CN106485728B - A kind of automatic measure of bar shaped main transformer oil level indicator - Google Patents
A kind of automatic measure of bar shaped main transformer oil level indicator Download PDFInfo
- Publication number
- CN106485728B CN106485728B CN201610814028.8A CN201610814028A CN106485728B CN 106485728 B CN106485728 B CN 106485728B CN 201610814028 A CN201610814028 A CN 201610814028A CN 106485728 B CN106485728 B CN 106485728B
- Authority
- CN
- China
- Prior art keywords
- image
- oil level
- level indicator
- bar shaped
- main transformer
- 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.)
- Active
Links
Classifications
-
- 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
- Image Analysis (AREA)
Abstract
The invention discloses a kind of automatic measures of bar shaped main transformer oil level indicator, belong to machine vision instrument and meter intelligent testing technology field, solve the problem of that there are backlight to take pictures in imaging process, uneven illumination is even etc. and cannot achieve automatic measurement.Utilization orientation histogram of gradients feature and support vector machines carry out Primary Location to the oil level indicator of bar shaped main transformer, color segmentation and Canny edge extracting method are recycled to be positioned to obtain the gray level image f for containing only complete oil level indicator to oil level indicator1(x, y) obtains containing only the gray level image f of complete observation window using the black outer rim of the removal oil level indicator such as expansion, holes filling and corrosion2(x, y) obtains the position of oil level indicator index line using Otsu threshold segmentation method, indicates line position and image f2Oil level indicator bound can be calculated oil level indicator percentage and mark in (x, y).Automatic measurement and reading for substation's bar shaped main transformer oil level indicator.
Description
Technical field
A kind of automatic measure of bar shaped main transformer oil level indicator, for substation's bar shaped main transformer oil level indicator it is automatic measurement and
Reading, belongs to machine vision instrument and meter intelligent testing technology field.
Background technique
Electric instrument is the important tool of people's detection and monitoring device, and accuracy is most important, therefore electric power
Department needs periodically to examine and determine instrument, and task is heavy.Using the very easy error of artificial calibrating mode and efficiency
Lowly, therefore people are excited to various instrument automation degree quantifier eliminations.Field is measured automatically in industrial instrument, it is mainly sharp
The automatic measurement of instrument is realized with the image processing method based on machine vision.Compared to artificial reading, automatic the advantages of measuring
Have：24 hours uninterrupted readings, save manual labor, stability is high, high-efficient.Therefore in industrial circle application, instrument
Table is measured automatically very high researching value.
Due to the function of industrial instrument, use is different, so that industrial instrument is many kinds of, structure is different, even if
It is the instrument of same model, there is also color, shape, the difference such as size.Thus, there is no a kind of omnipotent instrument to spend automatically
Amount method can accurately realize automatic measurement to all instrument.The instrument measured automatically is realized in this patent, is one
Kind bar shaped main transformer oil level indicator.Currently, for ellipse without grid oil level indicator and pointer dial plate, existing patent is able to achieve automatic degree
Amount, but it is directed to the oil level indicator of bar shaped main transformer, both patents all can not achieve automatic measurement.Due in imaging process,
There are backlight to take pictures, and uneven illumination is even, imaging noise, and focusing is fuzzy, and background is complicated, blocks, and tilts, and rotation etc. influences, in reality
When the automatic metric algorithm of existing instrument, need to consider several factors, in addition, shaping is precisely located in complicated image background
Shape oil level indicator is also a difficult point.
Summary of the invention
The present invention provides a kind of automatic measure of bar shaped main transformer oil level indicator in view of the above shortcomings, solves existing
For technology in imaging process, there are backlight to take pictures, uneven illumination is even, imaging noise, focusing is fuzzy, background is complicated, blocks, inclines
Tiltedly, the problem of rotation etc. influences, and cannot achieve the automatic measurement of Instrument image.
To achieve the goals above, the technical solution adopted by the present invention is：
A kind of automatic measure of bar shaped main transformer oil level indicator, which is characterized in that include the following steps：
Step 1, input original image f1(x,y)；
Step 2 extracts original image f1The histograms of oriented gradients feature (HOG) of (x, y) recycles trained branch
It holds vector machine (SVM) to classify, obtains the Primary Location image f of oil level indicator2(x,y)；
Step 3 chooses threshold value appropriate, to Primary Location image f2(x, y) carries out color segmentation, obtains image f3(x,
y)；
Step 4, to Primary Location image f2(x, y) carries out Canny edge extracting, obtains image f4(x, y), by image f3
(x, y) and image f4(x, y) is carried out or operation obtains image f5(x,y)；
Step 5, to image f5(x, y) successively carries out triple-expansion, a holes filling and ten etching operations and obtains figure
As f6(x,y)；
Step 6 extracts image f6Minimum rectangle in the connected component of (x, y), as bar shaped oil level indicator, in Primary Location
Image f2It cuts minimum rectangle on (x, y), obtains image f7(x,y)；
Step 7, Otsu Threshold segmentation handle image f7The position of index line is calculated in (x, y)；
Step 8 utilizes instruction line position and image f7(x, y) obtains the position of bar shaped oil level indicator bound, passes through bar shaped
The position of oil level indicator bound calculates the percentage of oil level indicator, and marks in Primary Location image f2On (x, y).
Further, the detailed process of the step 2 is：
(21) the bar shaped oil level indicator region for intercepting out several oil level indicator images, as positive sample, while intercepting
The HOG feature of positive negative sample is extracted as negative sample in background area, and the extraction of HOG feature is real by the library Opencv intrinsic function
Existing, input SVM is trained, and obtains trained SVM；
(22) original image f is inputted1(x, y) extracts original image f1The HOG feature of (x, y), the extraction of HOG feature pass through
The library Opencv intrinsic function is realized, is classified using trained SVM, is obtained the Primary Location image f of oil level indicator2(x,y)。
Further, in the step 3, to Primary Location image f2The formula that (x, y) carries out color segmentation is as follows：
Wherein R (x, y) indicates Primary Location image f2The channel R of the RGB channel of (x, y), G (x, y) indicate Primary Location
Image f2The channel G of the RGB channel of (x, y), B (x, y) indicate Primary Location image f2The channel B of the RGB channel of (x, y).
Further, in the step 4, by image f3(x, y) and image f4(x, y) is carried out or operation obtains image f5(x,y)
Calculation formula it is as follows：
Further, the detailed process of the step 5 is：
Indicate etching operation,Indicate expansion,Indicate holes filling；
Wherein A indicates original image, AcIndicate the supplementary set of A, B indicates structural elements, (B)z=w | and w=b+z, b ∈ B } it indicates B's
Origin translation to point z,Indicate all elements that the set is mapped about the origin of structural elements B, Xk
Maximum connected component is indicated, to the image f in step 45(x, y) carries out carrying out a secondary aperture after triple-expansion morphological operation
Hole padding finally carries out ten etching operations, obtains image f6(x,y)。
Further, the detailed process of the step 6 is：
(61) image f is extracted6The connected component of (x, y), formula are as follows：
In formula, A indicates original image, and B indicates structural elements, XKIndicate maximum connected component；
(62) it is gone with the smallest rectangle comprising maximum connected component, in f2Corresponding minimum rectangle is intercepted on (x, y), is obtained
To image f7(x, y), wherein minimum rectangle highest point ordinate positional value is P1, the minimum point value of ordinate is P2。
Further, the detailed process of the step 7 is：
(71) to image f7(x, y) carries out Otsu Threshold segmentation, and Otsu Threshold segmentation is real by the library Opencv intrinsic function
It is existing, maximum two pieces of white color components after segmentation are found, the ordinate of two white piece of highs and lows is obtained, totally 4 values；
(72) 4 values obtained in step (71) are ranked up, find out the ordinate average value of second with third,
Ordinate value P as instruction line position3。
Further, in the step 8, the formula for calculating the percentage of oil level indicator is：
Compared with the prior art, the advantages of the present invention are as follows：
One, in the positioning of oil level indicator, the present invention innovatively uses the classical HOG characteristics algorithm for pedestrian detection,
And in conjunction with SVM algorithm, the positioning of oil level indicator can substantially reach 100%, and accuracy is very high, well solve complicated back
The problem of being positioned under scape；
Two, present invention employs the image processing algorithm HOG feature of simple algorithm for pattern recognition SVM and some classics,
Otsu Threshold segmentation, Canny edge extracting etc. are fast implemented by calling the library Opencv can be simple；By largely surveying
An average picture operation time is calculated in 1s in examination, thus operand of the present invention is few, processing speed is fast, it is high-efficient with
And preferable robustness；
Three, due to that can be solved well using Morphology Algorithms, the present invention such as color segmentation, burn into expansion, holes filling
Certainly backlight is taken pictures, and uneven illumination is even, imaging noise problem, since the calculated result of oil level indicator is percentage, can be kept away well
The inclination exempted from, rotation bring influence；
Four, the present invention is realized on MATLAB software platform, and last successful implantation is flat to VS2013+Opencv
Platform illustrates that the transplantability of this algorithm is good, due to the C++ platform having been migrated under VS2013, can also be transplanted to other platforms,
Such as DSP platform.
Detailed description of the invention
Fig. 1 is algorithm flow schematic diagram of the invention；
Fig. 2 is the image of Primary Location of the invention；
Fig. 3 is the image after color segmentation of the invention；
Fig. 4 is the image after of the invention or operation；
Fig. 5 is the image after expansive working of the invention；
Fig. 6 is the image after holes filling operation of the invention；
Fig. 7 is the image after etching operation of the invention；
Fig. 8 is the pinpoint image of oil level indicator of the invention；
Fig. 9 is the image after Otsu segmentation of the invention；
Figure 10 is the oil level of the invention calculated and the image being labeled.
Specific embodiment
The present invention is further illustrated with reference to the accompanying drawings and examples.
A kind of automatic measure of bar shaped main transformer oil level indicator, includes the following steps：
Step 1, input original image f1(x,y)；
Step 2 extracts original image f1The histograms of oriented gradients feature (HOG) of (x, y) recycles trained branch
It holds vector machine (SVM) to classify, obtains the Primary Location image f of oil level indicator2(x,y)；Detailed process is：
(21) the bar shaped oil level indicator region for intercepting out several oil level indicator images, as positive sample, while intercepting
The HOG feature of positive negative sample is extracted as negative sample in background area, and the extraction of HOG feature is real by the library Opencv intrinsic function
Existing, input SVM is trained, and obtains trained SVM；
(22) original image f is inputted1(x, y) extracts original image f1The HOG feature of (x, y), the extraction of HOG feature pass through
The library Opencv intrinsic function is realized, is classified using trained SVM, is obtained the Primary Location image f of oil level indicator2(x,y)。
Step 3 chooses threshold value appropriate, to Primary Location image f2(x, y) carries out color segmentation, obtains image f3(x,
y)；To Primary Location image f2The formula that (x, y) carries out color segmentation is as follows：
Wherein R (x, y) indicates Primary Location image f2The channel R of the RGB channel of (x, y), G (x, y) indicate Primary Location
Image f2The channel G of the RGB channel of (x, y), B (x, y) indicate Primary Location image f2The channel B of the RGB channel of (x, y).
Step 4, to Primary Location image f2(x, y) carries out Canny edge extracting, obtains image f4(x, y), by image f3
(x, y) and image f4(x, y) is carried out or operation obtains image f5(x,y)；By image f3(x, y) and image f4(x, y) is carried out or fortune
Calculation obtains image f5The calculation formula of (x, y) is as follows：
Step 5, to image f5(x, y) successively carries out triple-expansion, a holes filling and ten etching operations and obtains figure
As f6(x,y)；Detailed process is：
Indicate etching operation,Indicate expansion,Indicate holes filling；
Wherein A indicates original image, AcIndicate the supplementary set of A, B indicates structural elements, (B)z=w | and w=b+z, b ∈ B } it indicates B's
Origin translation to point z,Indicate all elements that the set is mapped about the origin of structural elements B, Xk
Maximum connected component is indicated, to the image f in step 45(x, y) carries out carrying out a secondary aperture after triple-expansion morphological operation
Hole padding finally carries out ten etching operations, obtains image f6(x,y)。
Step 6 extracts image f6Minimum rectangle in the connected component of (x, y), as bar shaped oil level indicator, in Primary Location
Image f2It cuts minimum rectangle on (x, y), obtains image f7(x,y)；Detailed process is：
(61) image f is extracted6The connected component of (x, y), formula are as follows：
In formula, A indicates original image, and B indicates structural elements, XKIndicate maximum connected component；
(62) it is gone with the smallest rectangle comprising maximum connected component, in f2Corresponding minimum rectangle is intercepted on (x, y), is obtained
To image f7(x, y), wherein minimum rectangle highest point ordinate positional value is P1, the minimum point value of ordinate is P2。
Step 7, Otsu Threshold segmentation handle image f7The position of index line is calculated in (x, y)；Detailed process is：
(71) to image f7(x, y) carries out Otsu Threshold segmentation, and Otsu Threshold segmentation is real by the library Opencv intrinsic function
It is existing, maximum two pieces of white color components after segmentation are found, the ordinates of two white piece of highs and lows totally 4 values are obtained；
(72) 4 values obtained in step (71) are ranked up, find out the ordinate average value of second with third,
Ordinate value P as instruction line position3。
Step 8 utilizes instruction line position and image f7(x, y) obtains the position of bar shaped oil level indicator bound, passes through bar shaped
The position of oil level indicator bound calculates the percentage of oil level indicator, and marks in Primary Location image f2On (x, y).Calculate oil level indicator
The formula of percentage be：
Present invention utilizes histograms of oriented gradients features (HOG) and support vector machines (SVM), color segmentation, the side Canny
The classics image procossing such as Morphology Algorithms such as edge extracting method, Otsu threshold segmentation method, expansion, holes filling and corrosion is calculated
Method is obtained in conjunction with the concrete condition of bar shaped main transformer oil level indicator image by the continuous test of combination and various threshold values to algorithm
The present invention.
Claims (8)
1. a kind of automatic measure of bar shaped main transformer oil level indicator, which is characterized in that include the following steps：
Step 1, input original image f1(x,y)；
Step 2 extracts original image f1The histograms of oriented gradients feature (HOG) of (x, y), recycle it is trained support to
Amount machine (SVM) is classified, and the Primary Location image f of oil level indicator is obtained2(x,y)；
Step 3 chooses threshold value appropriate, to Primary Location image f2(x, y) carries out color segmentation, obtains image f3(x,y)；
Step 4, to Primary Location image f2(x, y) carries out Canny edge extracting, obtains image f4(x, y), by image f3(x,y)
With image f4(x, y) is carried out or operation obtains image f5(x,y)；
Step 5, to image f5(x, y) successively carries out triple-expansion, a holes filling and ten etching operations and obtains image f6
(x,y)；
Step 6 extracts image f6Minimum rectangle in the connected component of (x, y), as bar shaped oil level indicator, in Primary Location image
f2It cuts minimum rectangle on (x, y), obtains image f7(x,y)；
Step 7, Otsu Threshold segmentation handle image f7The position of index line is calculated in (x, y)；
Step 8 utilizes instruction line position and image f7(x, y) obtains the position of bar shaped oil level indicator bound, passes through bar shaped oil level indicator
The position of bound calculates the percentage of oil level indicator, and marks in Primary Location image f2On (x, y).
2. a kind of automatic measure of bar shaped main transformer oil level indicator according to claim 1, which is characterized in that the step
2 detailed process is：
(21) the bar shaped oil level indicator region for intercepting out several oil level indicator images, as positive sample, while intercepting some backgrounds
The HOG feature of positive negative sample is extracted, the extraction of HOG feature is realized by the library Opencv intrinsic function, defeated as negative sample in region
Enter SVM to be trained, obtains trained SVM；
(22) original image f is inputted1(x, y) extracts original image f1The HOG feature of (x, y), the extraction of HOG feature pass through
The library Opencv intrinsic function is realized, is classified using trained SVM, is obtained the Primary Location image f of oil level indicator2(x,y)。
3. a kind of automatic measure of bar shaped main transformer oil level indicator according to claim 1, which is characterized in that the step
In 3, to Primary Location image f2The formula that (x, y) carries out color segmentation is as follows：
Wherein R (x, y) indicates Primary Location image f2The channel R of the RGB channel of (x, y), G (x, y) indicate Primary Location image f2
The channel G of the RGB channel of (x, y), B (x, y) indicate Primary Location image f2The channel B of the RGB channel of (x, y).
4. a kind of automatic measure of bar shaped main transformer oil level indicator according to claim 1, which is characterized in that the step
In 4, by image f3(x, y) and image f4(x, y) is carried out or operation obtains image f5The calculation formula of (x, y) is as follows：
5. a kind of automatic measure of bar shaped main transformer oil level indicator according to claim 1, which is characterized in that the step
5 detailed process is：
Indicate etching operation,Indicate expansion,Indicate holes filling；
Wherein A indicates original image, AcIndicate the supplementary set of A, B indicates structural elements, (B)z=w | and w=b+z, b ∈ B } it indicates the origin of B
Point z is moved to,Indicate all elements that the set is mapped about the origin of structural elements B, XkIt indicates
Maximum connected component, to the image f in step 45Hole of progress is filled out after (x, y) carries out triple-expansion morphological operation
Operation is filled, ten etching operations is finally carried out, obtains image f6(x,y)。
6. a kind of automatic measure of bar shaped main transformer oil level indicator according to claim 1, which is characterized in that the step
6 detailed process is：
(61) image f is extracted6The connected component of (x, y), formula are as follows：
In formula, A indicates original image, and B indicates structural elements, XKIndicate maximum connected component；
(62) it is gone with the smallest rectangle comprising maximum connected component, in f2Corresponding minimum rectangle is intercepted on (x, y), obtains figure
As f7(x, y), wherein minimum rectangle highest point ordinate positional value is P1, the minimum point value of ordinate is P2。
7. a kind of automatic measure of bar shaped main transformer oil level indicator according to claim 6, which is characterized in that the step
7 detailed process is：
(71) to image f7(x, y) carries out Otsu Threshold segmentation, and Otsu Threshold segmentation is realized by the library Opencv intrinsic function, seeks
Maximum two pieces of white color components after dividing are looked for, the ordinate of two white piece of highs and lows is obtained, totally 4 values；
(72) 4 values obtained in step (71) are ranked up, find out the ordinate average value of second with third, as
Indicate the ordinate value P of line position3。
8. a kind of automatic measure of bar shaped main transformer oil level indicator according to claim 7, which is characterized in that the step
In 8, the formula for calculating the percentage of oil level indicator is：
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610814028.8A CN106485728B (en) | 2016-09-09 | 2016-09-09 | A kind of automatic measure of bar shaped main transformer oil level indicator |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610814028.8A CN106485728B (en) | 2016-09-09 | 2016-09-09 | A kind of automatic measure of bar shaped main transformer oil level indicator |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106485728A CN106485728A (en) | 2017-03-08 |
| CN106485728B true CN106485728B (en) | 2018-11-27 |
Family
ID=58273434
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610814028.8A Active CN106485728B (en) | 2016-09-09 | 2016-09-09 | A kind of automatic measure of bar shaped main transformer oil level indicator |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106485728B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107610128A (en) * | 2017-09-26 | 2018-01-19 | 山东鲁能智能技术有限公司 | The method for inspecting and device of a kind of oil level indicator |
| CN111272257A (en) * | 2020-02-20 | 2020-06-12 | 上海电机学院 | Transformer oil level monitoring system based on machine vision |
| CN114143446A (en) * | 2021-10-20 | 2022-03-04 | 深圳航天智慧城市系统技术研究院有限公司 | Histogram identification method, system, storage medium and equipment based on edge calculation |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7041059B2 (en) * | 2002-08-02 | 2006-05-09 | Diagnostic Ultrasound Corporation | 3D ultrasound-based instrument for non-invasive measurement of amniotic fluid volume |
| CN102254159B (en) * | 2011-07-07 | 2014-07-02 | 清华大学深圳研究生院 | Interpretation method for digital readout instrument |
| CN103955694B (en) * | 2014-04-09 | 2017-05-24 | 广州邦讯信息系统有限公司 | Image recognition meter reading system and method |
| CN105809168A (en) * | 2015-12-23 | 2016-07-27 | 佛山市质量计量监督检测中心 | Image fuzzy identification method of standard constant temperature bath glass thermometer |
-
2016
- 2016-09-09 CN CN201610814028.8A patent/CN106485728B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN106485728A (en) | 2017-03-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11506650B2 (en) | Method for automatic quantitative statistical distribution characterization of dendrite structures in a full view field of metal materials | |
| CN102176228B (en) | Machine vision method for identifying dial plate information of multi-pointer instrument | |
| CN110659636B (en) | Pointer instrument reading identification method based on deep learning | |
| CN106485728B (en) | A kind of automatic measure of bar shaped main transformer oil level indicator | |
| KR101917515B1 (en) | Object recognition apparatus, objection recognition method, and program | |
| CN107239742B (en) | Method for calculating scale value of instrument pointer | |
| CN111950330A (en) | Pointer instrument indicating number detection method based on target detection | |
| CN104256882B (en) | Based on reconstituted tobacco ratio measuring method in the pipe tobacco of computer vision | |
| CN109447061A (en) | Reactor oil level indicator recognition methods based on crusing robot | |
| CN109784250A (en) | The localization method and device of automatically guiding trolley | |
| CN114220757A (en) | Wafer detection alignment method, device and system and computer medium | |
| CN108305242A (en) | A kind of intelligent visual detection method, system and device for industrial production line | |
| CN104198325A (en) | Method for measuring ratio of cut stem to cut tobacco based on computer vision | |
| CN103218809A (en) | Image measuring method of pearl length parameter | |
| CN103150558B (en) | A kind of operation of the display terminal based on machine vision responses match detection method | |
| CN204202563U (en) | Fish morphological parameters self-operated measuring unit | |
| CN105371769B (en) | A kind of method and device for measuring Dynamic Crack length | |
| CN113435442B (en) | Water level measuring method and device, water gauge and electronic equipment | |
| CN112561989B (en) | Recognition method for hoisting object in construction scene | |
| CN113822836B (en) | Method for marking an image | |
| CN104655041B (en) | A kind of industrial part contour line multi-feature extraction method of additional constraint condition | |
| CN103955929B (en) | Image local edge pattern and non-edge mode judging method and judgment means | |
| CN115615992B (en) | Fireproof brick size measurement and defect detection method | |
| CN116486212A (en) | Water gauge identification method, system and storage medium based on computer vision | |
| CN105389818A (en) | Method and system for positioning element |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | 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 |