and stability of instrument information recognition, and provide strong support for industrial automation.

[0113] It should be understood that the present invention is not limited to the specific configurations and processes described above and illustrated in the figures. For the sake of brevity, a detailed description of known methods is omitted. In the above embodiments, several specific steps are described and illustrated as examples. However, the method of the present invention is not limited to the specific steps described and illustrated. Those skilled in the art may make various changes, modifications, and additions, or change the order of the steps after understanding the spirit of the present invention.

[0114] The functional blocks shown in the above-described block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in unit, a function card or the like. When implemented in software, the elements of the present invention are programs or code segments that are used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link by a data signal carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROMs, flash memories, erasable ROMs (EROMs), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.

[0115] It should also be noted that the exemplary embodiments described herein describe methods or systems based on a series of steps or devices. However, the present invention is not limited to the order of the steps described above. In other words, the steps may be performed in the order described in the embodiments, or in a different order, or several steps may be performed simultaneously.

[0116] The above description is only a specific embodiment of the present invention. Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the protection scope of the present invention is not limited to this. Any person skilled in the art can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention.

Claims

1. A method for identifying instrument information based on image recognition, characterized in that: include: S1: Obtain the information block diagram of the instrument device to be identified; S2: Cutting the information frame into a target frame of a preset size; S3: Recognize the content of the target frame image based on the training image set and output the recognition result; The recognition result includes the target block diagram, instrument digital reading information and instrument working status; The preset size range of the target frame diagram is within the range of 224 mm*224 mm.

2. The instrument information recognition method according to claim 1, characterized in that: Said S1 comprises: S11: placing a photographing device on the information display surface of the instrument device to be identified; S12: Adjusting the direction of the lens of the photographing device to be parallel to the information display surface of the instrument device to be identified; S13: photographing the information display surface of the instrument device to be identified by the photographing device, and outputting an information frame diagram.

3. The instrument information recognition method according to claim 1, characterized in that: The S2 also includes: S021: establishing a plane rectangular coordinate system on the information block diagram with the lower left corner of the information block diagram as the origin; S022: Identify the location of key information in the information block diagram; S023: Enlarging the target frame diagram within a preset size range based on the location of the key information; S024: intercepting the target block diagram.

4. The instrument information recognition method according to claim 3, characterized in that: The S3 also includes: S031: Collect image datasets of all instrumentation equipment in industrial sites in any state and perform inference training; S032: Obtaining a combined model through the inference training; S033: deploying the combined model to a server; S034: connecting the server to a camera at the industrial site, and receiving real-time image data from the camera via the server; Wherein, the combined model can simultaneously identify multiple instrument devices; The server includes a local server and / or a cloud server.

5. The instrument information recognition method according to claim 4, characterized in that: Said S3 further comprises: S31: Use FasterRCNN to identify the target block diagram of the instrument equipment to be detected; S32: Correcting the image information of the target frame image through the combined model; S33: Recognizing instrument information on the corrected target block diagram through the combined model; S34: Output recognition results; The recognition result further includes the coordinates of the four corner points of the target frame in the plane rectangular coordinate system.

6. An instrument information recognition device, characterized in that: include: The first information collection unit is used to collect image data sets of any state of all instrument equipment on the industrial site and make them into a set; The second information collection unit is used to collect the display information and working status of all instruments and equipment on the site; a combined model training unit, which performs recognition training on the information in the set generated by the first information acquisition unit, and compares the recognition training result with the final correct information corresponding to the recognition information; The information correction unit performs information identification on the instrument equipment information collected by the second information collection unit through the combined model training unit, and outputs correct information to ensure the accuracy of the information output of the instrument equipment collected by the second information collection unit.

Images