CN113537068A - Image-based digital meter reading identification method, storage medium and corollary equipment - Google Patents
Image-based digital meter reading identification method, storage medium and corollary equipment Download PDFInfo
- Publication number
- CN113537068A CN113537068A CN202110810033.2A CN202110810033A CN113537068A CN 113537068 A CN113537068 A CN 113537068A CN 202110810033 A CN202110810033 A CN 202110810033A CN 113537068 A CN113537068 A CN 113537068A
- Authority
- CN
- China
- Prior art keywords
- image
- digital
- reading
- sample set
- digital meter
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention relates to a reading identification method, a storage medium and a matched device of a digital meter based on an image, wherein the identification method comprises the following steps: automatically generating a digital synthesis training sample set; acquiring a required detection recognition model based on the digital synthesis training sample set and the artificially labeled digital real training sample set; and automatically recognizing the reading of the image containing the digital meter by using the trained detection and recognition model, detecting an area containing one and only one complete reading, detecting the number from the area, and sequencing the number to obtain the reading of the digital meter. The identification method is based on a deep learning algorithm, can accurately identify the readings of various different types of digital meters, has concise steps, reduces the dependence on manual marking data, and can be quickly adapted to new digital meter types.
Description
Technical Field
The invention relates to the technical field of transformer substation monitoring equipment, in particular to a reading identification method, a storage medium and corollary equipment of a transformer substation digital meter based on images.
Background
There are a large number of meters in substations for monitoring the operating state of equipment, and many of them are digital meters (i.e. using numbers as reading display modes, including mechanical rotating disk type and liquid crystal type). Fig. 1 is a photograph of several digital meters used in a transformer substation, which play an important role in ensuring the normal operation of the transformer substation, and meanwhile, due to the special high-voltage environment of the transformer substation, a remote meter reading system based on a wireless network such as 5G cannot be used, so that manual inspection and meter reading are performed by relying on a large amount of manpower.
In recent years, the problems of manual inspection and meter reading are researched and solved from the perspective of image recognition, namely, a camera is used for shooting a dial plate of a digital meter, and the reading of the digital meter is automatically recognized and reported by using an image intelligent algorithm taking a deep learning technology as a core. The problems existing in the prior art include that on one hand, a template matching method is used for positioning the numbers, which is highly dependent on the accuracy of camera installation and high requirements on the quality of the shot images, and on the other hand, only a specific type of digital meter can be identified, and the application range is narrow.
Disclosure of Invention
The invention aims to provide a reading identification method, a storage medium and matched equipment of a digital meter based on an image aiming at a transformer substation environment.
In order to achieve the above object, a first embodiment of the present invention provides an image-based digital meter reading identification method, including
Automatically generating a digital synthesis training sample set;
acquiring a required detection recognition model based on the digital synthesis training sample set and the artificially labeled digital real training sample set;
the method comprises the steps of utilizing a trained detection and identification model to automatically identify readings of an image containing a digital meter, specifically adopting two detection and identification models to identify the readings of the digital meter in the image containing the digital meter, namely detecting an area containing one and only one complete reading, detecting numbers from the area, and finally sequencing the numbers to obtain the readings of the digital meter.
According to an embodiment of the present invention, the automatically generating a digital synthesis training sample set specifically includes:
b1: preparing a computer font file of a commonly used print form and a liquid crystal font;
b2: randomly selecting a font from the computer font file;
b3: randomly selecting N numbers of 0-9 in the selected font, wherein N is the reading digit range of a common digital meter;
b4: drawing the selected numbers on the blank image, and recording the coordinates of each of the numbers, i.e. drawing the selected numbers on the blank image so that the numbers are in the same row.
B5: saving the drawn image, and writing corresponding numbers and coordinates (upper left corner and lower right corner) in the image into an annotation file to generate the digital synthesis training sample set (synthesis data set);
b6: judging whether the number of the generated samples in the digital synthesis training sample set reaches a preset value or not;
b7: if so, outputting the digital synthesis training sample set, otherwise, continuing to execute steps B2-B6.
According to an embodiment of the present invention, before said saving said rendered image, further comprises:
and adding random interference to the drawn image, wherein the random interference refers to the color and form change of the image and comprises Gaussian blur and color adjustment.
According to an embodiment of the present invention, after said adding random interference to said rendered image, further comprising:
and performing the same projective transformation on the coordinate of the drawn image and the coordinate of the number.
According to an embodiment of the present invention, the performing the same projective transformation on the coordinates of the rendered image and the number specifically includes:
and constructing a projection transformation matrix M by using random parameters, performing projection transformation on the drawn image by using the projection transformation matrix M to obtain a new image, and performing matrix multiplication operation on the projection transformation matrix M and the coordinates of each digit to obtain the coordinates of the digit in the new image.
According to an embodiment of the present invention, the acquiring a required detection recognition model based on the digital synthesis training sample set and the artificially labeled digital real training sample set specifically includes:
acquiring a digital meter image in a transformer substation shot by a camera;
manually marking a reading area on the digital meter image to form a reading area data sample set;
training by using the reading area data sample set to form a reading area detection model;
cutting the reading area to form a new image sample set;
manually labeling each number of the new image sample set to form a manually labeled digital real training sample set;
combining the artificially marked digital real training sample set with the digital synthesis training sample set to form a complete digital data sample set;
training with the complete digital data sample set to form a digital detection model.
According to an embodiment of the present invention, each number is manually labeled to the new image sample set, and the specific labeling content is a numerical value of each number and coordinates of an upper left corner and a lower right corner of a circumscribed rectangle.
According to an embodiment of the present invention, the recognizing the reading of the digital meter in the image including the digital meter by using two detection and recognition models specifically includes:
acquiring a digital meter image in a transformer substation shot by a camera;
detecting all reading areas of the digital meter in the cutting area by using the trained reading area detection model, wherein the obtained detection result is the coordinates of the upper left corner and the lower right corner of a circumscribed rectangle of each reading area;
cutting each reading area out of the cutting area;
inputting the reading areas into the digital detection model respectively, and outputting digital values and circumscribed rectangular coordinates of the reading areas;
rearranging all the numbers of the same reading area according to the sequence from left to right of the coordinates of the upper left corner of the circumscribed rectangle to obtain the final reading of the reading area;
outputting the final readings of all the reading areas.
In a second aspect, the embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method for recognizing the reading of the image-based digital meter of the above embodiment is implemented.
The computer-readable storage medium of the embodiment of the invention, when the computer program stored thereon and corresponding to the image-based digital meter reading identification method of the embodiment is executed by the processor, can accurately identify the readings of a plurality of different types of digital meters, has simple steps, reduces the dependence on manual labeling data, and can quickly adapt to new digital meter types.
In a third aspect, the present invention provides a kit, which includes a memory, a processor, and a computer program stored in the memory, and when the computer program is executed by the processor, the method for recognizing readings of an image-based digital meter according to the above embodiments is implemented.
The corollary equipment of the embodiment of the invention can accurately identify the readings of the digital meters of various different types when the computer program which is stored on the memory and corresponds to the image-based digital meter reading identification method of the embodiment is executed by the processor, has simple steps, reduces the dependence on manual labeling data, and can be quickly adapted to new digital meter types.
Compared with the prior art, the identification method provided by the embodiment of the invention uses a method for synthesizing training data to train the digital detection model, and does not depend on real data completely, so that the workload of manually marking data is reduced, and the training cost is reduced; on the other hand, the synthesized digital shape changes more, the sample distribution is more uniform, and the adaptability to different types of digital meters is better; in addition, two detection and identification models are adopted, the reading area is detected, and then each number in the reading area is detected. The reading area detection model can utilize global information to eliminate interference of other character information, only the reading is reserved, the number detection model only detects the number in the reading area, the accuracy is high, and the number is easily rearranged to obtain the final reading because all the numbers are ensured to be in one line in the previous step.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a block diagram of a hardware structure of a computer terminal of a reading identification method for an image-based digital meter according to an embodiment of the present invention;
FIG. 2 is a flow chart of a reading identification method for an image-based digital meter according to an embodiment of the present invention;
fig. 3 is a schematic flow chart of automatically generating a digital synthesis training sample set in the recognition method according to an embodiment of the present invention;
fig. 4 is a schematic flow chart illustrating a process of acquiring a required detection recognition model in the recognition method according to an embodiment of the present invention;
fig. 5 is a schematic flow chart illustrating the process of automatically recognizing and reading an image containing a digital meter by using the trained detection and recognition model in the recognition method according to an embodiment of the present invention.
Detailed Description
The image-based digital meter reading identification method, the storage medium and the supporting device according to the present invention will be described in further detail with reference to the accompanying drawings and specific embodiments. The advantages and features of the present invention will become more apparent from the following description. It is to be noted that the drawings are in a very simplified form and are not to precise scale, which is merely for the purpose of facilitating and distinctly claiming the embodiments of the present invention.
The method provided by the embodiment can be executed in a computer terminal or a similar arithmetic device. Taking the example of the computer terminal, referring to fig. 1, the computer terminal may include one or more processors 102 (only one is shown in fig. 1) (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), and a memory 104 for storing data, and optionally, the computer terminal may further include a transmission device 106 for communication function and an input/output device 108. It will be understood by those skilled in the art that the structure shown in fig. 1 is only an illustration and is not intended to limit the structure of the computer terminal. For example, the computer terminal may also include more or fewer components than shown in FIG. 1, or have a different configuration than shown in FIG. 1.
The memory 104 may be used to store software programs and modules of application software, such as program instructions/modules corresponding to the image-based digital meter reading identification method provided in the present application, and the processor 102 executes various functional applications and data processing by executing the software programs and modules stored in the memory 104, so as to implement the above-mentioned method. The memory 104 may include high speed random access memory and may also include non-volatile solid state memory. In some embodiments, the memory 104 may further include memory 104 located remotely from the processor 102, which may be connected to a computer terminal over a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The transmission device 106 is used for receiving or transmitting data via a network. Specific examples of such networks may include wireless networks provided by the communications provider of the computer terminal. In one embodiment, the transmission device includes a Network adapter (NIC) that can be connected to other Network devices through a base station to communicate with the internet. In one embodiment, the transmission device 106 may be a Radio Frequency (RF) module, which is used for communicating with the internet in a wireless manner.
Referring to fig. 2, the method for recognizing a reading of a digital meter based on an image according to the present embodiment includes the following steps:
step S1: automatically generating a digital synthesis training sample set;
step S2: acquiring a required detection recognition model based on the digital synthesis training sample set and the artificially labeled digital real training sample set;
step S3: the method comprises the steps of utilizing a trained detection and identification model to automatically identify readings of an image containing a digital meter, specifically adopting two detection and identification models to identify the readings of the digital meter in the image containing the digital meter, namely detecting an area containing one and only one complete reading, detecting numbers from the area, and finally sequencing the numbers to obtain the readings of the digital meter.
The reading identification method of the embodiment of the invention adopts a deep learning algorithm, can accurately identify the readings of various digital meters of different types, has simple steps, reduces the dependence on manually marked data, and can be quickly adapted to new digital meter types.
Further, as shown in fig. 3, the automatically generating a digital synthesis training sample set specifically includes:
b1: preparing a computer font file of a common print form and a liquid crystal font, wherein the font file preferably adopts a truetype format, the font file uses an operating system self-contained font or an open source font downloaded from the internet, and the font file selects a font similar to the reading of a real meter.
B2: randomly selecting a font in the computer font file.
B3: randomly selecting N numbers of 0-9 in the selected font, wherein N is the reading digit range of a common digital meter; the N is typically 1-6, e.g., randomly 2-5.
B4: drawing the selected digits on a blank image, and recording the coordinates of each digit, namely drawing the selected digits on the blank image to enable the digits to be in the same row; the size of the image is calculated from N, the character size, the pitch of each character, so that the image can just accommodate this line of numbers. The numerical value of each number and the coordinates of the upper left and lower right corners of the circumscribed rectangle are recorded. Wherein each character height is set to HCWidth of WCIf the character pitch is d, the image height H is HC+2 × d, width W ═ N × WCPlus (N +1) × d. Coordinate X of upper left corner of ith characteri0＝d+(d+WC)*i，Yi0D, lower right corner coordinate Xi1＝Xi0+WC，Yi1＝Yi0+HC。
B5: storing the drawn image, and writing the corresponding number and coordinates in the image into an annotation file to generate the digital synthesis training sample set (synthesis data set); the images are preferably stored in a jpg format, the format of the annotation file is an xml file compatible with a past voc data set, and each image corresponds to one annotation file.
B6: and judging whether the number of the generated samples in the digital synthesis training sample set reaches a preset value or not.
B7: if so, outputting the digital synthesis training sample set, otherwise, continuing to execute steps B2-B6.
It should be noted that, before the rendered image is saved, post-processing needs to be performed on the generated image and coordinates, specifically, coordinates of 4 corners X of the image in the step B4 are taken in the previous step, and are respectively subjected to random offset to obtain corresponding 4 new coordinates X ', a 3X 3 projection matrix M for mapping X to X' is obtained by using a getperspective transform algorithm in opencv, and a picture in the steps from a warp perspective algorithm in opencv and a parameter M to B4 is applied to obtain a geometrically transformed picture. For the coordinates [ x, y ] of each character in the step B4]Using matrix multiplication to calculate geometrically transformed character seatsThe label [ x ', y', 1]T＝M＊[x，，y，1]T。
In addition, in order to simulate image defects such as blur and dirt in a real meter image and enhance the robustness of a model, random interference needs to be added to the drawn image, wherein the random interference refers to color and form changes of the image and includes but is not limited to gaussian noise, color adjustment, random occlusion and the like. The specific implementation mode is to generate a random number x between 0 and 1, if x is less than 0, blurring is not performed, otherwise, Gaussian blurring is applied to the generated sample image, and the size of a Gaussian filtering kernel is a random integer between 3 and 7. Multiplying brightness, contrast and saturation of the image by a random coefficient k between 0.5 and 1.5 (k is 1 to represent that the image measure is kept unchanged, k is less than 1 to represent a reduction measure, and k is more than 1 to represent an increase measure).
In addition, in order to simulate the change of the shooting angle of a real camera and further increase the diversity of samples, after random interference is added to the drawn image, the same projection transformation needs to be carried out on the coordinates of the drawn image and the numbers.
In a specific example, the performing the same projective transformation on the coordinates of the rendered image and the number specifically includes:
and constructing a projection transformation matrix M by using random parameters, performing projection transformation on the drawn image by using the projection transformation matrix M to obtain a new image, and performing matrix multiplication operation on the projection transformation matrix M and the coordinates of each digit to obtain the coordinates of the digit in the new image. The variation matrix M is a homogeneous matrix of 3 multiplied by 3, the calculation method is to take the coordinates of 4 angles X of the image in the step B4, respectively carry out random offset (the offset does not exceed 1/4 of the image height) to obtain corresponding 4 new coordinates X ', obtain a 3 multiplied by 3 projection matrix M for mapping X to X' by using a getPerfect transform algorithm in opencv, and obtain an image after projection transformation by applying a warpPerfective algorithm and a parameter M in opencv to the image in the step B4. For the coordinates [ x, y ] of each character in the step B4]The projectively transformed character coordinates [ x ', y', 1 ] can be obtained using matrix multiplication]T＝M＊[x，，y，1]T。
Further, as shown in fig. 4, the acquiring a required detection recognition model based on the digital synthesis training sample set and the artificially labeled digital real training sample set specifically includes:
acquiring a digital meter image in a transformer substation shot by a camera;
manually marking a reading area on the digital meter image to form a reading area data sample set;
training by using the reading area data sample set to form a reading area detection model;
cutting the reading area to form a new image sample set;
manually labeling each number of the new image sample set to form a manually labeled digital real training sample set;
combining the artificially marked digital real training sample set with the digital synthesis training sample set to form a complete digital data sample set;
training with the complete digital data sample set to form a digital detection model.
Specifically, each number is manually marked on the new image sample set, and the specific marked content is a numerical value of each number and coordinates of the upper left corner and the lower right corner of the circumscribed rectangle. Illustratively, the artificially labeled digital realistic training sample set includes a numerical value of each number and coordinates of the upper left corner and the lower right corner of a circumscribed rectangle. Wherein the circumscribed rectangle is defined as: if the number is not a liquid crystal number, the external rectangle is a minimum rectangle completely containing the whole number, and four sides of the rectangle are parallel or vertical to coordinate axes; if the number is a liquid crystal number, the circumscribed rectangle is a circumscribed rectangle of the number '8'.
The reading area detection model adopts a yolov5 object detection model, a sample image is input during training, and the model predicts n detected digital areas and corresponding coordinates. The model learns the digital regions in the output image by comparing the detection results with the correct results in the sample labeling, minimizing the class error (softmax error) and the coordinate error (smoothl1 error).
The training mode of the digital detection model is the same as that of the model of the training digital region, and only the training data is different.
Further, as shown in fig. 5, the recognizing the reading of the digital meter in the image including the digital meter by using two detection and recognition models specifically includes:
acquiring a digital meter image in a transformer substation shot by a camera;
positioning the region position of the digital meter in the image by using a template matching method, and cutting the region;
detecting all reading areas of the digital meter in the cutting area by using the trained reading area detection model, wherein the obtained detection result is the coordinates of the upper left corner and the lower right corner of a circumscribed rectangle of each reading area;
cutting each reading area out of the cutting area;
inputting the reading areas into the digital detection model respectively, and outputting digital values and circumscribed rectangular coordinates of the reading areas;
rearranging all the numbers of the same reading area according to the sequence from left to right of the coordinates of the upper left corner of the circumscribed rectangle to obtain the final reading of the reading area;
outputting the final readings of all the reading areas.
It should be noted that, the acquiring of the digital meter image in the substation shot by the camera specifically includes extracting a key point in the image by using a Shi-Tomasi algorithm, and calculating an SIFT feature of the key point. And performing feature matching on the key points and the key points in the template image to find out the corresponding relation of the key points. Computing a basis matrix M using RANSAC algorithmTSo that the key points in the template image can be represented by the matrix MTTo a corresponding point in the acquired image. Then using the matrix MTAnd (4) cutting out the minimum circumscribed rectangle of the region according to the corresponding coordinates of the meter region in the template image in the acquired image.
Based on the image-based digital meter reading identification method of the embodiment, the invention provides a computer-readable storage medium.
In this embodiment, a computer readable storage medium has stored thereon a computer program which, when executed by a processor, implements the image-based digital meter reading identification method of the above-described embodiments.
The computer-readable storage medium of the embodiment of the invention, when the computer program stored thereon and corresponding to the image-based digital meter reading identification method of the embodiment is executed by the processor, can accurately identify the readings of a plurality of different types of digital meters, has simple steps, reduces the dependence on manual labeling data, and can quickly adapt to new digital meter types.
Based on the image-based digital meter reading identification method of the embodiment, the invention also provides a matched device.
In this embodiment, the companion device includes a memory, a processor, and a computer program stored on the memory, which when executed by the processor, implements the image-based digital meter reading identification method of the above-described embodiments.
The corollary equipment of the embodiment of the invention can accurately identify the readings of the digital meters of various different types when the computer program which is stored on the memory and corresponds to the image-based digital meter reading identification method of the embodiment is executed by the processor, has simple steps, reduces the dependence on manual labeling data, and can be quickly adapted to new digital meter types.
It should be noted that the logic and/or steps represented in the flowcharts or otherwise described herein, such as an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
In the description herein, reference to the description of the terms "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (10)
1. The digital meter reading identification method based on the image is characterized by comprising the following steps: the method comprises the following steps:
automatically generating a digital synthesis training sample set;
acquiring a required detection recognition model based on the digital synthesis training sample set and the artificially labeled digital real training sample set;
the method comprises the steps of utilizing a trained detection and identification model to automatically identify readings of an image containing a digital meter, specifically adopting two detection and identification models to identify the readings of the digital meter in the image containing the digital meter, namely detecting an area containing one and only one complete reading, detecting numbers from the area, and finally sequencing the numbers to obtain the readings of the digital meter.
2. The method of claim 1, wherein the automatically generating a digital composite training sample set comprises:
b1: preparing a computer font file of a commonly used print form and a liquid crystal font;
b2: randomly selecting a font from the computer font file;
b3: randomly selecting N numbers of 0-9 in the selected font, wherein N is the reading digit range of a common digital meter;
b4: drawing the selected numbers on a blank image, and recording the coordinates of each number;
b5: storing the drawn image, and writing corresponding numbers and coordinates in the image into an annotation file to generate the digital synthesis training sample set;
b6: judging whether the number of the generated samples in the digital synthesis training sample set reaches a preset value or not;
b7: if so, outputting the digital synthesis training sample set, otherwise, continuing to execute steps B2-B6.
3. The image-based digital meter reading identification method of claim 2, further comprising, prior to said saving of said rendered image:
adding random interference to the drawn image, wherein the random interference refers to the color and form change of the image, and includes but is not limited to Gaussian noise, color adjustment and random occlusion.
4. The image-based digital meter reading identification method of claim 3, further comprising, after said adding random interference to said rendered image:
and performing the same projective transformation on the coordinate of the drawn image and the coordinate of the number.
5. The method of claim 4, wherein performing the same projective transformation on the coordinates of the rendered image and the number comprises:
and constructing a projection transformation matrix M by using random parameters, performing projection transformation on the drawn image by using the projection transformation matrix M to obtain a new image, and performing matrix multiplication operation on the projection transformation matrix M and the coordinates of each digit to obtain the coordinates of the digit in the new image.
6. The method according to claim 5, wherein the obtaining of the required detection and identification model based on the digitally synthesized training sample set and the artificially labeled digitally authentic training sample set specifically comprises:
acquiring a digital meter image in a transformer substation shot by a camera;
manually marking a reading area on the digital meter image to form a reading area data sample set;
training by using the reading area data sample set to form a reading area detection model;
cutting the reading area to form a new image sample set;
manually labeling each number of the new image sample set to form a manually labeled digital real training sample set;
combining the artificially marked digital real training sample set with the digital synthesis training sample set to form a complete digital data sample set;
training with the complete digital data sample set to form a digital detection model.
7. The image-based digital meter reading identification method of claim 6, wherein each number is manually labeled to the new image sample set, and the specific labeling content is the numerical value of each number and the coordinates of the upper left corner and the lower right corner of the circumscribed rectangle.
8. The image-based digital meter reading identification method according to claim 7, wherein the identification of the reading of the digital meter in the image containing the digital meter is realized by using two detection and identification models, which specifically comprises:
acquiring a digital meter image in a transformer substation shot by a camera;
positioning the region position of the digital meter in the image by using a template matching method, and cutting the region;
detecting all reading areas of the digital meter in the cutting area by using the trained reading area detection model, wherein the obtained detection result is the coordinates of the upper left corner and the lower right corner of a circumscribed rectangle of each reading area;
cutting each reading area out of the cutting area;
inputting the reading areas into the digital detection model respectively, and outputting digital values and circumscribed rectangular coordinates of the reading areas;
rearranging all the numbers of the same reading area according to the sequence from left to right of the coordinates of the upper left corner of the circumscribed rectangle to obtain the final reading of the reading area;
outputting the final readings of all the reading areas.
9. Computer-readable storage medium of a method for image-based digital meter reading identification, on which a computer program is stored which, when being executed by a processor, carries out the method for image-based digital meter reading identification according to any one of claims 1 to 8.
10. Kit for a method for image-based digital meter reading identification, characterized in that it comprises a memory, a processor and a computer program stored on said memory, which computer program, when being executed by the processor, carries out the method for image-based digital meter reading identification according to any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110810033.2A CN113537068A (en) | 2021-07-18 | 2021-07-18 | Image-based digital meter reading identification method, storage medium and corollary equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110810033.2A CN113537068A (en) | 2021-07-18 | 2021-07-18 | Image-based digital meter reading identification method, storage medium and corollary equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113537068A true CN113537068A (en) | 2021-10-22 |
Family
ID=78128496
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110810033.2A Pending CN113537068A (en) | 2021-07-18 | 2021-07-18 | Image-based digital meter reading identification method, storage medium and corollary equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113537068A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116935388A (en) * | 2023-09-18 | 2023-10-24 | 四川大学 | Skin acne image auxiliary labeling method and system, and grading method and system |
-
2021
- 2021-07-18 CN CN202110810033.2A patent/CN113537068A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116935388A (en) * | 2023-09-18 | 2023-10-24 | 四川大学 | Skin acne image auxiliary labeling method and system, and grading method and system |
| CN116935388B (en) * | 2023-09-18 | 2023-11-21 | 四川大学 | Skin acne image auxiliary labeling method and system, and grading method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111127422B (en) | Image labeling method, device, system and host | |
| CN109360203A (en) | Method for registering images, image registration device and storage medium | |
| CN110706293B (en) | SURF feature matching-based electronic component positioning and detecting method | |
| CN103955694A (en) | Image recognition meter reading system and method | |
| CN112070137B (en) | Training data set generation method, target object detection method and related equipment | |
| CN111833237A (en) | Image registration method based on convolutional neural network and local homography transformation | |
| CN113936280B (en) | Automatic character recognition system and method for code disc of embedded instrument | |
| CN114872574B (en) | Acquisition and monitoring system and method based on charging pile | |
| CN113537068A (en) | Image-based digital meter reading identification method, storage medium and corollary equipment | |
| CN115187612A (en) | Plane area measuring method, device and system based on machine vision | |
| CN117095417A (en) | Screen shot form image text recognition method, device, equipment and storage medium | |
| CN117173223A (en) | Standard template generation method, device, equipment and medium for ammeter code-breaking screen | |
| CN115019297B (en) | Real-time license plate detection and identification method and device based on color augmentation | |
| CN114419037B (en) | Workpiece defect detection method and device | |
| CN115937839A (en) | Large-angle license plate image recognition method, calculation equipment and storage medium | |
| CN115861801A (en) | Pointer instrument identification method and system, electronic equipment and readable storage medium | |
| CN114882487A (en) | Pointer instrument reading method and system | |
| CN113177988B (en) | Spherical screen camera and laser calibration method, device, equipment and storage medium | |
| CN114037993A (en) | Substation pointer instrument reading method and device, storage medium and electronic equipment | |
| CN114549282A (en) | Method and system for realizing multi-meter reading based on affine transformation | |
| CN113850100A (en) | Method and device for correcting two-dimensional code | |
| CN114168695A (en) | Target position determining method, device, terminal and storage medium | |
| CN113505258A (en) | Method, system, device and medium for prestoring interface data of intelligent watch dial | |
| CN118429328B (en) | Defect detection method, device and equipment for flexible circuit board and storage medium | |
| CN110969063A (en) | Meter reading method, meter reading equipment, remote meter reading device, remote meter reading system and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |