CN109409355B - Novel transformer nameplate identification method and device - Google Patents
Novel transformer nameplate identification method and device Download PDFInfo
- Publication number
- CN109409355B CN109409355B CN201810918895.5A CN201810918895A CN109409355B CN 109409355 B CN109409355 B CN 109409355B CN 201810918895 A CN201810918895 A CN 201810918895A CN 109409355 B CN109409355 B CN 109409355B
- Authority
- CN
- China
- Prior art keywords
- image
- nameplate
- information
- character
- positioning
- 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.)
- Expired - Fee Related
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
- G06V30/14—Image acquisition
- G06V30/146—Aligning or centring of the image pick-up or image-field
- G06V30/1475—Inclination or skew detection or correction of characters or of image to be recognised
- G06V30/1478—Inclination or skew detection or correction of characters or of image to be recognised of characters or characters lines
-
- 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
Abstract
The disclosure discloses a novel transformer nameplate identification method and device. The device obtains transformer data plate information in real time through the rgb camera to pass into the PC end through USB, discern and save. Wherein 32-bit image information that rgb camera was gathered are transmitted to the PC end to the camera bottom, and the PC end carries out the preliminary treatment to the image information of gathering, carries out operations such as data plate information area location. And then positioning the text information in the preprocessed image, and identifying character information. And finally, the automatic identification and recording work of the transformer nameplate information is completed.
Description
Technical Field
The invention belongs to the technical field of measurement, and particularly relates to a novel transformer nameplate identification method and device.
Background
With the development of artificial intelligence, target detection, text information identification and the like are rapidly developed. Text recognition is an important technology in intelligent recognition technology. With the development of computer technology and image technology, the application field of character recognition is continuously expanded, and character recognition, license plate recognition and advertisement information recognition in scenes all relate to recognition of text information in images. Character recognition generally includes recognition of Chinese, English, and numeric characters, and character recognition also actually solves the problem of character classification.
With the development of text recognition technology, the rapid development of OCR technology and the like, and the recognition of characters, the life and work of people are also continuously intelligentized, the character recognition in a life scene is beneficial to classifying and recycling the texts, and the text recognition in an industrialized image is beneficial to automatically and rapidly acquiring machine information and is beneficial to production and processing. These are constantly intelligent, facilitate life, work.
Disclosure of Invention
The invention aims to provide a transformer nameplate identification method and device, and aims to improve the industrial production efficiency through an intelligent automatic identification device.
The invention provides a transformer nameplate identification method, which comprises the following steps:
a novel transformer nameplate identification method comprises the following steps:
s100, acquiring an image, namely acquiring a nameplate image of the transformer by using an RGB (red, green and blue) camera;
s200, positioning an image, positioning a nameplate information area in the nameplate image, and cutting the background area;
s300, enhancing the image, namely enhancing the image obtained in the step S200 by adopting a multi-scale Retinex algorithm;
s400, image correction, namely performing horizontal inclination correction and perspective deformation correction on the nameplate image subjected to image enhancement;
s500, positioning a nameplate text, namely positioning the text information in the corrected image nameplate to enable the related text information to form a vocabulary entry and positioning all the vocabulary entries;
s600, identifying the nameplate text, identifying all the positioned entries, identifying all the text information and displaying the text information on a PC (personal computer) terminal.
The invention also discloses a novel transformer nameplate recognition device, which comprises:
an image acquisition module: collecting a nameplate image of the transformer by using an RGB camera;
an image positioning module: the nameplate information area is used for positioning the nameplate image;
an image enhancement module: the nameplate image locating device is used for carrying out image enhancement on the nameplate image with the nameplate information area located;
an image rectification module: the nameplate image correction device is used for correcting the nameplate image subjected to image enhancement;
nameplate text positioning module: the system is used for positioning the text information in the corrected image nameplate, so that the related text information forms a vocabulary entry and all the vocabulary entries are positioned;
and the nameplate text recognition module is used for recognizing all the positioned entries, recognizing all the text information and displaying the text information at the PC end.
The present disclosure has the following beneficial effects:
this is disclosed through the discernment to transformer data plate information, has reached purposes such as intelligence, swift in the industrial production experiment, and effectual data plate information identification does benefit to in the industrial production to the quantity function statistics of production facility, can improve production efficiency for the industrial production makes intellectuality more.
Drawings
FIG. 1 is a model diagram of a novel transformer nameplate identification method and apparatus of the present invention;
fig. 2 is an overall flow diagram of an embodiment of the invention.
Detailed Description
The invention is described in detail below with reference to the drawings and examples, but the invention is not limited thereto.
In one embodiment, a method for identifying a new type of transformer nameplate includes the following steps:
s100, acquiring an image, namely acquiring a nameplate image of the transformer by using an RGB (red, green and blue) camera;
s200, positioning an image, positioning a nameplate information area in the nameplate image, and cutting the background area;
s300, enhancing the image, namely enhancing the image obtained in the step S200 by adopting a multi-scale Retinex algorithm;
s400, image correction, namely performing horizontal inclination correction and perspective deformation correction on the nameplate image subjected to image enhancement;
s500, positioning a nameplate text, namely positioning the text information in the corrected image nameplate to enable the related text information to form a vocabulary entry and positioning all the vocabulary entries;
s600, identifying the nameplate text, identifying all the positioned entries, identifying all the text information and displaying the text information on a PC (personal computer) terminal.
According to the method, the purposes of intelligence, rapidness and the like are achieved in industrial production experiments by identifying the information of the nameplate of the transformer, the identification of the information of the nameplate is effective, the statistics of the quantity and functions of production equipment in industrial production is facilitated, the production efficiency can be improved, and the industrial production is more intelligent.
In one embodiment, in the step S200, an edge detection is performed by using a canny operator, so as to locate the nameplate information region.
In the embodiment, the canny operator has the advantages that other operators do not have, unnecessary points on the edge are eliminated on the basis of other operators, and the edge closure is also detected and realized, so that the method is most convenient and suitable for practical engineering.
The canny operator comprises the following steps:
1) the image is smoothed, noise filtered, using a gaussian filter.
2) And calculating the gradient strength and the direction of each pixel point in the image.
3) Non-maxima suppression is applied to eliminate spurious responses from edge detection.
4) Double threshold detection is applied to determine true and potential edges.
5) Final completion of edge detection by suppressing encouraged weak edges
In one embodiment, the multi-scale Retinex algorithm in step S300 includes the following steps:
s301, performing multi-scale convolution filtering operation on each color channel of the image in a surrounding function respectively;
s302, performing linear weighted summation on the image processed in the step S301 under multiple scales;
s303, obtaining an enhanced image after calculation through the following formula,
wherein i is RGB color channel, convolution operator, N is scale degree, and wjIs the weight corresponding to the enhancement result in the j-th scale, Si(x, y) represents the original image, G is a gaussian surround function:
delta represents a scale parameter, and x and y respectively represent pixel point coordinates in the image.
In this embodiment, in step S302, the image is subjected to linear weighted summation at multiple scales to ensure that the enhanced image has the advantages of different scales, and generally takes three scales, i.e., high, medium, and low. When the value of the variable delta is large, the image focuses on color preservation, and detailed information is easy to ignore, and when the value of the delta is small, the detailed information is highlighted, but color distortion is easy to cause.
In one embodiment, the horizontal tilt correction comprises the steps of:
s401, performing expansion operation on the image;
s402, performing edge detection by using a sobel operator to obtain edge points;
and S403, hough transformation is carried out on the edge points, a rotation angle is found, and horizontal tilt correction is carried out by utilizing the rotation angle.
In this embodiment, a longest line segment that can be formed by all edge points is found, and an included angle between the longest line segment and the horizontal direction is a rotation angle.
Performing horizontal tilt correction using the rotation angle includes: and rotating the image by combining a rotation function in the opencv library function, and displaying the image to obtain the horizontally corrected image.
The basic idea is as follows: a) solving the diagonal length of the original image, making a square image (a temporary new image) with the diagonal as the side length, obtaining the coordinate of the upper left corner of the original image in the new image, and copying the original image to the new temporary image; b) and taking the central point of the new temporary graph as a rotation point, combining the opencv library function and the rotation angle to obtain a two-dimensional rotation affine transformation matrix, and performing affine transformation to obtain a rotated graph.
In one embodiment, the perspective transformation correction comprises the steps of:
s4001, extracting feature points of the marked nameplate image by using a Harris operator;
s4002, recording the positions of Harris corner points closest to four boundaries of the original image, and performing deformation correction by taking the four corner points as reference points.
In this embodiment, the basic idea of the harris operator is: a local detection window is designed in an image, when the window moves slightly in all directions, the average energy change of the window is considered, and when the energy change exceeds a set threshold value, the central pixel point of the window is extracted as an angular point.
In one embodiment, the step S500 includes the steps of:
s501, decomposing the image into a plurality of different connected components by using a color clustering algorithm;
and S502, verifying the information in the connected domain according to the aspect ratio and the area ratio of the text information, so as to screen out the entry text information.
The embodiment utilizes the characteristics of the text information to distinguish the text information from other information, so that the text information is screened out from the nameplate image.
In one embodiment, the step S600 includes the steps of: s601, performing character segmentation on each entry text message, and segmenting different characters;
s602, character normalization processing, wherein characters are normalized to 25 × 50;
s603, refining the image by using a Rosenfeld skeleton refining algorithm;
s604, extracting character features, namely extracting character feature values based on stroke slope cumulative feature extraction, inflection point amplitude cumulative feature extraction, character contour depth feature extraction and character jumping point statistics;
s605, training the character characteristic values by using a BP neural network classifier algorithm to obtain the classification characteristic of each character;
and S606, repeating the steps to identify the character information in each entry and displaying the character information on the PC side.
In one embodiment, the present disclosure discloses a transformer nameplate identification device, the device comprising:
an image acquisition module: collecting a nameplate image of the transformer by using an RGB camera;
an image positioning module: the nameplate information area is used for positioning the nameplate image;
an image enhancement module: the nameplate image locating device is used for carrying out image enhancement on the nameplate image with the nameplate information area located;
an image rectification module: the nameplate image correction device is used for correcting the nameplate image subjected to image enhancement;
nameplate text positioning module: the system is used for positioning the text information in the corrected image nameplate, so that the related text information forms a vocabulary entry and all the vocabulary entries are positioned;
and the nameplate text recognition module is used for recognizing all the positioned entries, recognizing all the text information and displaying the text information at the PC end.
In this embodiment, the device has reached purposes such as intelligence, swift through the discernment to transformer data plate information in the industrial production experiment, and effectual data plate information identification does benefit to the statistics of the quantity function of production facility in the industrial production, can improve production efficiency for the industrial production makes intellectuality more.
In one embodiment, the image rectification module comprises a horizontal rectification module and a perspective deformation rectification module;
the horizontal correction module is used for performing horizontal inclination correction on the nameplate image;
the perspective deformation module is used for extracting characteristic points for marking the nameplate image by using a Harris operator, recording the positions of Harris angular points closest to four boundaries of the original nameplate image, and performing perspective deformation correction by taking the four angular points as reference points.
In one embodiment, fig. 1 is an overall flow diagram of the present embodiment.
The utility model provides an identification of novel transformer data plate device, includes image acquisition module and information display module.
The image acquisition module enables the rgb camera to acquire nameplate information in real time, and transmits the image information to the PC end through the USB for information processing. And the information display module is mainly used for carrying out upper computer display on the identified information in the nameplate.
In one embodiment, fig. 2 is an overall flow chart of the present embodiment.
The following describes each flow in detail:
s100, a camera positioned in front of a transformer collects nameplate information in real time and transmits the collected image information to a PC (personal computer) end through a USB (universal serial bus);
s200, name plate information area extraction is carried out on the collected image, namely, an important name plate information area is intercepted, and irrelevant background areas are removed. The present invention utilizes a method of edge detection. The basic idea is that a) edge detection is performed by using a canny operator b) accurate positioning and error point removal are performed by using information such as colors.
S300, image enhancement: because the data plate is mostly made by metal material, can appear illumination inequality and reflector lamp phenomenon to the data plate image of shooing. The image is preprocessed by removing illumination through a multi-scale Retinex algorithm. The basic idea of the multi-scale Retinex algorithm is that a) each color channel of an image is subjected to multi-scale convolution filtering operation on a surrounding function, b) the filtered image is subjected to linear weighted summation under multiple scales, c) illumination components are subtracted through logarithm operation, and the obtained reflectivity image is an enhanced image.
Where i is the RGB color channel, w is the convolution operatorjG is a gaussian surrounding function:
s400, image rectification: respectively carrying out horizontal correction and perspective deformation correction.
And (4) horizontal tilt correction, wherein the target area is subjected to tilt correction by adopting hough change. The basic idea is as follows: a) performing expansion operation on the image; b) performing edge detection on the image (by using a Sobel operator); c) and (4) hough transformation is carried out on the edge points, the angle (included angle with the horizontal direction) of the longest line segment is found, namely the rotation angle, and horizontal inclination correction is carried out.
The perspective deformation correction method comprises the following specific steps: a) extracting characteristic points of the marked nameplate image by using a Harris operator, b) recording the positions of Harris angular points closest to four side points of the original image, and performing deformation correction by using the four points as reference points of a world plane in a four-point method.
S500, a positioning method of the connected components is adopted, and due to the fact that colors of text information and background information in the nameplate are different, the texts have similar colors.
The method comprises the following specific implementation steps:
a) the image is decomposed into different connected components using a color clustering algorithm,
b) and comparing the information in the connected domain according to the self characteristic length-width ratio and the area of the text, and verifying, thereby screening out entry text information.
S600, performing text recognition on each entry information in the positioned nameplate, and displaying the text information at the PC terminal. The method comprises the following specific steps:
s601, performing character segmentation on each entry information, and segmenting different characters;
s601, character normalization processing, wherein characters are normalized to 25 × 50;
s602, thinning the image, and using a Rosenfeld skeleton thinning method, wherein Rosenfeld is a parallel thinning method, and the character skeleton after the Rosenfeld processing is in 8-connection form and is used for a 0-1 binary image.
And S603, extracting character features, wherein the features such as slope features of character strokes, depth of character side surfaces and the like are counted to be used as the features for extracting the characters. Specifically, feature values are extracted based on stroke slope cumulative feature extraction, inflection point amplitude cumulative feature extraction, character outline depth feature extraction and character jumping point statistics.
S604, training the extracted features by applying a BP neural network classifier algorithm.
And S605, repeating the steps to identify the character information in each entry and displaying the character information on the PC side.
The above description is only for the specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.
Claims (7)
1. The identification method of the novel transformer nameplate is characterized by comprising the following steps:
s100, acquiring an image, namely acquiring a nameplate image of the transformer by using an RGB (red, green and blue) camera;
s200, positioning an image, positioning a nameplate information area in the nameplate image, and cutting the background area;
s300, enhancing the image, namely enhancing the image obtained in the step S200 by adopting a multi-scale Retinex algorithm;
s400, image correction, namely performing horizontal inclination correction and perspective deformation correction on the nameplate image subjected to image enhancement; wherein the horizontal tilt correction comprises the steps of:
s401, performing expansion operation on the image;
s402, performing edge detection by using a sobel operator to obtain edge points;
s403, hough transformation is carried out on the edge points, a rotation angle is found, and horizontal inclination correction is carried out by utilizing the rotation angle; further, the horizontal tilt correction using the rotation angle includes: combining the rotation function in the opencv library function to rotate the image, and displaying the image to obtain a horizontally corrected image, which specifically comprises: a) solving the diagonal length of the original image, making a square image with the diagonal as the side length as a temporary new image, obtaining the coordinate of the upper left corner of the original image in the temporary new image, and copying the original image to a new temporary image; b) taking the central point of the new temporary graph as a rotation point, combining an opencv library function and a rotation angle, obtaining a two-dimensional rotation affine transformation matrix, and performing affine transformation to obtain a rotated graph;
s500, positioning a nameplate text, namely positioning the text information in the corrected image nameplate to enable the related text information to form a vocabulary entry, and positioning all vocabulary entries in the image nameplate;
s600, identifying a nameplate text, namely identifying the positioned entry, identifying text information in the nameplate and displaying the text information on a PC (personal computer) terminal;
wherein the content of the first and second substances,
the multi-scale Retinex algorithm in step S300 includes the following steps:
s301, performing multi-scale convolution filtering operation on each color channel of the image in a surrounding function respectively;
s302, performing linear weighted summation on the image processed in the step S301 under multiple scales;
s303, obtaining an enhanced image after calculation through the following formula,
wherein i is RGB color channel, convolution operator, N is scale degree, and wjIs the weight corresponding to the enhancement result in the j-th scale, Si(x, y) represents the original image, G is a gaussian surround function:
delta represents a scale parameter, and x and y respectively represent pixel point coordinates in the image;
wherein, in the step S302, the multiple scales are high, medium and low;
the step S600 includes the steps of:
s601, performing character segmentation on each entry text message, and segmenting different characters;
s602, carrying out normalization processing on each character, and normalizing the character to 25 × 50;
s603, processing the image by using a Rosenfeld skeleton thinning algorithm;
s604, obtaining a characteristic value of each character based on stroke slope cumulative characteristic extraction, inflection point amplitude cumulative characteristic extraction, character outline depth characteristic extraction and character jumping point statistics;
s605, training the character characteristic values by using a BP neural network classifier algorithm to obtain the classification characteristic of each character;
s606, repeating the steps to identify the character information in each entry, and displaying the character information on the PC;
the method is used for: and counting the number of the transformer equipment produced in the industrial production through nameplate information identification.
2. The method according to claim 1, wherein in step S200, edge detection is performed by using canny operator, so as to locate the nameplate information area.
3. The method of claim 1, wherein the perspective deformation correction comprises the steps of:
s4001, extracting and marking feature points of the nameplate image by using a Harris operator;
s4002, recording the positions of Harris corner points closest to the four boundaries of the image processed in the step S200, and performing perspective deformation correction by taking the four corner points as reference points.
4. The method according to claim 1, wherein the step S500 comprises the steps of:
s501, decomposing the image into a plurality of different connected domains by using a color clustering algorithm;
and S502, verifying the information in the connected domain according to the aspect ratio and the area ratio of the text information, so as to screen out the entry text information.
5. The method according to claim 1, characterized in that the rotation angle is obtained by:
and finding the longest line segment which can be formed by all the edge points, wherein the included angle between the longest line segment and the horizontal direction is the rotation angle.
6. A transformer nameplate identification device based on the method of claim 1, the device comprising:
an image acquisition module: collecting a nameplate image of the transformer by using an RGB camera;
an image positioning module: the nameplate information area is used for positioning the nameplate image;
an image enhancement module: the nameplate image locating device is used for carrying out image enhancement on the nameplate image with the nameplate information area located;
an image rectification module: the nameplate image correction device is used for correcting the nameplate image subjected to image enhancement;
nameplate text positioning module: the system is used for positioning the text information in the corrected image nameplate, so that the related text information forms a vocabulary entry and all the vocabulary entries are positioned;
and the nameplate text recognition module is used for recognizing all the positioned entries, recognizing all the text information and displaying the text information at the PC end.
7. The apparatus of claim 6, wherein the image rectification module comprises a horizontal rectification module and a perspective deformation rectification module;
the horizontal correction module is used for performing horizontal inclination correction on the nameplate image;
the perspective deformation module is used for extracting characteristic points for marking the nameplate image by using a Harris operator, recording the positions of Harris angular points closest to four boundaries of the original nameplate image, and performing perspective deformation correction on the four angular points.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810918895.5A CN109409355B (en) | 2018-08-13 | 2018-08-13 | Novel transformer nameplate identification method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810918895.5A CN109409355B (en) | 2018-08-13 | 2018-08-13 | Novel transformer nameplate identification method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109409355A CN109409355A (en) | 2019-03-01 |
| CN109409355B true CN109409355B (en) | 2021-09-14 |
Family
ID=65464313
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810918895.5A Expired - Fee Related CN109409355B (en) | 2018-08-13 | 2018-08-13 | Novel transformer nameplate identification method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109409355B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110334647A (en) * | 2019-07-03 | 2019-10-15 | 云南电网有限责任公司信息中心 | A kind of parameter format method based on image recognition |
| CN110490185A (en) * | 2019-08-23 | 2019-11-22 | 北京工业大学 | One kind identifying improved method based on repeatedly comparison correction OCR card information |
| CN110991448A (en) * | 2019-11-27 | 2020-04-10 | 云南电网有限责任公司电力科学研究院 | Text detection method and device for nameplate image of power equipment |
| CN111291752A (en) * | 2020-01-22 | 2020-06-16 | 山东浪潮通软信息科技有限公司 | Invoice identification method, equipment and medium |
| CN111401289B (en) * | 2020-03-24 | 2024-01-23 | 国网上海市电力公司 | Intelligent identification method and device for transformer component |
| CN111245103A (en) * | 2020-03-31 | 2020-06-05 | 贵州电网有限责任公司 | Display and storage system of power grid transformer nameplate based on neural computing rod |
| CN112446370B (en) * | 2020-11-24 | 2024-03-29 | 东南大学 | Method for identifying text information of nameplate of power equipment |
| CN113537147B (en) * | 2021-08-09 | 2022-04-12 | 桂林电子科技大学 | Night lane line detection method based on illumination compensation |
| CN114267038B (en) * | 2022-03-03 | 2022-05-20 | 南京甄视智能科技有限公司 | Nameplate type identification method and device, storage medium and equipment |
| CN115187881A (en) * | 2022-09-08 | 2022-10-14 | 国网江西省电力有限公司电力科学研究院 | Power equipment nameplate identification and platform area compliance automatic checking system and method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108009547A (en) * | 2017-12-26 | 2018-05-08 | 深圳供电局有限公司 | A kind of nameplate recognition methods of substation equipment and device |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20120088350A (en) * | 2011-01-31 | 2012-08-08 | 한국전자통신연구원 | Apparatus for generating high resolution image |
| CN104200209B (en) * | 2014-08-29 | 2017-11-03 | 南京烽火星空通信发展有限公司 | A kind of pictograph detection method |
-
2018
- 2018-08-13 CN CN201810918895.5A patent/CN109409355B/en not_active Expired - Fee Related
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108009547A (en) * | 2017-12-26 | 2018-05-08 | 深圳供电局有限公司 | A kind of nameplate recognition methods of substation equipment and device |
Non-Patent Citations (4)
| Title |
|---|
| opencv利用仿射变换函数对图像进行任意角度旋转;fu_shuwu;《https://blog.csdn.net/fu_shuwu/article/details/77688411?utm_source=blogxgwz7》;20170829;第1页 * |
| 基于无监督学习的铭牌文字定位和识别;孙晔;《中国学位论文全文数据库》;20160831;1-77 * |
| 细化算法;maginy;《https://blog.csdn.net/maginy/article/details/37762525》;20140714;1-3 * |
| 车牌校正中水平方向的边缘提取;阿迪spring;《https://blog.csdn.net/wangzengdi/article/details/25690473?locationNum=12》;20140513;1-2 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109409355A (en) | 2019-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109409355B (en) | Novel transformer nameplate identification method and device | |
| CN107545239B (en) | Fake plate detection method based on license plate recognition and vehicle characteristic matching | |
| CN107491730A (en) | A kind of laboratory test report recognition methods based on image procossing | |
| CN107424142B (en) | Weld joint identification method based on image significance detection | |
| WO2018018788A1 (en) | Image recognition-based meter reading apparatus and method thereof | |
| CN104751142B (en) | A kind of natural scene Method for text detection based on stroke feature | |
| CN107909081B (en) | Method for quickly acquiring and quickly calibrating image data set in deep learning | |
| CN111915704A (en) | Apple hierarchical identification method based on deep learning | |
| CN103824091B (en) | A kind of licence plate recognition method for intelligent transportation system | |
| CN101807257A (en) | Method for identifying information of image tag | |
| CN109687382B (en) | Relay protection pressing plate switching state identification method based on color template matching | |
| CN114549981A (en) | Intelligent inspection pointer type instrument recognition and reading method based on deep learning | |
| US20140301608A1 (en) | Chemical structure recognition tool | |
| CN110287787B (en) | Image recognition method, image recognition device and computer-readable storage medium | |
| CN108961262B (en) | Bar code positioning method in complex scene | |
| CN111861866A (en) | Panoramic reconstruction method for substation equipment inspection image | |
| CN103886319A (en) | Intelligent held board recognizing method based on machine vision | |
| CN110503051B (en) | Precious wood identification system and method based on image identification technology | |
| CN115731493A (en) | Rainfall micro physical characteristic parameter extraction and analysis method based on video image recognition | |
| CN109271882B (en) | Method for extracting color-distinguished handwritten Chinese characters | |
| CN113283439B (en) | Intelligent counting method, device and system based on image recognition | |
| Kumar | An efficient text extraction algorithm in complex images | |
| CN113052234A (en) | Jade classification method based on image features and deep learning technology | |
| CN111539330B (en) | Transformer substation digital display instrument identification method based on double-SVM multi-classifier | |
| Rastegar et al. | An intelligent control system using an efficient License Plate Location and Recognition Approach |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210914 |