CN106548211A - Transformer station's point for measuring temperature automatic identifying method and device based on infrared image - Google Patents
Transformer station's point for measuring temperature automatic identifying method and device based on infrared image Download PDFInfo
- Publication number
- CN106548211A CN106548211A CN201610982127.7A CN201610982127A CN106548211A CN 106548211 A CN106548211 A CN 106548211A CN 201610982127 A CN201610982127 A CN 201610982127A CN 106548211 A CN106548211 A CN 106548211A
- Authority
- CN
- China
- Prior art keywords
- infrared image
- image
- detected
- point
- measuring temperature
- 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
- 238000000034 method Methods 0.000 title claims abstract description 75
- 238000013210 evaluation model Methods 0.000 claims abstract description 4
- 230000009466 transformation Effects 0.000 claims description 12
- 239000000203 mixture Substances 0.000 claims description 5
- 238000002203 pretreatment Methods 0.000 claims description 4
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 238000011524 similarity measure Methods 0.000 claims description 2
- 238000007689 inspection Methods 0.000 description 8
- 238000012549 training Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 2
- 238000003331 infrared imaging Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 241000894007 species Species 0.000 description 2
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 description 1
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 230000003340 mental effect Effects 0.000 description 1
- 231100000572 poisoning Toxicity 0.000 description 1
- 230000000607 poisoning effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000001931 thermography Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/757—Matching configurations of points or features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/245—Classification techniques relating to the decision surface
- G06F18/2451—Classification techniques relating to the decision surface linear, e.g. hyperplane
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Radiation Pyrometers (AREA)
Abstract
The invention discloses a kind of transformer station's point for measuring temperature automatic identifying method and device based on infrared image.Methods described includes:Gather the infrared image to be detected of transformer station's point for measuring temperature to be monitored;Use pattern evaluation model determines the type of the infrared image to be detected, and the type includes feature protruding type infrared image and non-feature protruding type infrared image;When the type of the infrared image to be detected is characterized protruding type infrared image, using the image matching method of feature based, the infrared image to be detected is matched with the benchmark image of known transformer station's point for measuring temperature to be detected, to recognize transformer station's point for measuring temperature in the infrared image to be detected;When the type of the infrared image to be detected is non-feature protruding type infrared image, using the image matching method based on area information, the infrared image to be detected is matched with the benchmark image, to recognize transformer station's point for measuring temperature in the infrared image to be detected.
Description
Technical field
The present invention relates to computer vision, Digital Image Processing and mode identification technology, relate in particular to one kind
Transformer station's point for measuring temperature automatic identifying method and device based on infrared image.
Background technology
With electrical network continuous enlarging and develop rapidly, power equipment safety security problem becomes increasingly conspicuous, and traditional is artificial
Routine inspection mode, is affected by factors such as tour personnel's labor intensity, professional skill, the mental status and responsibility, missing inspection,
Flase drop situation happens occasionally, and is that national economy and people's lives bring hidden trouble, has been increasingly difficult to meet modern society pair
The needs of substation safety operation, unmanned and intelligent patrol detection will become the inexorable trend of transformer station's development.
Using the noncontact of infrared thermography, night monitoring, sensitivity is good, diagnosis efficiency is high the advantages of, by intelligence
Crusing robot carries out infrared imaging temperature measuring and fault alarm to high-tension transformation equipment heating sensitive part, is current research heat
Point.According to the demand of substation safety monitoring, point for measuring temperature to be monitored is determined in advance, these points for measuring temperature and relevant parameter have been made
It is bound in the task list of intelligent inspection robot in advance for presetting bit, in task implementation procedure, intelligent inspection robot enters
Row fixed point is patrolled and examined.Transformer station's point for measuring temperature to be monitored, generally selects the key equipment or portion that can reflect electrical equipment fault
Position, such as spark gap, transformer, sleeve pipe, disconnecting switch joint, contact of breaker etc..
Ideally, when crusing robot carries out fixed point according to the point for measuring temperature of advance bookbinding and patrols and examines, thermal infrared imager with
Point for measuring temperature is that imaging center gathers image, and point for measuring temperature is by positioned at the center of infrared image.However, in practical implementation,
When infrared imaging is carried out to point for measuring temperature by presetting bit, by crusing robot self poisoning precision, head pointing accuracy etc. it is various because
The impact of element, point for measuring temperature often deviates the center of infrared image, and is affected by imaged viewing angle precision, and what is gathered is red
Outer image is also typically present certain deformation compared with ideal situation;And due to the randomness of error, when patrolling and examining every time, adopted
The departure degree of the deformation extent and same point for measuring temperature of collection infrared image is also not quite similar, therefore, when patrolling and examining every time, it is required for
Transformer station's point for measuring temperature is identified from the infrared image of collection.
Additionally, the power equipment species that transformer station's point for measuring temperature is covered is various, profile is varied, it is difficult to from infrared image
The common trait of all kinds of points for measuring temperature of transformer station is extracted, it is current many using man-machine interactively mode, specify in infrared image or indicate
Go out point for measuring temperature to be monitored, the inefficient working method of this repeatability have impact on transformer station's unmanned and automated process
Propulsion.
The content of the invention
In order to overcome, current infrared image transformer station point for measuring temperature technology of identification man-machine interactively is loaded down with trivial details, automaticity is low, labor
The strong deficiency of dynamic repeatability, the present invention utilize computer vision, image procossing and mode identification technology, there is provided a kind of automatical and efficient
The point for measuring temperature recognition methodss of infrared image transformer station and device.
A kind of transformer station's point for measuring temperature automatic identifying method based on infrared image, bag are provided according to an aspect of the present invention
Include:
Gather the infrared image to be detected of transformer station's point for measuring temperature to be monitored;
Use pattern evaluation model determines the type of the infrared image to be detected, and the type includes that feature protruding type is red
Outer image and non-feature protruding type infrared image;
When the type of the infrared image to be detected is characterized protruding type infrared image, using the image of feature based
Method of completing the square, is matched with the benchmark image of known transformer station's point for measuring temperature to be detected to the infrared image to be detected,
To recognize transformer station's point for measuring temperature in the infrared image to be detected;
When the type of the infrared image to be detected is non-feature protruding type infrared image, using based on area information
Image matching method, is matched with the benchmark image to the infrared image to be detected, described to be detected infrared to recognize
Transformer station's point for measuring temperature in image.
Wherein, it is described when the type of the infrared image to be detected is characterized protruding type infrared image, using based on spy
The image matching method levied, the benchmark image to the infrared image to be detected and known transformer station's point for measuring temperature to be detected
Matched, to recognize transformer station's point for measuring temperature in the infrared image to be detected, including:
Extract the characteristics of image of the infrared image to be detected and the benchmark image;
The character pair between the infrared image to be detected and the benchmark image, shape are obtained using feature matching method
Into feature same place to collection;The feature same place includes multiple feature same places pair to collection, and each feature same place is to including
The character pair of the infrared image to be measured and the benchmark image;
Training is got between the infrared image to be detected and the benchmark image using the feature same place
Image transform model;
According to described image transformation model and the benchmark image, identify that transformer station surveys in the infrared image to be measured
Warm spot region.
Wherein, it is described when the type of the infrared image to be detected is non-feature protruding type infrared image, adopt and be based on
The image matching method of area information, is matched with the benchmark image to the infrared image to be detected, described to recognize
Transformer station's point for measuring temperature in infrared image to be detected, including:
Under each yardstick and rotation angle parameter, the benchmark image is calculated based on sliding window method to be checked with described
Survey similarity of the infrared image at correspondence sliding window;
The benchmark image and infrared image under statistics current scale and rotation angle parameter, at the different sliding windows
Maximum similarity;
Obtain under all yardsticks and rotation angle parameter, the maximum similarity of the benchmark image and infrared image, and will
Corresponding sliding window position, yardstick and rotation angle parameter as the benchmark image with the infrared image to be detected most
Good match parameter;
According to the optimum matching parameter, transformer station's point for measuring temperature area is identified on the infrared image to be detected
Domain.
Wherein, it is described according to the optimum matching parameter, the transformer station is identified on the infrared image to be detected
Point for measuring temperature region, including:
According to the optimum matching parameter, four summits of the benchmark image are found in the infrared image to be detected
Four corresponding vertexs, and four corresponding vertexs are coupled together the closed area of composition, it is identified as described to be detected infrared
Transformer station's point for measuring temperature region on image.
Wherein, methods described also includes:
Gather the infrared image to be calibrated of transformer station's point for measuring temperature to be detected;
Pretreatment is carried out to the infrared image to be calibrated;
The region of transformer station's point for measuring temperature to be detected is demarcated on infrared image described to be calibrated after the pre-treatment, and then
Generate the benchmark image of transformer station's point for measuring temperature to be detected.
According to a further aspect of the invention, there is provided a kind of transformer station's point for measuring temperature automatic identification equipment based on infrared image,
Including:
Acquisition module, is configured to gather the infrared image to be detected of transformer station's point for measuring temperature to be monitored;
Sort module, is configured to, with the type that Types Assessment model determines the infrared image to be detected, the class
Type includes feature protruding type infrared image and non-feature protruding type infrared image;
First identification module, the type for being configured as the infrared image to be detected are characterized protruding type infrared image
When, using the image matching method of feature based, the infrared image to be detected is surveyed with the known transformer station to be detected
The benchmark image of warm spot is matched, to recognize transformer station's point for measuring temperature in the infrared image to be detected;
Second identification module, the type for being configured as the infrared image to be detected are non-feature protruding type infrared image
When, using the image matching method based on area information, the infrared image to be detected is matched with the benchmark image,
To recognize transformer station's point for measuring temperature in the infrared image to be detected.
Wherein, first identification module, including:
Extracting sub-module, is configured to extract the characteristics of image of the infrared image to be detected and the benchmark image;
Matched sub-block, is configured to, with feature matching method and obtains the infrared image to be detected and the reference map
Character pair as between, forms feature same place to collection;The feature same place includes multiple feature same places pair to collection, often
Individual feature same place is to the character pair including the infrared image to be measured and the benchmark image;
Submodule is solved, the feature same place is configured to, with and the infrared image to be detected and described is solved to collection
Image transform model between benchmark image;
First identification submodule, is configured to according to described image transformation model and the benchmark image, described to be measured
Transformer station's point for measuring temperature region is identified in infrared image.
Wherein, second identification module, including:
Similarity Measure submodule, is configured under each yardstick and rotation angle parameter, based on sliding window method
Calculate similarity of the benchmark image with the infrared image to be detected at corresponding sliding window;
Statistic submodule, is configured to count under current scale and rotation angle parameter, described at different sliding windows
The maximum similarity of benchmark image and infrared image;
Acquisition submodule, is configured to obtain under all yardsticks and rotation angle parameter, the benchmark image and infrared figure
The maximum similarity of picture, and using corresponding sliding window position, yardstick and rotation angle parameter as the benchmark image and institute
State the optimum matching parameter of infrared image to be detected;
Second identification submodule, is configured to according to the optimum matching parameter, knows on the infrared image to be detected
Do not go out transformer station's point for measuring temperature region.
Wherein, the second identification submodule, including:
3rd identification submodule, is configured to according to the optimum matching parameter, finds four summits of the benchmark image
Four corresponding vertexs in the infrared image to be detected, and four corresponding vertexs are coupled together the enclosed area of composition
Domain, the transformer station's point for measuring temperature region being identified as on the infrared image to be detected.
Wherein, described device also includes:
Second acquisition module, is configured to gather the infrared image to be calibrated of transformer station's point for measuring temperature to be detected;
Pretreatment module, is configured to carry out pretreatment to the infrared image to be calibrated;
Demarcating module, is configured to the transformer station to be detected is demarcated on infrared image described to be calibrated after the pre-treatment
The region of point for measuring temperature, and then generate the benchmark image of transformer station's point for measuring temperature to be detected.
Transformer station's point for measuring temperature automatic identifying method based on infrared image and device of the present invention, with intelligent inspection robot
For application platform, the converting station electric power equipment infrared image gathered by cordless during patrolling and examining to fixed point, using meter
Calculation machine vision and image processing techniquess, realize the transformer station's point for measuring temperature automatic identification in infrared image.
The point for measuring temperature of transformer station, can cover various key equipments that can reflect electrical equipment fault or position, such as keep away
Thunder device, transformer, sleeve pipe, disconnecting switch joint, contact of breaker etc..These transformer station's point for measuring temperature species are various, various shapes,
Show in infrared image different.Thinking of the present invention using matching, builds point for measuring temperature reference image library in advance, by realizing surveying
Warm spot benchmark image is matched with the infrared image of Real-time Collection, is difficult the difficult point extracted to solve point for measuring temperature common trait.
The present invention proposes the thinking of image content-based adaptive analysis, is judged by the analysis of image type, selects automatically
Corresponding image matching method is selected, the automatic identification of infrared image transformer station point for measuring temperature is realized.It is this " divide and rule, to the ill under
The process strategy of medicine ", can both ensure the point for measuring temperature recognition effect of feature protruding type infrared image, it is also ensured that non-feature is dashed forward
Go out the recognition effect of type infrared image.
The method and apparatus that the present invention is provided, can support the automatic identification of various types of transformer station's points for measuring temperature, fit
Answer scope wide;The analysis judgement of image type is automatically obtained, for different types of infrared image, different methods is automatically selected
The identification of transformer station's point for measuring temperature is realized, can be ensured;Once bind, for a long time
Effectively;The dilatation of transformer station's point for measuring temperature is supported, even if increasing new infrared measurement of temperature point, without redesign algorithm.Infrared figure
As the automatic identification of transformer station's point for measuring temperature, it is subsequently to be carried based on the point for measuring temperature temperature extraction of infrared image, accident analysis and warning
Important technical support is supplied.
Description of the drawings
Fig. 1 is multiple transformer station's point for measuring temperature schematic diagrams to be monitored;
Fig. 2 is the transformer station's point for measuring temperature recognition methodss flow chart in the present invention based on infrared image;
Fig. 3 (a) is the benchmark image schematic diagram that the present invention is demarcated;
Fig. 3 (b) is the transformer station's point for measuring temperature area schematic identified in identification process of the present invention;
Fig. 4 is transformer station's point for measuring temperature identification process figure of feature protruding type infrared image in the present invention;
Fig. 5 is transformer station's point for measuring temperature identification process figure of non-feature protruding type infrared image in the present invention.
Specific embodiment
The present invention is described in detail below in conjunction with accompanying drawing, it is noted that described embodiment is only intended to just
In the understanding of the present invention, and any restriction effect is not played to which.
The method that the present invention is used can be installed on personal computer, industrial computer and server in the form of software and be held
OK, also method can be made embedded chip to be implemented in hardware.
The present invention proposes a kind of transformer station's point for measuring temperature automatic identifying method based on infrared image, including:
Gather the infrared image to be detected of transformer station's point for measuring temperature to be monitored;
Use pattern evaluation model determines the type of the infrared image to be detected, and the type includes that feature protruding type is red
Outer image and non-feature protruding type infrared image;
When the type of the infrared image to be detected is characterized protruding type infrared image, using the image of feature based
Method of completing the square, is matched with the benchmark image of known transformer station's point for measuring temperature to be detected to the infrared image to be detected,
To recognize transformer station's point for measuring temperature in the infrared image to be detected;
When the type of the infrared image to be detected is non-feature protruding type infrared image, using based on area information
Image matching method, is matched with the benchmark image to the infrared image to be detected, described to be detected infrared to recognize
Transformer station's point for measuring temperature in image.
Technical scheme is discussed in detail below by specific embodiment.
In one embodiment, the present invention proposes a kind of infrared image transformer station point for measuring temperature towards intelligent inspection robot
Recognition methodss.
Intelligent inspection robot gathers to be monitored survey by presetting bit when carrying out fixed point by programme path to transformer station and patrolling and examining
The infrared image of warm spot, includes transformer station's point for measuring temperature to be monitored, and the benchmark image of the point for measuring temperature is in the infrared image
It is bound in infrared image transformer station point for measuring temperature identifying system in advance.Some changes to be monitored of the present embodiment are given in Fig. 1
Power station point for measuring temperature schematic diagram.
Fig. 2 gives the transformer station's point for measuring temperature recognition methodss flow chart based on infrared image of embodiment.As shown in Fig. 2
Based on transformer station's point for measuring temperature recognition methodss handling process of infrared image, including two big process of registration process and identification process.
Registration process is as follows:Before transformer station's infrared image point for measuring temperature automatic identification is carried out, need in advance to infrared figure
As transformer station's point for measuring temperature identifying system registers transformer station's point for measuring temperature to be monitored, the infrared image of transformer station's point for measuring temperature is gathered, and
Infrared image to gathering makees pretreatment, is then manually demarcated, and will demarcate reference map of the region out as point for measuring temperature
As (as shown in Fig. 3 (a)), it is stored in point for measuring temperature reference image library.Transformer station's point for measuring temperature of Real-time Collection during to patrolling and examining
When infrared image carries out point for measuring temperature identification, the benchmark image of the point for measuring temperature according to index, can be retrieved.
When intelligent inspection robot carries out fixed point patrols and examines, gather after the infrared image of transformer station's point for measuring temperature by presetting bit,
Just enter identification process:
In order to reduce effect of noise, the infrared image to gathering first carries out pretreatment, and such as image denoising and image increases
It is strong etc.;
Then, image type judgement is carried out to pretreated infrared image, according to image clearly degree, by infrared image
It is divided into feature protruding type infrared image and non-feature protruding type infrared image.The spectral discrimination process, can adopt supervised training
The method of study, such as SVM, integrated learning approach, deep learning method etc..Below, with based on integrated study Boosting methods
Image type decision process as a example by describe in detail.First, substantial amounts of transformer station's point for measuring temperature infrared image is gathered as instruction
Practice sample, by way of artificial demarcation, these training samples are divided and is characterized protruding type infrared image and non-feature protruding type
Infrared image;Then, various characteristics of image are extracted to training sample, is trained using Boosting methods, from various images
Effective feature is selected in feature, and obtains corresponding grader, the validity feature for obtaining and grader are trained as supervision
The learning outcome of training.To train the validity feature for obtaining and grader as foundation, for the thermometric for treating that image type judges
Point infrared image, extracts validity feature, and the classification obtained using supervised training is classified, and point for measuring temperature infrared image is divided into
Feature protruding type infrared image or non-feature protruding type infrared image.
For feature protruding type infrared image, which often shows as, and details is abundant, edge clear, can adopt feature based
Image matching method, by the local message at control point, realizes matching for infrared image and point for measuring temperature benchmark image;For non-spy
Levy protruding type infrared image, in view of its contrast is weak, edge details are few, shape and structure not substantially or many factors such as repeatability is strong
Restriction, it tends to be difficult to extract reliable and stable control point, thus can using the image matching method based on area information, lead to
Global information is crossed, matching for infrared image and point for measuring temperature benchmark image is realized;
Finally, point for measuring temperature region is indicated automatically with significantly mark on infrared image, shown in such as Fig. 3 (b), closing
Black wire frame region representation point for measuring temperature region.
Fig. 4 gives the flow process of transformer station's point for measuring temperature identification process of the feature protruding type infrared image of one embodiment
Figure.Transformer station's point for measuring temperature identification of the feature protruding type infrared image in the present invention can be realized in the following way:
Characteristics of image is extracted respectively to infrared image to be detected and point for measuring temperature benchmark infrared image during first to patrolling and examining,
The characteristics of image for being extracted includes a feature and line feature, the such as feature such as angle point, line cross point, edge, profile, enclosed region,
Feature set is formed, the feature of selective extraction meets in the real-time infrared image for obtaining as far as possible and is respectively provided with benchmark infrared image
Significance, and quantity enough, distribution uniform.
Then, in infrared image to be detected and the corresponding feature focus utilization Feature Correspondence Algorithm of point for measuring temperature benchmark image,
The feature that there is corresponding relation is found out, in order to describe conveniently, will be deposited in infrared image to be detected and point for measuring temperature benchmark image below
In the feature of corresponding relation, it is referred to as being characterized same place pair.
Select the images relations transformation model between infrared image to be detected and point for measuring temperature benchmark image, such as affine transformation mould
Type, by the feature same place pair obtained during characteristic matching, asks for the parameter of images relations transformation model.
Finally, according to transformation model and point for measuring temperature benchmark image, transformer station's point for measuring temperature region is oriented in infrared image,
Realize the identification of infrared image transformer station point for measuring temperature.Implement process as follows:Upper left, upper right to point for measuring temperature benchmark image,
Lower-left, the summit of bottom right four, according to the infrared image to be detected asked for and point for measuring temperature benchmark image transformation model, calculate this four
Coordinate position of the individual picture point on infrared image to be detected, this line of four picture point on infrared image to be detected are formed
Closed area, be the transformer station's point for measuring temperature region in infrared image to be detected, i.e. by images match mode, realize infrared
The identification of image multiclass transformer station point for measuring temperature.
Fig. 5 gives transformer station's point for measuring temperature identification process of the non-feature protruding type infrared image of one embodiment.This
Transformer station's point for measuring temperature identification of the non-feature protruding type infrared image in bright can be by realizing as follows:In arbitrary yardstick
Under rotation angle parameter, infrared image is carried out into corresponding scaling and rotation first, the infrared image after being converted,
On infrared image after the transformation, continuous sliding window, to the infrared image sliding window region after conversion and point for measuring temperature base
Quasi- image carries out similarity measurement, e.g., can be using correlation coefficient as measuring similarity;Count the yardstick and anglec of rotation ginseng
Maximum similarity under several at the different sliding windows;Find under all yardsticks and rotation angle parameter, point for measuring temperature benchmark image
With the maximum similarity of infrared image, and corresponding position, yardstick and the anglec of rotation are recorded, position, yardstick and the anglec of rotation
This 3 parameters are collectively as point for measuring temperature benchmark image and the optimum matching parameter of infrared image;According to position, yardstick and the anglec of rotation
The optimum matching parameter that this 3 parameters are collectively constituted is spent, the inverse transformation process for scaling by infrared image and rotating finds thermometric
Correspondence position of four summits of point benchmark image in original infrared image, the enclosed area of four vertex correspondence position line compositions
Domain, is the transformer station's point for measuring temperature region on infrared image.
The above, the only specific embodiment in the present invention, but protection scope of the present invention is not limited thereto, and appoints
What be familiar with the people of the technology disclosed herein technical scope in, it will be appreciated that the conversion expected or replacement, should all cover
The present invention include within the scope of, therefore, protection scope of the present invention should be defined by the protection domain of claims.
Claims (10)
1. a kind of transformer station's point for measuring temperature automatic identifying method based on infrared image, it is characterised in that include:
Gather the infrared image to be detected of transformer station's point for measuring temperature to be monitored;
Use pattern evaluation model determines the type of the infrared image to be detected, and the type includes the infrared figure of feature protruding type
Picture and non-feature protruding type infrared image;
When the type of the infrared image to be detected is characterized protruding type infrared image, using the images match side of feature based
Method, is matched with the benchmark image of known transformer station's point for measuring temperature to be detected to the infrared image to be detected, to know
Transformer station's point for measuring temperature in not described infrared image to be detected;
When the type of the infrared image to be detected is non-feature protruding type infrared image, using the image based on area information
Matching process, is matched with the benchmark image to the infrared image to be detected, to recognize the infrared image to be detected
Middle transformer station's point for measuring temperature.
2. the method for claim 1, it is characterised in that it is described when the type of the infrared image to be detected be characterized it is prominent
When going out type infrared image, using the image matching method of feature based, the infrared image to be detected described is treated with known
The benchmark image of detection transformer station point for measuring temperature is matched, to recognize transformer station's point for measuring temperature in the infrared image to be detected, bag
Include:
Extract the characteristics of image of the infrared image to be detected and the benchmark image;
The character pair between the infrared image to be detected and the benchmark image is obtained using feature matching method, forms special
Same place is levied to collection;The feature same place includes multiple feature same places pair to collection, and each feature same place is to including described
The character pair of infrared image to be measured and the benchmark image;
The image conversion between the infrared image to be detected and the benchmark image is solved using the feature same place to collection
Model;
According to described image transformation model and the benchmark image, transformer station's point for measuring temperature is identified in the infrared image to be measured
Region.
3. the method for claim 1, it is characterised in that described when the type of the infrared image to be detected is non-feature
During protruding type infrared image, using the image matching method based on area information, to the infrared image to be detected and the base
Quasi- image is matched, to recognize transformer station's point for measuring temperature in the infrared image to be detected, including:
Under each yardstick and rotation angle parameter, the benchmark image is calculated based on sliding window method to be detected red with described
Similarity of the outer image at correspondence sliding window;
The maximum of the benchmark image and infrared image under statistics current scale and rotation angle parameter, at the different sliding windows
Similarity;
Obtain under all yardsticks and rotation angle parameter, the maximum similarity of the benchmark image and infrared image, and will correspondence
Sliding window position, yardstick and rotation angle parameter be used as optimal of the benchmark image and the infrared image to be detected
With parameter;
According to the optimum matching parameter, transformer station's point for measuring temperature region is identified on the infrared image to be detected.
4. method as claimed in claim 3, it is characterised in that described according to the optimum matching parameter, described to be detected
Transformer station's point for measuring temperature region is identified on infrared image, including:
According to the optimum matching parameter, four in the infrared image to be detected, four summits of the benchmark image are found
Corresponding vertex, and four corresponding vertexs are coupled together the closed area of composition, it is identified as the infrared image to be detected
On transformer station's point for measuring temperature region.
5. the method for claim 1, it is characterised in that also include:
Gather the infrared image to be calibrated of transformer station's point for measuring temperature to be detected;
Pretreatment is carried out to the infrared image to be calibrated;
The region of transformer station's point for measuring temperature to be detected is demarcated on infrared image described to be calibrated after the pre-treatment, and then is generated
The benchmark image of transformer station's point for measuring temperature to be detected.
6. a kind of transformer station's point for measuring temperature automatic identification equipment based on infrared image, it is characterised in that include:
Acquisition module, is configured to gather the infrared image to be detected of transformer station's point for measuring temperature to be monitored;
Sort module, is configured to, with the type that Types Assessment model determines the infrared image to be detected, the type bag
Include feature protruding type infrared image and non-feature protruding type infrared image;
First identification module, when being configured as the type of the infrared image to be detected and being characterized protruding type infrared image, adopts
With the image matching method of feature based, to the infrared image to be detected and known transformer station's point for measuring temperature to be detected
Benchmark image is matched, to recognize transformer station's point for measuring temperature in the infrared image to be detected;
Second identification module, when the type for being configured as the infrared image to be detected is non-feature protruding type infrared image,
Using the image matching method based on area information, the infrared image to be detected is matched with the benchmark image, with
Recognize transformer station's point for measuring temperature in the infrared image to be detected.
7. device as claimed in claim 6, it is characterised in that first identification module, including:
Extracting sub-module, is configured to extract the characteristics of image of the infrared image to be detected and the benchmark image;
Matched sub-block, be configured to, with feature matching method obtain the infrared image to be detected and the benchmark image it
Between character pair, formed feature same place to collection;The feature same place includes multiple feature same places pair to collection, and each is special
Same place is levied to the character pair including the infrared image to be measured and the benchmark image;
Submodule is solved, the feature same place is configured to, with and the infrared image to be detected and the benchmark is solved to collection
Image transform model between image;
First identification submodule, is configured to according to described image transformation model and the benchmark image, described to be measured infrared
Transformer station's point for measuring temperature region is identified in image.
8. device as claimed in claim 6, it is characterised in that second identification module, including:
Similarity Measure submodule, is configured under each yardstick and rotation angle parameter, is calculated based on sliding window method
Similarity of the benchmark image with the infrared image to be detected at corresponding sliding window;
Statistic submodule, is configured to count under current scale and rotation angle parameter, the benchmark at the different sliding windows
The maximum similarity of image and infrared image;
Acquisition submodule, is configured to obtain under all yardsticks and rotation angle parameter, the benchmark image and infrared image
Maximum similarity, and corresponding sliding window position, yardstick and rotation angle parameter are treated as the benchmark image with described
The optimum matching parameter of detection infrared image;
Second identification submodule, is configured to according to the optimum matching parameter, identifies on the infrared image to be detected
Transformer station's point for measuring temperature region.
9. device as claimed in claim 8, it is characterised in that the second identification submodule, including:
3rd identification submodule, is configured to according to the optimum matching parameter, finds four summits of the benchmark image in institute
State four corresponding vertexs in infrared image to be detected, and four corresponding vertexs coupled together the closed area of composition,
The transformer station's point for measuring temperature region being identified as on the infrared image to be detected.
10. device as claimed in claim 6, it is characterised in that also include:
Second acquisition module, is configured to gather the infrared image to be calibrated of transformer station's point for measuring temperature to be detected;
Pretreatment module, is configured to carry out pretreatment to the infrared image to be calibrated;
Demarcating module, is configured to transformer station's thermometric to be detected is demarcated on infrared image described to be calibrated after the pre-treatment
The region of point, and then generate the benchmark image of transformer station's point for measuring temperature to be detected.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610982127.7A CN106548211A (en) | 2016-11-08 | 2016-11-08 | Transformer station's point for measuring temperature automatic identifying method and device based on infrared image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610982127.7A CN106548211A (en) | 2016-11-08 | 2016-11-08 | Transformer station's point for measuring temperature automatic identifying method and device based on infrared image |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106548211A true CN106548211A (en) | 2017-03-29 |
Family
ID=58395550
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610982127.7A Pending CN106548211A (en) | 2016-11-08 | 2016-11-08 | Transformer station's point for measuring temperature automatic identifying method and device based on infrared image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106548211A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108764057A (en) * | 2018-05-03 | 2018-11-06 | 武汉高德智感科技有限公司 | A kind of far infrared human type of face detection method and system based on deep learning |
CN109855739A (en) * | 2019-01-04 | 2019-06-07 | 三峡大学 | Power equipment infrared measurement of temperature method and device based on affine transformation |
CN111061208A (en) * | 2019-12-31 | 2020-04-24 | 福建睿思特科技股份有限公司 | Night infrared identification and illumination system of transformer substation |
CN111160305A (en) * | 2019-12-31 | 2020-05-15 | 武汉工程大学 | Transformer bushing abnormity monitoring method and computer readable medium |
CN116379937A (en) * | 2023-06-06 | 2023-07-04 | 武汉大学 | Method and device for monitoring shaking of power transmission tower |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102798469A (en) * | 2012-06-29 | 2012-11-28 | 山东鲁能智能技术有限公司 | Angular-point-characteristic-based method for accurately measuring temperature of electrical equipment |
CN104331521A (en) * | 2014-11-27 | 2015-02-04 | 国家电网公司 | Automatic power transformation equipment abnormity identification method based on image processing |
CN105512623A (en) * | 2015-12-02 | 2016-04-20 | 吉林大学 | Foggy-day driving visual enhancement and visibility early warning system and method based on multiple sensors |
CN105844257A (en) * | 2016-04-11 | 2016-08-10 | 吉林大学 | Early warning system based on machine vision driving-in-fog road denoter missing and early warning method |
CN103617469B (en) * | 2013-12-18 | 2016-09-14 | 重庆大学 | Power system device failure prediction method and system |
-
2016
- 2016-11-08 CN CN201610982127.7A patent/CN106548211A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102798469A (en) * | 2012-06-29 | 2012-11-28 | 山东鲁能智能技术有限公司 | Angular-point-characteristic-based method for accurately measuring temperature of electrical equipment |
CN103617469B (en) * | 2013-12-18 | 2016-09-14 | 重庆大学 | Power system device failure prediction method and system |
CN104331521A (en) * | 2014-11-27 | 2015-02-04 | 国家电网公司 | Automatic power transformation equipment abnormity identification method based on image processing |
CN105512623A (en) * | 2015-12-02 | 2016-04-20 | 吉林大学 | Foggy-day driving visual enhancement and visibility early warning system and method based on multiple sensors |
CN105844257A (en) * | 2016-04-11 | 2016-08-10 | 吉林大学 | Early warning system based on machine vision driving-in-fog road denoter missing and early warning method |
Non-Patent Citations (2)
Title |
---|
刘强: "多源信息融合框架下辅助导航***的景象匹配算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
张兴华: "变电站红外遥视图像配准研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108764057A (en) * | 2018-05-03 | 2018-11-06 | 武汉高德智感科技有限公司 | A kind of far infrared human type of face detection method and system based on deep learning |
CN109855739A (en) * | 2019-01-04 | 2019-06-07 | 三峡大学 | Power equipment infrared measurement of temperature method and device based on affine transformation |
CN111061208A (en) * | 2019-12-31 | 2020-04-24 | 福建睿思特科技股份有限公司 | Night infrared identification and illumination system of transformer substation |
CN111160305A (en) * | 2019-12-31 | 2020-05-15 | 武汉工程大学 | Transformer bushing abnormity monitoring method and computer readable medium |
CN111061208B (en) * | 2019-12-31 | 2020-11-20 | 福建睿思特科技股份有限公司 | Night infrared identification and illumination system of transformer substation |
CN116379937A (en) * | 2023-06-06 | 2023-07-04 | 武汉大学 | Method and device for monitoring shaking of power transmission tower |
CN116379937B (en) * | 2023-06-06 | 2023-11-24 | 武汉大学 | Method and device for monitoring shaking of power transmission tower |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106548211A (en) | Transformer station's point for measuring temperature automatic identifying method and device based on infrared image | |
CN103901072B (en) | Method for diagnosing equipment overheating defects by utilizing infrared spectrum analysis | |
CN109101906A (en) | A kind of converting station electric power equipment infrared image exception real-time detection method and device | |
CN110850723B (en) | Fault diagnosis and positioning method based on transformer substation inspection robot system | |
CN111242123B (en) | Power equipment fault diagnosis method based on infrared image | |
CN111751002B (en) | Intelligent charged equipment fault diagnosis method based on infrared thermal imaging | |
CN111624229B (en) | Intelligent charged equipment fault diagnosis method based on infrared imaging | |
CN109800697B (en) | Transformer target detection and appearance defect identification method based on VGG-net style migration | |
CN108537154A (en) | Transmission line of electricity Bird's Nest recognition methods based on HOG features and machine learning | |
CN112085089A (en) | Intelligent temperature measurement method for transformer substation equipment based on deep learning algorithm | |
CN104994334A (en) | Automatic substation monitoring method based on real-time video | |
CN109034272A (en) | A kind of transmission line of electricity heat generating components automatic identifying method | |
CN112697798B (en) | Infrared image-oriented diagnosis method and device for current-induced thermal defects of power transformation equipment | |
CN112734698A (en) | Cable terminal abnormity diagnosis method and device based on infrared image | |
CN111738156A (en) | Intelligent inspection management method and system for state of high-voltage switchgear | |
CN108280449A (en) | Power equipment image collecting method based on multispectral sensor group | |
CN113418617A (en) | Three-dimensional temperature field measuring and diagnosing method for box type transformer of wind power plant | |
Wang et al. | Online monitoring of electrical equipment condition based on infrared image temperature data visualization | |
CN108898187A (en) | A kind of method and device of automatic identification power distribution room indicating equipment image | |
CN112818806A (en) | Transformer substation inspection robot auxiliary navigation method based on deep learning | |
EP3706270B1 (en) | Artificial intelligence monitoring system using infrared images to identify hotspots in a switchgear | |
CN106248215A (en) | A kind of real-time calculating system of power transformer critical component running temperature | |
Taib et al. | Thermal imaging for qualitative-based measurements of thermal anomalies in electrical components | |
US20210397837A1 (en) | Apparatus for monitoring a switchgear | |
CN116188505A (en) | Insulation power equipment on-line monitoring system based on ultraviolet and infrared collaborative detection |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170329 |