CN112464761A - Identification system and method for pressing plate signboard - Google Patents
Identification system and method for pressing plate signboard Download PDFInfo
- Publication number
- CN112464761A CN112464761A CN202011281996.XA CN202011281996A CN112464761A CN 112464761 A CN112464761 A CN 112464761A CN 202011281996 A CN202011281996 A CN 202011281996A CN 112464761 A CN112464761 A CN 112464761A
- Authority
- CN
- China
- Prior art keywords
- module
- image
- character
- characters
- recognition
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/32—Digital ink
- G06V30/333—Preprocessing; Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/03—Arrangements for converting the position or the displacement of a member into a coded form
- G06F3/041—Digitisers, e.g. for touch screens or touch pads, characterised by the transducing means
- G06F3/044—Digitisers, e.g. for touch screens or touch pads, characterised by the transducing means by capacitive means
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/148—Segmentation of character regions
- G06V30/153—Segmentation of character regions using recognition of characters or words
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2203/00—Indexing scheme relating to G06F3/00 - G06F3/048
- G06F2203/041—Indexing scheme relating to G06F3/041 - G06F3/045
- G06F2203/04104—Multi-touch detection in digitiser, i.e. details about the simultaneous detection of a plurality of touching locations, e.g. multiple fingers or pen and finger
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Character Input (AREA)
Abstract
The invention relates to a recognition system for a pressure plate signboard, which comprises a character and graph recognition module, a communication module, a reset module, an image acquisition module, a storage module, a power supply module and a server, wherein the character and graph recognition module is used for recognizing characters and graphs; the output end of the character and figure recognition module is electrically connected with the input end of the server through the communication module, the output end of the reset module is electrically connected with the output end of the character and figure recognition module and the output end of the power supply module, the image acquisition module and the storage module are electrically connected with the character and figure recognition module in a bidirectional mode, and the power supply module provides electric energy for the character and figure recognition module. Through the information such as mark tablet name of shooting the pressure plate of continuing, mark the tablet name to website relay protection and automatics clamp plate and discern, realize automatically that the clamp plate marks the automatic modeling of tablet and build the storehouse, compare the function automatically according to clamp plate mode table, accessible communication module uploads to the server, and this system marks the tablet discernment rate of accuracy height to the clamp plate, and identification time is short, has saved manpower and materials, has improved work efficiency.
Description
Technical Field
The invention relates to the field of relay protection, in particular to a system and a method for identifying a pressure plate signboard.
Background
The pressing plate is a bridge and a link of a relay protection device or an automatic device which are connected with external wiring, and in order to ensure that the relay protection device or the automatic device is correctly switched on and off according to a pressing plate mode table, at present, inspection personnel need to check the pressing plate of the relay protection device or the automatic device of the transformer substation every month. Because the clamp plates are large in quantity and dense in arrangement, the clamp plates are small in nameplate character body, manual checking and recording are time-consuming and labor-consuming, working efficiency is low, accuracy cannot be guaranteed, and checking and recording errors of the clamp plates are easily caused.
In the prior art, chinese invention patent CN111046986A discloses a "power cable line measurement system and method based on two-dimensional code signboard", which is published as 21/04/2020, and includes a two-dimensional code signboard, a two-dimensional code scanning terminal and a power cable line measurement server, wherein the two-dimensional code signboard is suspended at two ends of a power cable, and the two-dimensional code has identification information of the power cable; the two-dimensional code scanning terminal comprises an image acquisition module, an image identification module, a communication module, a display module, a power supply module and a microcontroller, wherein the power supply module is respectively connected with the image acquisition module, the image identification module, the communication module, the display module and the power supply end of the microcontroller; the output end of the image acquisition module is connected with the input end of the image recognition module, and the output end of the image recognition module is connected with the input end of the microcontroller; the microcontroller is connected with the image acquisition module, the image identification module, the communication module and the display module; in the invention, the system utilizes the two-dimension code recognition model to recognize the two-dimension code information in the image, an image processing algorithm is not used, the recognized image has low complexity, and the character information cannot be effectively recognized.
Disclosure of Invention
The invention provides a system and a method for identifying a signboard of a pressing plate, which aim to solve the technical problems that manual checking and recording are time-consuming and labor-consuming due to the large number of the pressing plates and the accuracy is low at present.
In order to realize the purpose, the technical scheme is as follows:
a recognition system for a pressure plate signboard comprises a character and pattern recognition module, a communication module, a reset module, an image acquisition module, a storage module, a power supply module and a server; the output of character and figure recognition module with communication module input electric connection, the communication module output with the input wireless communication of server is connected, the output of module that resets with the equal electric connection of character and figure recognition module output and power module output, image acquisition module and storage module all with the two-way electric connection of character and figure recognition module, power module does the character and figure recognition module provides the electric energy.
Among the above-mentioned scheme, through information such as mark tablet name, website name of shooting relay protection and automatic device clamp plate, mark the tablet name to website relay protection and automatic device clamp plate and discern, realize automatically that the clamp plate marks the automatic modeling of tablet and build the storehouse, according to the automatic function of comparing of clamp plate mode table, accessible communication module uploads to the server, and this system marks the tablet to the clamp plate and discerns the rate of accuracy height, and identification time is short, has saved manpower and materials, has improved work efficiency simultaneously.
The character and figure recognition module is electrically connected with the touch display module in a bidirectional mode, and the touch display module is in a capacitive touch mode.
According to the scheme, a multi-point touch function can be realized, and a user can perform operations such as selection, attribute information editing, analysis data information reading and the like through a touch display function.
The output end of the clock module is electrically connected with the input end of the character and figure recognition module.
The identification method for the pressure plate signboard is applied to an identification system for the pressure plate signboard and comprises the following steps:
s1: the image acquisition module shoots to obtain a character image;
s2: the character pattern recognition module carries out photoelectric conversion detection on the character image;
s3: the character and pattern recognition module carries out image segmentation on the image subjected to the photoelectric conversion detection;
s4: the character and pattern recognition module recognizes and preprocesses the segmented image;
s5: the character and pattern recognition module performs feature extraction on the image after recognition preprocessing;
s6: and the character and pattern recognition module is used for recognizing and judging the image extracted by the features to obtain a recognition result.
In step S2, a photoelectric conversion process is performed on the text content image of the signboard on the relay protection screen by using photoelectric conversion detection, and a digital signal with a certain gray scale is obtained through analog-to-digital conversion.
In step S3, the graph segmentation is to locate text regions in the image, separate real texts, and identify one by one; due to the regularity of the text, the width of a few pixels exists between characters, the broadband is approximately between 3-8 pixels, most characters are cut out by the character image segmentation according to the characteristic, but some characters are cut into one part and two parts, some characters are adhered together due to noise and the like, the characters of non-Chinese characters with the adhered characters need to be distinguished and segmented, and the segmentation of the Chinese characters and the non-Chinese characters is completed correctly.
The character of the non-Chinese character is judged and segmented, namely, the graph after binarization is normalized based on height, the character is segmented based on a vertical projection method, the widths of the number and the English character are in a smaller range, the narrowest number 1 character is judged whether to accord with the length-width ratio of a single number and the English character through the comparison of the length-width ratio, and judgment in a recognition model is carried out, wherein the character can be judged as the non-Chinese character only when the judgment needs to reach a high confidence coefficient, so that the character is judged as an array or a letter, and the rest parts are combined to be used as the Chinese character for recognition; the segmentation of the final character is based on the text line extracted from the text region in the original image, so that the influence of image processing performed in pre-segmentation on the image quality can be ensured, the original quality of the image is retained to the maximum extent, and the maximum accuracy of the subsequent recognition process is ensured.
In step S4, the identification preprocessing is to process the uncorrected identification image of the signboard by using an adaptive threshold and a binary method of a sliding window, and then perform image refinement, where the processed image is a binary image with white background and black characters.
The binarization method is that the image shot by the pressing plate is grayed, and then the gray value of the point of the image is set to two values, namely 0 or 255, by using binarization, so that the whole image presents a black-and-white effect, and the processed image is a binary image of a black character with a white background; graying is to apply different weights to R, G, and B using a weighted average method to a color platen image and to weight the values of R, G, and B equally, i.e., R ═ G ═ B ═ ω (ω ═ G ═ B ═ G ═ B-r×R+ωg×G+ωbX B)/3, wherein omegar,ωg,ωbAre weighting coefficients of R, G, B, respectively, and ωr+ωg+ωb＝1。ωr,ωg,ωbThe gray images obtained by taking different values are different, and human eyes have different sensitivities to three colors, wherein the human eyes are most sensitive to green, the red and the blue, and therefore omega is enabled to be the most sensitiveg>ωr>ωbAnd obtaining a reasonable gray image.
In the scheme, the binarization is to segment the character graph by selecting a proper self-adaptive threshold, process the gray level graphs of 256 brightness levels into a binary graph, and determine an optimal threshold T by using a correlation algorithm; the gray scale value of the pixel is set to 255 for being larger than the optimal threshold value and 0 for being smaller than the optimal threshold value. The image processed by the above process only has black and white colors, the gray scale range of the image is divided into two types of target and background, and the binarization processing of the image is completed, and the principle formula is as follows:
the image refinement is that the identification processing part uses image refinement to carry out refinement processing on the binarized relay protection pressing plate image, a 3 x 3 refinement template is established, the refinement template is placed in the relay protection pressing plate nameplate image, and the motion of one point by one point is carried out according to the rank sequence, and meanwhile, the pixel points of the template area are marked; the marking results are P0, P1, P2, … and P8, wherein P0 is a central pixel point; when P0 is 1 and the following four conditions are all satisfied, deleting the P0 point; firstly, a,
The situation that P0 is an endpoint and P0 is an inner point is avoided; step two, S (P0) is 1, wherein the condition that P0 is a thin line connecting point with one pixel in width is avoided; ③ P1 & P3 & P7 is 0 or S (P7) ≠ 1, avoiding the situation that P0 is left, upper end point and upper left corner point; the fourth step of avoiding the situation that P0 is a right corner point, a lower corner point and a right lower corner point, wherein P3, P5, P7 is 0 or S (P5) is not equal to 1; the operation is carried out on each pixel point in the relay protection pressing plate nameplate image one by one, when all the pixel points cannot be deleted, the operation is stopped, invalid parts such as black points, blanks and the like in the image are removed after the image is refined, the text image information is enhanced, signals with the same size, position and stroke are obtained, and therefore the difficulty of character recognition is greatly reduced.
In step S6, the identification is to optimize the contents of the template library by the extracted binary image, and finally identify each character by using the convolutional neural network, and similarly, identify all the single texts one by one.
In the scheme, the character and pattern recognition module is provided with an ARM Cortex-A72 framework, a six-core 64-bit high-performance processor and an AI embedded neural network processor NPU, and the peak calculation power is greater than or equal to 5.6Tops, so that strong calculation capability is provided for relay protection and automatic device pressure plate nameplate recognition.
The reset module is connected with the character and pattern recognition module and can reset the character and pattern recognition module and the power supply module, so that the normal operation of a circuit of the device is ensured, and a program is prevented from sending wrong instructions and executing wrong operations; the reset module monitors the power supply voltage when the power supply module is normal, and if the power supply is abnormal, the reset module can carry out forced reset so as to enable the power supply voltage of each module of the device to be in a normal state.
The storage module is connected with the character and pattern recognition module, the storage module consists of a DDR3 memory bank and an EMMC memory and provides data operation and storage for the character and pattern recognition unit, the EMMC interface speed is as high as 52MB per second, and the EMMC has the performances of small volume, large capacity, quickness and upgradability.
The power module mainly comprises a polymer lithium battery and a voltage conversion circuit, the voltage conversion circuit adopts a buck-boost charging management IC to control the charging and discharging of the polymer lithium battery, and simultaneously has the functions of input overvoltage protection, output overvoltage overcurrent protection and the like, the power voltage output by the polymer lithium battery is converted into stable 3.3V working voltage through the voltage conversion circuit to provide a stable and reliable working power supply for each functional module of the device, the polymer lithium battery is light in weight, large in capacity, and strong in working endurance capacity of the system, an NTC temperature probe is attached to the polymer lithium battery, the probe is buried in silica gel to conduct heat more accurately, and the power supply output of the power module is disconnected when abnormal overheating occurs, so that the overheating protection effect is achieved.
Compared with the prior art, the invention has the beneficial effects that:
the invention provides a system and a method for identifying a pressure plate signboard, which are used for identifying the signboard name of a station relay protection and automatic device pressure plate by shooting information such as the signboard name, the station name and the like of the relay protection and automatic device pressure plate, automatically realizing automatic modeling and database building of the pressure plate signboard, automatically comparing the function according to a pressure plate mode table and uploading the result to a server through a communication module.
Drawings
FIG. 1 is a schematic block diagram of a system of the present invention;
FIG. 2 is a flow chart of a method of the present invention;
FIG. 3 is a graphical representation of the identification of an uncorrected sign of the present invention;
fig. 4 is a diagram of the corrected sign recognition result of the present invention.
Detailed Description
The drawings are for illustrative purposes only and are not to be construed as limiting the patent;
the invention is further illustrated below with reference to the figures and examples.
Example 1
As shown in fig. 1, a recognition system for a pressure plate signboard comprises a character and figure recognition module, a communication module, a reset module, an image acquisition module, a storage module, a power supply module and a server; the output of character and figure recognition module with communication module input electric connection, the communication module output with the input wireless communication of server is connected, the output of module that resets with the equal electric connection of character and figure recognition module output and power module output, image acquisition module and storage module all with the two-way electric connection of character and figure recognition module, power module does the character and figure recognition module provides the electric energy.
Among the above-mentioned scheme, through information such as mark tablet name, website name of shooting relay protection and automatic device clamp plate, mark the tablet name to website relay protection and automatic device clamp plate and discern, realize automatically that the clamp plate marks the automatic modeling of tablet and build the storehouse, according to the automatic function of comparing of clamp plate mode table, accessible communication module uploads to the server, and this system marks the tablet to the clamp plate and discerns the rate of accuracy height, and identification time is short, has saved manpower and materials greatly, has improved work efficiency simultaneously.
The character and figure recognition module is electrically connected with the touch display module in a bidirectional mode, and the touch display module is in a capacitive touch mode.
According to the scheme, a multi-point touch function can be realized, and a user can perform operations such as selection, attribute information editing, analysis data information reading and the like through a touch display function.
The output end of the clock module is electrically connected with the input end of the character and figure recognition module.
In the scheme, the character and pattern recognition module is provided with an ARM Cortex-A72 framework, a six-core 64-bit high-performance processor and an AI embedded neural network processor NPU, and the peak calculation power is greater than or equal to 5.6Tops, so that strong calculation capability is provided for relay protection and automatic device pressure plate nameplate recognition.
The reset module is connected with the character and pattern recognition module and can reset the character and pattern recognition module and the power supply module, so that the normal operation of a circuit of the device is ensured, and a program is prevented from sending wrong instructions and executing wrong operations; the reset module monitors the power supply voltage when the power supply module is normal, and if the power supply is abnormal, the reset module can carry out forced reset so as to enable the power supply voltage of each module of the device to be in a normal state.
The storage module is connected with the character and pattern recognition module, the storage module consists of a DDR3 memory bank and an EMMC memory and provides data operation and storage for the character and pattern recognition unit, the EMMC interface speed is as high as 52MB per second, and the EMMC has the performances of small volume, large capacity, quickness and upgradability.
The power module mainly comprises a polymer lithium battery and a voltage conversion circuit, the voltage conversion circuit adopts a buck-boost charging management IC to control the charging and discharging of the polymer lithium battery, and simultaneously has the functions of input overvoltage protection, output overvoltage overcurrent protection and the like, the power voltage output by the polymer lithium battery is converted into stable 3.3V working voltage through the voltage conversion circuit to provide a stable and reliable working power supply for each functional module of the device, the polymer lithium battery is light in weight, large in capacity, and strong in working endurance capacity of the system, an NTC temperature probe is attached to the polymer lithium battery, the probe is buried in silica gel to conduct heat more accurately, and the power supply output of the power module is disconnected when abnormal overheating occurs, so that the overheating protection effect is achieved.
Example 2
As shown in fig. 2, 3 and 4, a method for identifying a signboard of a press plate is applied to an identification system for a signboard of a press plate, comprising the steps of:
s1: the image acquisition module shoots to obtain a character image;
s2: the character pattern recognition module carries out photoelectric conversion detection on the character image;
s3: the character and pattern recognition module carries out image segmentation on the image subjected to the photoelectric conversion detection;
s4: the character and pattern recognition module recognizes and preprocesses the segmented image;
s5: the character and pattern recognition module performs feature extraction on the image after recognition preprocessing;
s6: and the character and pattern recognition module is used for recognizing and judging the image extracted by the features to obtain a recognition result.
In step S2, a photoelectric conversion process is performed on the text content image of the signboard on the relay protection screen by using photoelectric conversion detection, and a digital signal with a certain gray scale is obtained through analog-to-digital conversion.
In step S3, the graph segmentation is to locate text regions in the image, separate real texts, and identify one by one; due to the regularity of the text, the width of a few pixels exists between characters, the broadband is approximately between 3-8 pixels, most characters are cut out by the character image segmentation according to the characteristic, but some characters are cut into one part and two parts, some characters are adhered together due to noise and the like, the characters of non-Chinese characters with the adhered characters need to be distinguished and segmented, and the segmentation of the Chinese characters and the non-Chinese characters is completed correctly.
The character of the non-Chinese character is judged and segmented, namely, the graph after binarization is normalized based on height, the character is segmented based on a vertical projection method, the widths of the number and the English character are in a smaller range, the narrowest number 1 character is judged whether to accord with the length-width ratio of a single number and the English character through the comparison of the length-width ratio, and judgment in a recognition model is carried out, wherein the character can be judged as the non-Chinese character only when the judgment needs to reach a high confidence coefficient, so that the character is judged as an array or a letter, and the rest parts are combined to be used as the Chinese character for recognition; the segmentation of the final character is based on the text line extracted from the text region in the original image, so that the influence of image processing performed in pre-segmentation on the image quality can be ensured, the original quality of the image is retained to the maximum extent, and the maximum accuracy of the subsequent recognition process is ensured.
In step S4, the identification preprocessing is to process the uncorrected identification image of the signboard by using an adaptive threshold and a binary method of a sliding window, and then perform image refinement, where the processed image is a binary image with white background and black characters.
The binarization method comprises graying the image photographed by the pressing plate, and then performing binarizationSetting the gray value of the image point to two values, namely 0 or 255, by using binarization, so that the whole image presents a black-and-white effect, and the processed image is a binary image of a black character with a white background; graying is to apply different weights to R, G, and B using a weighted average method to a color platen image and to weight the values of R, G, and B equally, i.e., R ═ G ═ B ═ ω (ω ═ G ═ B ═ G ═ B-r×R+ωg×G+ωbX B)/3, wherein omegar,ωg,ωbAre weighting coefficients of R, G, B, respectively, and ωr+ωg+ωb＝1。ωr,ωg,ωbThe gray images obtained by taking different values are different, and human eyes have different sensitivities to three colors, wherein the human eyes are most sensitive to green, the red and the blue, and therefore omega is enabled to be the most sensitiveg>ωr>ωbAnd obtaining a reasonable gray image.
In the scheme, the binarization is to segment the character graph by selecting a proper self-adaptive threshold, process the gray level graphs of 256 brightness levels into a binary graph, and determine an optimal threshold T by using a correlation algorithm; the gray scale value of the pixel is set to 255 for being larger than the optimal threshold value and 0 for being smaller than the optimal threshold value. The image processed by the above process only has black and white colors, the gray scale range of the image is divided into two types of target and background, and the binarization processing of the image is completed, and the principle formula is as follows:
the image refinement is that the identification processing part uses image refinement to carry out refinement processing on the binarized relay protection pressing plate image, a 3 x 3 refinement template is established, the refinement template is placed in the relay protection pressing plate nameplate image, and the motion of one point by one point is carried out according to the rank sequence, and meanwhile, the pixel points of the template area are marked; the marking results are P0, P1, P2, … and P8, wherein P0 is a central pixel point; when P0 is 1 and the following four conditions are all satisfied, deleting the P0 point; firstly, a,
Avoiding P0 as an endpoint andp0 is the case of interior points; step two, S (P0) is 1, wherein the condition that P0 is a thin line connecting point with one pixel in width is avoided; ③ P1 & P3 & P7 is 0 or S (P7) ≠ 1, avoiding the situation that P0 is left, upper end point and upper left corner point; the fourth step of avoiding the situation that P0 is a right corner point, a lower corner point and a right lower corner point, wherein P3, P5, P7 is 0 or S (P5) is not equal to 1; the operation is carried out on each pixel point in the relay protection pressing plate nameplate image one by one, when all the pixel points cannot be deleted, the operation is stopped, invalid parts such as black points, blanks and the like in the image are removed after the image is refined, the text image information is enhanced, signals with the same size, position and stroke are obtained, and therefore the difficulty of character recognition is greatly reduced.
In step S6, the identification is to optimize the contents of the template library by the extracted binary image, and finally identify each character by using the convolutional neural network, and similarly, identify all the single texts one by one.
It should be understood that the above-described embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. Other variations and modifications will be apparent to persons skilled in the art in light of the above description. And are neither required nor exhaustive of all embodiments. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the claims of the present invention.
Claims (10)
1. A recognition system for a pressure plate signboard is characterized by comprising a character and pattern recognition module, a communication module, a reset module, an image acquisition module, a storage module, a power supply module and a server; the output of character and figure recognition module with communication module input electric connection, the communication module output with the input wireless communication of server is connected, the output of module that resets with the equal electric connection of output of character and figure recognition module and power module's output, image acquisition module and storage module all with the two-way electric connection of character and figure recognition module, power module does character and figure recognition module provides the electric energy.
2. The system of claim 1, further comprising a touch display module, wherein the touch display module is electrically connected to the text and graphic recognition module in a bi-directional manner, and the touch display module is capacitive.
3. The identification system for a platen signboard of claim 2 further comprising a clock module, wherein an output of the clock module is electrically connected to an input of the text and graphic identification module.
4. A method for recognizing a signboard of a press plate, which is applied to the recognition system for a signboard of a press plate according to claim 3, comprising the steps of:
s1: the image acquisition module shoots to obtain a character image;
s2: the character pattern recognition module carries out photoelectric conversion detection on the character image;
s3: the character and pattern recognition module carries out image segmentation on the image subjected to the photoelectric conversion detection;
s4: the character and pattern recognition module recognizes and preprocesses the segmented image;
s5: the character and pattern recognition module performs feature extraction on the image after recognition preprocessing;
s6: and the character and pattern recognition module is used for recognizing and judging the image extracted by the features to obtain a recognition result.
5. The method for identifying a signboard of claim 4, wherein in step S2, a photoelectric conversion process is performed on the textual content image of the signboard on the relay protection screen by using photoelectric conversion detection, and a digital signal with a certain gray scale is obtained by analog-to-digital conversion.
6. The method as claimed in claim 5, wherein in step S3, the graph segmentation is to locate text regions in the image, separate real text, and perform recognition one by one; due to the regularity of the text, the width of a few pixels exists between characters, the broadband is approximately between 3-8 pixels, most characters are cut out by the character image segmentation according to the characteristic, but some characters are cut into one part and two parts, some characters are adhered together due to noise and the like, the characters of non-Chinese characters with the adhered characters need to be distinguished and segmented, and the segmentation of the Chinese characters and the non-Chinese characters is completed correctly.
7. The method as claimed in claim 6, wherein the distinguishing and segmenting of the characters other than the chinese characters is to normalize the binarized graph based on height, segment the characters based on a vertical projection method, determine whether the widths of the numbers and the english characters are in a relatively small range, compare the aspect ratio of the narrowest number 1 character with the aspect ratio to determine whether the widths of the numbers and the english characters meet the aspect ratio of a single number and the english characters, and perform the determination in the recognition model, wherein the determination requires a high confidence level to determine the characters as the non-chinese characters, thereby determining the characters as the arrays or the letters, and performing the combination of the rest parts as the chinese characters; the segmentation of the final character is based on the text line extracted from the text region in the original image, so that the influence of image processing performed in pre-segmentation on the image quality can be ensured, the original quality of the image is retained to the maximum extent, and the maximum accuracy of the subsequent recognition process is ensured.
8. The method as claimed in claim 7, wherein in step S4, the identification pre-process is to apply adaptive threshold and sliding window binarization to the uncorrected signboard identification image, and then perform image refinement, wherein the processed image is a binary image with black white characters.
9. The method for identifying the signboard of the pressing plate as claimed in claim 8, wherein the binarization method is to graye the image photographed by the pressing plate, and then to set the gray value of the image point to two values, i.e. 0 or 255, by using binarization, so that the whole image exhibits black and white effect, and the processed image is a binary image of black and white characters; graying is to apply different weights to R, G, and B using a weighted average method to a color platen image and to weight the values of R, G, and B equally, i.e., R ═ G ═ B ═ ω (ω ═ G ═ B ═ G ═ B-r×R+ωg×G+ωbX B)/3, wherein omegar,ωg,ωbAre weighting coefficients of R, G, B, respectively, and ωr+ωg+ωb＝1；ωr,ωg,ωbThe gray images obtained by taking different values are different, and human eyes have different sensitivities to three colors, wherein the human eyes are most sensitive to green, the red and the blue, and therefore omega is enabled to be the most sensitiveg>ωr>ωbAnd obtaining a reasonable gray image.
10. The method as claimed in claim 8, wherein in step S6, the recognition is performed by optimizing the contents of the template library with the extracted binary image, and finally recognizing each character with the convolutional neural network, and similarly, recognizing all the single texts one by one.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011281996.XA CN112464761A (en) | 2020-11-16 | 2020-11-16 | Identification system and method for pressing plate signboard |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011281996.XA CN112464761A (en) | 2020-11-16 | 2020-11-16 | Identification system and method for pressing plate signboard |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112464761A true CN112464761A (en) | 2021-03-09 |
Family
ID=74837106
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011281996.XA Pending CN112464761A (en) | 2020-11-16 | 2020-11-16 | Identification system and method for pressing plate signboard |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112464761A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105117721A (en) * | 2015-08-17 | 2015-12-02 | 国网山东省电力公司日照供电公司 | Transformer substation equipment pressing plate fast checking system and checking method |
| CN110992302A (en) * | 2019-10-14 | 2020-04-10 | 广西电网有限责任公司贺州供电局 | Method and device for automatically checking pressure plate |
| CN111754496A (en) * | 2020-06-29 | 2020-10-09 | 国网陕西省电力公司延安供电公司 | Relay protection circuit protection pressing plate state monitoring system and monitoring method |
-
2020
- 2020-11-16 CN CN202011281996.XA patent/CN112464761A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105117721A (en) * | 2015-08-17 | 2015-12-02 | 国网山东省电力公司日照供电公司 | Transformer substation equipment pressing plate fast checking system and checking method |
| CN110992302A (en) * | 2019-10-14 | 2020-04-10 | 广西电网有限责任公司贺州供电局 | Method and device for automatically checking pressure plate |
| CN111754496A (en) * | 2020-06-29 | 2020-10-09 | 国网陕西省电力公司延安供电公司 | Relay protection circuit protection pressing plate state monitoring system and monitoring method |
Non-Patent Citations (1)
| Title |
|---|
| 董五洲: "基于统计分类的混排字符切分算法的研究", 《中国优秀博硕士学位论文全文数据库 (硕士)信息科技辑》, no. 08, pages 138 - 616 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021000702A1 (en) | Image detection method, device, and system | |
| CN105427320A (en) | Image segmentation and extraction method | |
| CN104268814A (en) | Method and device for generating electronic medical record | |
| CN104778453A (en) | Night pedestrian detection method based on statistical features of infrared pedestrian brightness | |
| CN112668541A (en) | Intelligent dictation system based on dot matrix pen and method thereof | |
| US20120257834A1 (en) | Computer vision-based methods for enhanced jbig2 and generic bitonal compression | |
| CN107122775A (en) | A kind of Android mobile phone identity card character identifying method of feature based matching | |
| CN112014413A (en) | Mobile phone glass cover plate window area defect detection method based on machine vision | |
| CN105184329A (en) | Cloud-platform-based off-line handwriting recognition method | |
| CN106709474A (en) | Handwritten telephone number identification, verification and information sending system | |
| CN108960142A (en) | Pedestrian based on global characteristics loss function recognition methods again | |
| CN114549993A (en) | Method, system and device for scoring line segment image in experiment and readable storage medium | |
| CN104391857A (en) | Equipment terminal and method for displaying webpage content according to age information | |
| CN106960188B (en) | Weather image classification method and device | |
| CN112464761A (en) | Identification system and method for pressing plate signboard | |
| CN113159029A (en) | Method and system for accurately capturing local information in picture | |
| CN111611883A (en) | Table layout analysis method, system and equipment based on minimum cell clustering | |
| CN111046770A (en) | Automatic annotation method for photo file figures | |
| CN104268536B (en) | A kind of image method for detecting human face | |
| CN115359505A (en) | Electric power drawing detection and extraction method and system | |
| CN115578739A (en) | Training method and device for realizing IA classification model by combining RPA and AI | |
| CN115035313A (en) | Black-neck crane identification method, device, equipment and storage medium | |
| WO2023101232A1 (en) | Digitization device and method | |
| US10019772B1 (en) | Document image orientation, assessment and correction | |
| CN112418265A (en) | Label detection method, device 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 |