1. A method of identifying a pointer meter reading, the method comprising:

acquiring the pointer instrument image;

inputting the pointer instrument image into a depth neural network model to obtain a target characteristic image;

carrying out contour recognition processing on the target characteristic image to obtain an environment contour recognition result and a reading contour recognition result;

judging whether the environment contour recognition result meets a preset environment contour condition or not;

and if the environment contour recognition result meets the preset environment contour condition, calculating the reading of the pointer instrument according to the reading contour recognition result.

2. The method for recognizing the reading of the pointer instrument according to claim 1, wherein the environment contour recognition result includes an outer contour of a dial, and the step of performing the contour recognition processing on the target feature image to obtain the environment contour recognition result and the reading contour recognition result includes:

judging whether the outer contour of the dial plate meets a preset dial plate shape or not, and judging whether the outer contour size of the dial plate meets a first preset size or not;

and when the dial outer contour meets the preset dial shape and the dial outer contour size meets the first preset size, calculating the reading of the pointer instrument according to the reading contour recognition result.

3. The method for recognizing the reading of the pointer instrument according to claim 1, wherein the environment contour recognition result includes a dial outer contour and an oil-air boundary, and the step of performing the contour recognition processing on the target feature image to obtain the environment contour recognition result and the reading contour recognition result includes:

acquiring the distance between the oil air boundary and the outer contour of the dial plate;

judging whether the distance is not greater than a second preset distance threshold value or not, and judging whether the length of the oil-liquid air boundary is not greater than a second preset size or not;

and when the distance is not greater than the second preset distance threshold and the length is not greater than the second preset size, calculating the reading of the pointer instrument according to the reading profile identification result.

4. The method of claim 1, wherein the reading profile recognition result includes a pointer and a scale, and the step of calculating the pointer meter reading from the reading profile recognition result includes:

judging whether the axis of the pointer is superposed with the circle center of the scale;

and when the axis of the pointer coincides with the circle center of the scale, the reading of the pointer instrument is calculated according to the pointer and the scale.

5. The method of claim 4, wherein the step of calculating the pointer meter reading from the pointer and the scale comprises:

acquiring a first connecting line between zero scale in the scales and the axis of the pointer and a second connecting line between the position pointed by the pointer and the axis of the pointer;

determining an included angle between the first connecting line and the second connecting line;

and calculating to obtain the reading of the pointer instrument by utilizing the ratio of the included angle to a preset included angle, wherein the preset included angle is the included angle between zero scale and the maximum scale in the scales of the pointer instrument.

6. The method of identifying pointer meter readings of claim 1 wherein said step of obtaining said pointer meter image comprises:

acquiring a pointer instrument image, wherein the pointer instrument image comprises deflection data;

and carrying out image correction processing on the original image according to the deflection data to obtain an image of the pointer instrument.

7. The method of claim 6, wherein the step of performing image correction processing on the raw image according to the deflection data to obtain an image of the pointer instrument comprises:

determining equation parameters of matrix perspective transformation according to the image deflection data to obtain a matrix perspective transformation equation;

processing the pointer instrument image according to the matrix perspective transformation equation to obtain an orthographic shot image;

and carrying out image preprocessing on the front-view shot image to obtain a pointer instrument image.

8. An identification system for pointer meter readings, the identification system comprising:

the acquisition unit is used for acquiring the pointer instrument image;

the network model unit is used for inputting the pointer instrument image into a depth neural network model to obtain a target characteristic image;

the contour recognition unit is used for carrying out contour recognition processing on the target characteristic image to obtain an environment contour recognition result and a reading contour recognition result;

the judging unit is used for judging whether the environment contour recognition result meets a preset environment contour condition or not;

and the calculation unit is used for calculating the reading of the pointer instrument according to the reading contour recognition result if the environment contour recognition result meets the preset environment contour condition.

9. A terminal device, comprising a memory and a processor coupled to the memory;

wherein the memory is for storing program data and the processor is for executing the program data to implement the method of identifying a pointer meter reading according to any one of claims 1 to 7.

10. A computer storage medium for storing program data which, when executed by a processor, is adapted to implement a method of identifying a pointer meter reading according to any one of claims 1 to 7.

Background

Current machine vision techniques aimed at character class recognition are well established. For example, the digital instrument reading is identified and processed by using a deep neural network model. However, the problem of low reading accuracy of the pointer instrument can exist in the process of directly calculating the reading of the pointer instrument based on the recognition image output by the deep neural network model.

Disclosure of Invention

The application provides a method, a system and equipment for identifying the reading of a pointer instrument and a computer storage medium, which aim to solve the problem that the reading of the pointer instrument is inaccurate in the prior art.

In order to solve the technical problem, the present application provides a method for identifying a reading of a pointer instrument, the method including:

acquiring the pointer instrument image;

inputting the pointer instrument image into a depth neural network model to obtain a target characteristic image;

carrying out contour recognition processing on the target characteristic image to obtain an environment contour recognition result and a reading contour recognition result;

judging whether the environment contour recognition result meets a preset environment contour condition or not;

and if the environment contour recognition result meets the preset environment contour condition, calculating the reading of the pointer instrument according to the reading contour recognition result.

In order to solve the above technical problem, the present application provides an identification system for reading of a pointer instrument, the identification system including:

the acquisition unit is used for acquiring the pointer instrument image;

the network model unit is used for inputting the pointer instrument image into a depth neural network model to obtain a target characteristic image;

the contour recognition unit is used for carrying out contour recognition processing on the target characteristic image to obtain an environment contour recognition result and a reading contour recognition result;

the judging unit is used for judging whether the environment contour recognition result meets a preset environment contour condition or not;

and the calculation unit is used for calculating the reading of the pointer instrument according to the reading contour recognition result if the environment contour recognition result meets the preset environment contour condition.

To solve the above technical problem, the present application provides a terminal device, which includes a memory and a processor coupled to the memory;

the memory is used for storing program data and the processor is used for executing the program data to realize the identification method of the reading of the pointer instrument.

In order to solve the above technical problem, the present application further provides a computer storage medium for storing program data, which when executed by a processor, is used to implement the identification method of reading of a pointer instrument as described above.

According to the pointer instrument reading identification method, the target characteristic image output by the deep neural network model is subjected to contour identification processing to determine the environment contour identification result and the reading contour identification result, and when the environment contour identification result meets the preset environment contour condition, the reading contour identification result is used for calculating the pointer instrument reading, so that errors caused by the fact that the target characteristic image output by the deep neural network model is directly used for calculating the pointer instrument reading are avoided, and the accuracy of the pointer instrument reading is improved.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts. Wherein:

FIG. 1 is a schematic flow chart diagram of an embodiment of a method for identifying pointer meter readings provided herein;

FIG. 2 is a schematic flow chart of an embodiment after S103 in the method for identifying a reading of a pointer meter shown in FIG. 1;

FIG. 3 is a schematic flow chart of an embodiment after S103 in the method for identifying a reading of a pointer meter shown in FIG. 1;

FIG. 4 is a schematic flow chart of an embodiment of S105 of the method for identifying a reading of a pointer meter shown in FIG. 1;

FIG. 5 is a schematic flow chart diagram of an embodiment of S402 in the method for identifying pointer meter readings shown in FIG. 4;

FIG. 6 is a schematic flow chart of an embodiment of S101 in the method for identifying a reading of a pointer meter shown in FIG. 1;

FIG. 7 is a schematic diagram of an embodiment of an identification system provided herein;

fig. 8 is a schematic structural diagram of an embodiment of a terminal device provided in the present application;

FIG. 9 is a schematic structural diagram of an embodiment of a computer storage medium provided in the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

The present application provides a method for recognizing a reading of a pointer instrument, and particularly please refer to fig. 1, where fig. 1 is a schematic flowchart of an embodiment of the method for recognizing a reading of a pointer instrument provided in the present application. The identification method of the pointer meter reading in the embodiment can be applied to an identification device, and the identification device of the present application can be a server, a mobile device, or a system in which the server and the mobile device are matched with each other. Accordingly, each part, such as each unit, sub-unit, module, and sub-module, included in the mobile device may be all disposed in the server, may also be all disposed in the mobile device, and may also be disposed in the server and the mobile device, respectively.

Further, the server may be hardware or software. When the server is hardware, it may be implemented as a distributed server cluster formed by multiple servers, or may be implemented as a single server. When the server is software, it may be implemented as a plurality of software or software modules, for example, software or software modules for providing distributed servers, or as a single software or software module, and is not limited herein.

The method for identifying the reading of the pointer instrument in the embodiment specifically comprises the following steps:

s101: and acquiring an image of the pointer instrument.

The reading of the pointer instrument is obtained through image recognition. Specifically, the recognition device may capture a pointer instrument image through a camera mounted in front of the pointer instrument. And the pointer instrument image can also be obtained through a mobile terminal carried by a meter reading worker. The method for acquiring the pointer instrument image by the camera arranged in front of the pointer instrument is adopted, and the pointer instrument image can be acquired at preset time intervals. The method for acquiring the pointer instrument image by the mobile terminal carried by the meter reading staff is adopted, and the pointer instrument image can be acquired when the meter reading staff arrives at the position of the pointer instrument every time.

In a specific embodiment, when the mobile terminal carried by the meter reading staff is used for acquiring the pointer instrument image, the acquired pointer instrument image is within a preset deflection angle range on the front surface of the pointer instrument, so that the problem that the pointer instrument reading identification is inaccurate due to the fact that the deflection angle of the pointer instrument image is too large is avoided. Wherein, the preset deflection angle range is within 20 degrees of the front surface of the pointer instrument.

Furthermore, in consideration of the installation environment of the pointer instrument, in order to improve the acquisition efficiency of the image of the pointer instrument, a shooting frame can be arranged according to the outline of the pointer instrument, so that a meter reading worker can conveniently and quickly frame the pointer instrument to obtain the image of the pointer instrument with a declination angle meeting the preset condition.

S102: and inputting the pointer instrument image into the depth neural network model to obtain a target characteristic image.

In order to identify the reading of the pointer instrument, the identification device of the embodiment inputs the image of the pointer instrument into the deep neural network model. Specifically, the recognition device inputs the target image into the deep neural network model to obtain a target characteristic image.

Wherein, the deep neural network model is a trained network model. In a specific embodiment, the recognition device can utilize a large number of training images and annotation images to train the deep neural network model to obtain a training characteristic image, calculate a loss function between the training characteristic image and a target characteristic image, and train the deep neural network model with the loss function becoming smaller as a target to obtain the deep neural network model meeting the requirements.

In a specific embodiment, the identification means may input the pointer instrument image into the first sub-network model, resulting in a multi-dimensional sub-feature image. And inputting the multi-dimensional sub-feature image into a second sub-network model to obtain a target feature image.

S103: and carrying out contour recognition processing on the target characteristic image to obtain an environment contour recognition result and a reading contour recognition result.

The high difficulty and inaccuracy of the reading of the pointer instrument in the prior art are considered by directly utilizing the target feature image output by the deep neural network model. The identification device of the embodiment considers the identification of the environment contour and the reading contour of the pointer instrument in the target characteristic image, so that the reading contour identification result is used for calculating the reading of the pointer instrument when the environment contour meets the preset environment contour condition. Specifically, the recognition device needs to perform contour recognition processing on the target feature image to obtain an environment contour recognition result and a reading contour recognition result. The environment contour recognition result can comprise the outer contour of the dial, the inner contour, an oil-air boundary and the like. The reading profile recognition result can comprise scales, pointers and the like.

In a specific embodiment, the identification device may perform contour identification processing on the target feature image by using a contour identification technology to obtain an environment contour identification result and a reading contour identification result. The contour identification technology can be opencv contour identification technology.

S104: and judging whether the environment contour recognition result meets the preset environment contour condition.

In order to improve the reading accuracy of the pointer instrument, the pointer of the pointer instrument is considered to be positioned in the contour of the dial plate. The recognition apparatus of this embodiment executes S105 when the environment contour recognition result satisfies the preset environment contour condition by determining whether the environment contour recognition result satisfies the preset environment contour condition. Otherwise, the environment contour recognition result and the reading contour recognition result are considered to be wrong.

S105: the pointer meter reading is calculated from the reading profile recognition result.

When the recognition device judges that the environment contour recognition result meets the preset environment contour condition, the reading contour recognition result is used for calculating the reading of the pointer instrument.

In the scheme, the recognition device carries out contour recognition processing on the target characteristic image output by the deep neural network model to determine an environment contour recognition result and a reading contour recognition result, and when the environment contour recognition result meets a preset environment contour condition, the reading contour recognition result is used for calculating the reading of the pointer instrument, so that errors caused by the fact that the reading of the pointer instrument is calculated by directly using the target characteristic image output by the deep neural network model are avoided, and the reading accuracy of the pointer instrument is improved.

Referring to fig. 2, fig. 2 is a schematic flowchart of the embodiment after S103 in the method for identifying reading of a pointer meter shown in fig. 1. On the basis of the above embodiment, the following steps are further included after S103:

s201: and judging whether the outer contour of the dial plate meets the preset dial plate shape or not, and judging whether the outer contour size of the dial plate meets a first preset size or not.

In order to improve the reading accuracy of the pointer instrument, the identification device needs to determine whether the environment contour identification result meets the preset environment contour condition. At this time, the environment contour recognition result includes the dial outer contour. Specifically, the identification device determines whether the dial outer contour meets a preset dial shape and determines whether the dial outer contour size meets a first preset size, and executes S202 when the dial outer contour meets the preset dial shape and the dial outer contour size meets the first preset size. When the dial plate outer contour does not meet the preset dial plate shape and the dial plate outer contour size does not meet the first preset size, or the dial plate outer contour meets the preset dial plate shape and the dial plate outer contour size does not meet the first preset size, and the dial plate outer contour does not meet the preset dial plate shape and the dial plate outer contour size meets the first preset size, the dial plate outer contour error identified by the contour identification technology is shown, and accurate pointer instrument reading can not be obtained based on the corresponding reading contour identification result.

S202: and calculating the reading of the pointer instrument according to the reading profile identification result.

The pointer meter reading is calculated based on the reading profile recognition result.

In the scheme, the identification device determines the accuracy of the dial outer contour identified by the contour identification technology by judging whether the dial outer contour meets the preset dial shape and judging whether the dial outer contour size meets the first preset size, so that the accuracy of the pointer instrument reading obtained by calculating based on the dial outer contour is improved.

Referring to fig. 3, fig. 3 is a schematic flowchart of the embodiment after S103 in the method for identifying reading of a pointer meter shown in fig. 1. On the basis of the above embodiment, the following steps are further included after S103:

s301: and acquiring the distance between the oil air boundary and the outer contour of the dial plate.

It is considered that an oil-air boundary may exist in the pointer reading, and the oil-air boundary has a great influence on the identification of the pointer reading. For this purpose, the identification device takes into account the influence of the oil-air boundary on the reading of the pointer instrument.

Specifically, the environment contour recognition result of the present embodiment includes a dial outer contour and an oil-air boundary. In order to avoid the influence of the oil-air boundary line on the reading accuracy of the pointer instrument, the recognition device obtains the distance between the oil-air boundary line and the outer contour of the dial plate. The distance between the oil-liquid air boundary and the outer contour of the dial is an inscribed distance.

S302: and judging whether the distance is not greater than a second preset distance threshold value or not, and judging whether the length of the oil-liquid-air boundary is not greater than a second preset size or not.

In order to determine the accuracy of the oil-air boundary and the dial outer contour identified by the contour identification technology, the identification device determines whether the inscribed distance between the oil-air boundary and the dial outer contour is not greater than a second preset distance threshold, determines whether the length of the oil-air boundary is not greater than a second preset size, and executes S303 when the distance is not greater than the second preset distance threshold and the length is not greater than the second preset size. When the distance is not greater than a second preset distance threshold but the length is smaller than a second preset size, or the distance is smaller than the second preset distance threshold but the length is not greater than the second preset size, and the distance is smaller than the second preset distance threshold and the length is smaller than the second preset size, it is determined that the oil-air boundary line or the dial outer contour identified by the contour identification technology is wrong, and accurate pointer instrument reading cannot be obtained based on the corresponding reading contour identification result.

S303: and calculating the reading of the pointer instrument according to the reading profile identification result.

The pointer meter reading is calculated based on the reading profile recognition result.

According to the scheme, the recognition device determines the accuracy of the oil-air boundary and the dial outer contour recognized by the recognition contour technology by acquiring the distance between the oil-air boundary and the dial outer contour, judging whether the distance is not greater than a second preset distance threshold value or not, and judging whether the length of the oil-air boundary is not greater than a second preset size or not, so that the accuracy of the pointer instrument reading obtained by calculation based on the oil-air boundary and the dial outer contour is improved.

Referring to fig. 4, fig. 4 is a flowchart illustrating an embodiment of the step S105 in the method for identifying reading of a pointer meter shown in fig. 1. On the basis of the above embodiment, S105 further includes the steps of:

s401: and judging whether the axis of the pointer is coincident with the circle center of the scale.

In order to avoid inaccuracy of the reading of the pointer instrument calculated based on the reading profile recognition result, the recognition device of the present embodiment judges the accuracy of the reading profile recognition result. The reading outline recognition result comprises a pointer and scales.

In an embodiment, the identification device determines whether the axis of the pointer coincides with the center of the circle where the scale is located, and executes S402 when the axis of the pointer coincides with the center of the circle where the scale is located. Otherwise, the error of the pointer and scale identified by the contour identification technology is indicated.

S402: and calculating the reading of the pointer instrument according to the pointer and the scales.

Optionally, this embodiment may adopt the embodiment of fig. 5 to implement S402, which specifically includes S501 to S503:

s501: and acquiring a first connecting line between the zero scale in the scales and the axis of the pointer and a second connecting line between the position pointed by the pointer and the axis of the pointer.

Based on the acquired pointer and scale, the recognition device acquires a first line between the zero scale in the scale and the axis of the pointer, and a second line between the position pointed by the pointer and the axis of the pointer.

S502: an angle between the first line and the second line is determined.

The identification device determines an included angle between the first connecting line and the second connecting line by using the first connecting line and the second connecting line.

S503: and calculating by using the ratio of the included angle to the preset included angle to obtain the reading of the pointer instrument.

Since the total reading of the scale is determined, the identification device calculates the reading of the pointer instrument by using the ratio of the included angle between the first connecting line and the second connecting line to the preset included angle. In particular, the identification means calculates the pointer meter reading using the product between the ratio and the total reading. The preset included angle is an included angle between zero scale and the maximum scale in the scales of the pointer instrument.

In the scheme, the identification device determines the accuracy of the pointer and the scales identified by the contour identification technology by judging whether the axis coincides with the circle center of the scales, so that the reading of the pointer instrument is calculated based on the pointer and the scales to obtain the reading of the pointer instrument, and the accuracy of the reading of the pointer instrument is improved.

Continuing to refer to fig. 6, fig. 6 is a flowchart illustrating an embodiment of S101 in the method for identifying reading of a pointer meter shown in fig. 1. On the basis of the above embodiment, S101 further includes the steps of:

s601: and determining equation parameters of matrix perspective transformation according to the image deflection data to obtain a matrix perspective transformation equation.

In consideration of the great influence of the image quality of the pointer instrument on the reading identification of the pointer instrument, the problem that the quality of the obtained image of the pointer instrument is poor due to the problems of ambient light of the pointer instrument or the angle of shooting the image of the pointer instrument is solved, and therefore the reading accuracy of the pointer instrument identified by the image of the pointer instrument is low. The recognition device determines an image conversion relation parameter between the pointer instrument image and the target image according to the deflection data of the pointer instrument image. Specifically, the identification device takes deflection data of the pointer instrument image as a matrix perspective transformation equation parameter to obtain a matrix perspective transformation equation.

In a particular embodiment, the identification device may utilize a gyroscope to obtain deflection data for the pointer instrument image.

S602: and processing the original image according to a matrix perspective transformation equation to obtain an orthographic shot image.

Based on the matrix perspective transformation equation acquired in S601, the recognition apparatus obtains an orthographic view taken image corresponding to the pointer instrument image by calculation using the matrix perspective transformation equation.

S603: and carrying out image preprocessing on the orthographic shot image to obtain a pointer instrument image.

The shooting angle of the front-view shot image meets the requirement of the target image, but the quality problem of the target image caused by shooting ambient light or other reasons exists. Therefore, the recognition device carries out image preprocessing on the orthographic shot image to obtain a pointer instrument image.

In a specific embodiment, the identification device can obtain the shot image after the gray level is removed by carrying out gray level processing on the front-view shot image; carrying out median filtering processing on the shot image subjected to the gray level removal to obtain a filtered shot image; and smoothing the filtered shot image based on the filtered shot image to obtain a pointer instrument image.

In the scheme, the identification device obtains the matrix perspective transformation equation by using the image deflection data as the matrix perspective transformation equation parameters, obtains the front-view shot image corresponding to the pointer instrument image according to the matrix perspective transformation equation, and performs image preprocessing on the front-view shot image to obtain the target image, so that the reading accuracy of the pointer instrument is improved, and the problem of low reading accuracy of the pointer instrument due to poor image quality of the pointer instrument is solved.

To implement the method for recognizing the reading of the pointer instrument in the foregoing embodiment, the present application provides a recognition system, and specifically refer to fig. 7, where fig. 7 is a schematic structural diagram of an embodiment of the recognition system provided in the present application.

The recognition system 700 includes an obtaining unit 71, a network model unit 72, an outline recognition unit 73, a judgment unit 74, and a calculation unit 75.

Specifically, the recognition system 700 includes: an acquisition unit 71 for acquiring a pointer instrument image.

And the network model unit 72 is configured to input the pointer instrument image into the deep neural network model to obtain a target feature image.

And the contour recognition unit 73 is used for performing contour recognition processing on the target feature image to obtain an environment contour recognition result and a reading contour recognition result.

And a judging unit 74, configured to judge whether the environment contour recognition result satisfies a preset environment contour condition.

And a calculating unit 75, configured to calculate, if the environment contour recognition result meets a preset environment contour condition, a reading of the pointer instrument according to the reading contour recognition result.

To implement the method for recognizing the reading of the pointer instrument in the foregoing embodiment, the present application provides a terminal device, and specifically refer to fig. 8, where fig. 8 is a schematic structural diagram of an embodiment of the terminal device provided in the present application.

The terminal device 800 comprises a memory 81 and a processor 82, wherein the memory 81 and the processor 82 are coupled.

The memory 81 is used for storing program data and the processor 82 is used for executing the program data to realize the identification method of the reading of the pointer meter of the above embodiment.

In the present embodiment, the processor 82 may also be referred to as a CPU (Central Processing Unit). The processor 82 may be an integrated circuit chip having signal processing capabilities. The processor 82 may also be a general purpose processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components. A general purpose processor may be a microprocessor or the processor 82 may be any conventional processor or the like.

The present application further provides a computer storage medium 900, as shown in fig. 9, the computer storage medium 900 is used for storing program data 91, and the program data 91, when executed by a processor, is used for implementing the method for identifying a reading of a pointer meter as described in the method embodiment of the present application.

The method involved in the embodiment of the identification method for the reading of the pointer meter is realized in the form of a software functional unit, and can be stored in a device, such as a computer readable storage medium, when the method is sold or used as a stand-alone product. Based on such understanding, the technical solution of the present application may be substantially implemented or contributed by the prior art, or all or part of the technical solution may be embodied in a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) or a processor (processor) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.

The above description is only for the purpose of illustrating embodiments of the present application and is not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings of the present application or are directly or indirectly applied to other related technical fields, are also included in the scope of the present application.