Intelligent identification method and system for industrial instrument

Abstract

The invention provides an intelligent identification method and system for an industrial instrument, and particularly relates to the field of intelligent identification of the industrial instrument, and the specific technical scheme is as follows: S1, obtaining and preprocessing an industrial instrument meter picture subjected to video monitoring to obtain a grayscale picture; s2, performing feature extraction on an outer dial in the gray scale picture to obtain a scale line annular belt region, positioning a reading region, and performing binarization processing on the scale line annular belt region to obtain a main scale line; and S3, marking an inner dial plate in the gray scale picture based on a Hough straight line to obtain a pointer feature, calculating a pointer slope based on a pointer feature numerical value to obtain a pointer skewness degree, and combining the pointer skewness degree and the main scale line to obtain an industrial instrument reading. According to the invention, through intelligent identification of pointer reading, the problems of extremely high labor cost and low reading efficiency and precision are improved, and the instrument pointer identification accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent identification of industrial instruments, and in particular to an intelligent identification method and system for industrial instruments. Background Art

[0002] Industrial instruments are widely used in various fields of industry. The data collected by instruments can provide important basis for technicians to master various information in the production process. However, human eye observation is easily interfered by various subjective factors, which often results in low accuracy and poor stability of the instrument reading. In addition, due to the fatigue of human eyes, it is easy to ignore the sudden change of the instrument during operation.

[0003] In industrial scenarios, non-intelligent instruments are widely used, and these instruments do not have data transmission functions. Since traditional instruments such as pointer instruments include dials, pointers, scales and other parts, and their structures are more complex than digital instruments, there is relatively little research on automatic identification of pointer instruments. Currently, most inspections are carried out manually, which has extremely high labor costs and low work efficiency and accuracy.

[0004] Therefore, it is urgent to develop a solution to solve the above problems. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent identification method and system for industrial instruments, so as to improve the problems of extremely high labor cost, low reading efficiency and accuracy, and improve the accuracy of instrument pointer identification.

[0006] The present invention provides an intelligent identification method and system for industrial instruments using the following technical solutions:

[0007] In a first aspect, an intelligent identification method for industrial instruments includes: acquiring and preprocessing an industrial instrument meter image of a video surveillance to obtain a grayscale image;

[0008] Feature extraction of the outer dial in the grayscale image to obtain a scale ring area, positioning the reading area and binarizing the scale ring area to obtain a main scale line;

[0009] The inner dial in the grayscale image is marked based on the Hough straight line to obtain the pointer feature, the pointer slope is calculated based on the pointer feature value to obtain the pointer deflection, and the industrial instrument reading is obtained by combining the pointer deflection and the main scale line.

[0010] The beneficial effects of an industrial instrument intelligent identification method provided by the present invention are as follows: the industrial instrument intelligent identification method provided by the present invention obtains meter images through a video monitoring system, performs artificial intelligence analysis on the images, and reads pointer parameters; digital technology is used to effectively improve labor costs, work efficiency, and accuracy, and intelligent inspections of industrial scenes are achieved by applying image recognition technology, thereby reducing the operation monitoring pressure of front-line staff and liberating a large amount of human resources.

[0011] Optionally, obtaining and preprocessing the video surveillance industrial instrument meter image to obtain a grayscale image includes:

[0012] Perform Gaussian filtering and blurring based on the original image;

[0013] Grayscale processing is used to convert the image information of the three channels of red, green and blue of the original image into the grayscale information of a single channel to obtain a grayscale image.

[0014] Optionally, feature extraction of a dial outline in a grayscale image to obtain a scale ring area includes:

[0015] Get a circular dial based on Hough circle detection;

[0016] The circular area in the middle of the hollow circular dial is the scale ring area.

[0017] Optionally, before locating the reading area and binarizing the scale line band area to obtain the main scale lines, the scale line band area is subjected to image inversion processing and image corrosion processing.

[0018] Optionally, the pointer feature is obtained based on the inner dial in the grayscale image identified by the Hough straight line, including:

[0019] The inner dial image is grayscaled and blurred, the image is inverted and the inner dial is corroded, the corroded image is thresholded, and the pointer is identified and marked based on Hough line detection to obtain the pointer feature.

[0020] Optionally, calculating the pointer slope based on the pointer characteristic value to obtain the pointer deflection includes:

[0021] Get the coefficient of the pointer in the polar coordinate system;

[0022] The coordinates of the two end points are obtained based on the numerical calculation of the coefficients;

[0023] Compare the horizontal and vertical coordinate values ​​of the two end points. If the vertical coordinate values ​​are equal, the pointer deflection is a right angle, and if the horizontal coordinate values ​​are equal, the pointer deflection is a straight angle.

[0024] When neither of the two conditions is satisfied, the pointer skewness is calculated based on the skewness calculation formula.

[0025] Optionally, when obtaining the industrial instrument reading in combination with the pointer deflection and the main scale line, it includes:

[0026] Perform character image recognition on the main scale lines to obtain the unit scale value and the deflection range of the dial pointer;

[0027] The industrial instrument reading is calculated based on the unit scale value and the dial pointer deflection range combined with the pointer deflection value.

[0028] Optionally, when applied to a service platform in a substation, it includes:

[0029] Build a wireless LAN in the substation to achieve network coverage within the substation;

[0030] Install wireless cameras to monitor power terminals such as meters, switch position indicators, etc.

[0031] Collect power terminal images regularly and transmit them to the edge IoT agent module through the network;

[0032] Apply the intelligent recognition method of industrial instruments in the edge IoT agent module to perform intelligent analysis and obtain image recognition results;

[0033] The substation meter pointer, shape and switch position recognition application is connected to the substation business system, and the image recognition results are transmitted to the substation business system to issue abnormal alarms and joint disposal.

[0034] In a second aspect, an intelligent identification system for industrial instruments includes:

[0035] An image collection module is used to obtain and pre-process images of industrial instruments and meters monitored by video to obtain grayscale images;

[0036] A feature extraction module is used to extract features of the outer dial in the grayscale image to obtain a scale ring area, locate the reading area and binarize the scale ring area to obtain a main scale line;

[0037] The reading module obtains the pointer feature based on the inner dial in the grayscale image marked by the Hough straight line, calculates the pointer slope based on the pointer feature value to obtain the pointer deflection, and obtains the industrial instrument reading by combining the pointer deflection and the main scale line.

[0038] The beneficial effects of the second aspect can refer to the description of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a flow chart of the overall steps of an industrial instrument intelligent identification method provided by the present invention;

[0040] Figure 2It is a schematic diagram of the application implementation of an industrial instrument intelligent identification method provided by the present invention. DETAILED DESCRIPTION

[0041] 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. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Unless otherwise defined, the technical terms or scientific terms used herein should be understood by people with general skills in the field to which the present invention belongs. "Including" and similar words used in this article mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects.

[0042] The embodiment of the present invention provides an industrial instrument intelligent identification method, see Figure 1 , the specific steps include:

[0043] Step S1, obtaining and preprocessing the industrial instrument meter images of the video surveillance to obtain grayscale images;

[0044] Step S2, feature extraction of the outer dial in the grayscale image to obtain a scale ring area, positioning the reading area and binarizing the scale ring area to obtain a main scale line;

[0045] Step S3, obtaining pointer features based on the inner dial in the grayscale image using the Hough straight line, calculating the pointer slope based on the pointer feature value to obtain the pointer deflection, and obtaining the industrial instrument reading by combining the pointer deflection and the main scale line.

[0046] In some embodiments, when executing step S1, obtaining and preprocessing the video surveillance industrial instrument meter image to obtain a grayscale image includes:

[0047] Perform Gaussian filtering and blurring based on the original image;

[0048] Grayscale processing is used to convert the image information of the three channels of red, green and blue of the original image into the grayscale information of a single channel to obtain a grayscale image.

[0049] Specifically, the use of Gaussian filtering can remove random noise in the image, especially Gaussian noise, thereby improving the quality of the image; through Gaussian filtering, small noise points in the image will be replaced by the weighted average of the surrounding pixels, making the image smoother and providing better input for subsequent feature extraction.

[0050] Specifically, the images of industrial instruments and meters obtained through video surveillance are color images in the BGR color mode. By adjusting the strength of the three primary colors, the images can display various colors. Grayscale images are single-channel images, that is, black and white images. The purpose of grayscale preprocessing the template image is to convert the image information of the three channels of red, green and blue into the grayscale information of a single channel, which is equivalent to reducing the amount of calculation by two-thirds and speeding up the calculation efficiency of the template matching algorithm.

[0051] Furthermore, the grayscale processing is to remove the color features of the image without losing the edge features of the image itself and keep the grayscale distribution uniform.

[0052] In some embodiments, when executing step S2, feature extraction of the dial outline in the grayscale image to obtain the scale ring area includes:

[0053] Get a circular dial based on Hough circle detection;

[0054] The circular area in the middle of the hollow circular dial is the scale ring area.

[0055] Specifically, Hough circle detection is used to detect and locate circular targets in images and identify the outline of the outer dial. It can still maintain a high recognition accuracy when processing images with complex backgrounds and noise.

[0056] In some embodiments, before locating the reading area and binarizing the scale line band area to obtain the main scale lines when executing step S2, the scale line band area is subjected to image inversion processing and image corrosion processing.

[0057] Specifically, by performing a negation operation on each pixel in the scale line ring area, the brightness of the image is reversed to improve the visual effect of the image. In particular, when selecting connected areas of a binary image, if the background color and the object color are opposite, the processing process can be simplified by performing a negation operation.

[0058] Specifically, the corrosion process is used to eliminate small objects and noise in the image to make the target area smoother; by performing logical operations on the pixels in the image, the pixels that do not match the structural elements are gradually corroded, reduced or deleted.

[0059] In some embodiments, when executing step S3, obtaining the pointer feature by identifying the inner dial in the grayscale image based on the Hough straight line includes:

[0060] The inner dial image is grayscaled and blurred, the image is inverted and the inner dial is corroded, the corroded image is thresholded, and the pointer is identified and marked based on Hough line detection to obtain the pointer feature.

[0061] Specifically, the image is segmented according to a set threshold value, the pixel value greater than the preset threshold value is the target area, and the pixel value less than the preset threshold value is the background area.

[0062] In some embodiments, when executing step S3, calculating the pointer slope based on the pointer characteristic value to obtain the pointer deflection includes:

[0063] Get the coefficient of the pointer in the polar coordinate system;

[0064] The coordinates of the two end points are obtained based on the numerical calculation of the coefficients;

[0065] Compare the horizontal and vertical coordinate values ​​of the two end points. If the vertical coordinate values ​​are equal, the pointer deflection is a right angle, and if the horizontal coordinate values ​​are equal, the pointer deflection is a straight angle.

[0066] When neither of the two conditions is satisfied, the pointer skewness is calculated based on the skewness calculation formula.

[0067] Specifically, the skewness calculation formula is:

[0068] result = 90 + tan(k)*δ;

[0069] result is the pointer deflection, k is the pointer slope, and δ is the exponential constant.

[0070] In some embodiments, when executing step S3, combining the pointer deflection and the main scale line to obtain the industrial instrument reading includes:

[0071] Perform character image recognition on the main scale lines to obtain the unit scale value and the deflection range of the dial pointer;

[0072] Calculates industrial meter readings based on unit scale values ​​and dial pointer deflection range combined with pointer deflection values.

[0073] Specifically, the reading calculation formula is:

[0074] meter=result / range*unit;

[0075] meter is the reading, result is the pointer deflection, range is the dial pointer deflection range, and unit is the unit scale value.

[0076] In some embodiments, when executing step S3, when applied to the service platform in the substation, it includes:

[0077] Build a wireless LAN in the substation to achieve network coverage within the substation;

[0078] Install wireless cameras to monitor power terminals such as meters, switch position indicators, etc.

[0079] Collect power terminal images regularly and transmit them to the edge IoT agent module through the network;

[0080] Apply the intelligent recognition method of industrial instruments in the edge IoT agent module to perform intelligent analysis and obtain image recognition results;

[0081] The substation meter pointer, shape and switch position recognition application is connected to the substation business system, and the image recognition results are transmitted to the substation business system to issue abnormal alarms and joint disposal.

[0082] For details, see Figure 2 The wireless camera can take pictures or videos of power terminals such as meters and switch indicators at regular intervals and upload them to the edge IoT agent module through the WAPI wireless network. The edge IoT agent device has edge computing capabilities and can realize the application of edge-side artificial intelligence technology. The algorithm of the present invention is deployed in the edge IoT agent device, and through real-time monitoring and analysis of video streams, key images are captured. The captured images are identified and inferred by the algorithm, and the images are classified, monitored, segmented, and identified, and then the edge analysis data is uploaded to the power business platform.

[0083] The embodiment of the present invention provides an intelligent identification system for industrial instruments, which specifically includes:

[0084] An image collection module is used to obtain and pre-process images of industrial instruments and meters monitored by video to obtain grayscale images;

[0085] A feature extraction module is used to extract features of the outer dial in the grayscale image to obtain a scale ring area, locate the reading area and binarize the scale ring area to obtain a main scale line;

[0086] The reading module obtains the pointer feature based on the inner dial in the grayscale image marked by the Hough straight line, calculates the pointer slope based on the pointer feature value to obtain the pointer deflection, and obtains the industrial instrument reading by combining the pointer deflection and the main scale line.

[0087] Although the embodiments of the present invention are described in detail above, it is obvious to those skilled in the art that various modifications and variations can be made to these embodiments. However, it should be understood that such modifications and variations are within the scope and spirit of the present invention as described in the claims. Moreover, the present invention described herein may have other embodiments and may be implemented or realized in a variety of ways.

Claims

1. An intelligent identification method for industrial instruments, characterized in that: include: Acquire and preprocess the images of industrial instruments and meters under video surveillance to obtain grayscale images; Feature extraction of the outer dial in the grayscale image to obtain a scale ring area, positioning the reading area and binarizing the scale ring area to obtain a main scale line; The inner dial in the grayscale image is marked based on the Hough straight line to obtain the pointer feature, the pointer slope is calculated based on the pointer feature value to obtain the pointer deflection, and the industrial instrument reading is obtained by combining the pointer deflection and the main scale line.

2. The intelligent identification method of industrial instruments according to claim 1 is characterized in that: When obtaining and preprocessing the video surveillance industrial instrument meter image to obtain the grayscale image, it includes: Perform Gaussian filtering and blurring based on the original image; Grayscale processing is used to convert the image information of the three channels of red, green and blue of the original image into the grayscale information of a single channel to obtain a grayscale image.

3. The intelligent identification method of industrial instruments according to claim 1 is characterized in that: Feature extraction: The dial outline in the grayscale image is used to obtain the scale ring area, including: Get a circular dial based on Hough circle detection; The circular area in the middle of the hollow circular dial is the scale ring area.

4. The intelligent identification method of industrial instruments according to claim 1 is characterized in that: Before locating the reading area and binarizing the scale line ring area to obtain the main scale lines, the scale line ring area is subjected to image inversion processing and image corrosion processing.

5. The intelligent identification method of industrial instruments according to claim 1 is characterized in that: The pointer features are obtained by marking the inner dial in the grayscale image based on the Hough straight line, including: The inner dial image is grayscaled and blurred, the image is inverted and the inner dial is corroded, the corroded image is thresholded, and the pointer is identified and marked based on Hough line detection to obtain the pointer feature.

6. The intelligent identification method of industrial instruments according to claim 1 is characterized in that: Calculating the pointer slope based on the pointer characteristic value to obtain the pointer deflection includes: Get the coefficient of the pointer in the polar coordinate system; The coordinates of the two end points are obtained based on the numerical calculation of the coefficients; Compare the horizontal and vertical coordinate values ​​of the two end points. If the vertical coordinate values ​​are equal, the pointer deflection is a right angle, and if the horizontal coordinate values ​​are equal, the pointer deflection is a straight angle. When neither of the two conditions is satisfied, the pointer skewness is calculated based on the skewness calculation formula.

7. The intelligent identification method of industrial instruments according to claim 1, characterized in that: Combining the pointer deflection and the main scale line to obtain the industrial instrument reading includes: Perform character image recognition on the main scale lines to obtain the unit scale value and the deflection range of the dial pointer; The industrial instrument reading is calculated based on the unit scale value and the dial pointer deflection range combined with the pointer deflection value.

8. The intelligent identification method of industrial instruments according to claim 1, characterized in that: When applied to the business platform in the substation, it includes: Build a wireless LAN in the substation to achieve network coverage within the substation; Install wireless cameras to monitor power terminals such as meters, switch position indicators, etc. Collect power terminal images regularly and transmit them to the edge IoT agent module through the network; Apply the intelligent recognition method of industrial instruments in the edge IoT agent module to perform intelligent analysis and obtain image recognition results; The substation meter pointer, shape and switch position recognition application is connected to the substation business system, and the image recognition results are transmitted to the substation business system to issue abnormal alarms and joint disposal.

9. An intelligent identification system for industrial instruments, characterized in that: include: An image collection module is used to obtain and pre-process images of industrial instruments and meters monitored by video to obtain grayscale images; A feature extraction module is used to extract features of the outer dial in the grayscale image to obtain a scale ring area, locate the reading area and binarize the scale ring area to obtain a main scale line; The reading module obtains the pointer feature based on the inner dial in the grayscale image marked by the Hough straight line, calculates the pointer slope based on the pointer feature value to obtain the pointer deflection, and obtains the industrial instrument reading by combining the pointer deflection and the main scale line.

Images