CN106778723B - A kind of pneumatic equipment bladess surface image extracting method in complex background environment - Google Patents
A kind of pneumatic equipment bladess surface image extracting method in complex background environment Download PDFInfo
- Publication number
- CN106778723B CN106778723B CN201611065764.4A CN201611065764A CN106778723B CN 106778723 B CN106778723 B CN 106778723B CN 201611065764 A CN201611065764 A CN 201611065764A CN 106778723 B CN106778723 B CN 106778723B
- Authority
- CN
- China
- Prior art keywords
- image
- blade
- mark
- camera
- leaf
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30164—Workpiece; Machine component
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses the pneumatic equipment bladess surface image extracting method in a kind of complex background environment, it comprises the following steps:(1) square black and white case marking figure is set respectively at the windward side of each blade of wind energy conversion system, and the position of stickup is plane and parallel with blade windward side, and then utilization marks the posture of figure to replace Leaf orientation;(2) wind energy conversion system operation image is acquired using camera, and by the Image Real-time Transmission collected into industrial computer;(3) image information collected to camera carries out image procossing to obtain blade surface image, and obtains pneumatic equipment bladess information according to the mark figure on blade.The extraction to running pneumatic equipment bladess image in wild environment can be achieved in the present invention, simplifies analyzing and processing process, accelerates analyze speed.
Description
Technical field
The invention belongs to wind energy conversion system operation maintenance and technical field of image processing, more particularly, to a kind of complex background
Pneumatic equipment bladess surface image extracting method in environment.
Background technology
The failure on blade of wind-driven generator surface includes crackle, sand holes, surface contamination, icing etc., if leaving its development
Serious consequence, the operation safety of leaf destruction, decrease in power generation efficiency and influence driving-chain etc. can be caused.Pass through shooting
Equipment is monitored to blade surface image, initial stage it can be found in failure and being capable of reasonable arrangement maintenance, it is to avoid enter one
Step loss.
How patent CN201410172548.4, CN201410157662.X, CN201410171787.8 are devised to wind
Power machine blade surface carries out region division and extracts the species of blade surface failure and the method for position, and it is assuming that blade
Surface image is carried out on the basis of having obtained, and does not involve how to effectively extract blade figure from the background image of field
The content of picture.And the background rejecting effect of leaf image has important shadow for next step analysis blade surface failure mode
Ring, because the complexity of wild environment, and different pneumatic equipment bladess images appear in doing caused by field of view simultaneously
The reason such as disturbing causes to extract leaf image very difficult.
Carried out at present for object identification is general using substantial amounts of subject image come the method for training image grader, the party
Method needs substantial amounts of view data and prolonged training in order to obtain higher precision.And for pneumatic equipment bladess
Speech, first acquisition view data is relatively difficult, secondly because pneumatic equipment bladess length is larger, it is difficult to ensure that shooting in identification process
Complete leaf image is resulted in the visual field, and then causes the decline of accuracy of identification.
The content of the invention
For the disadvantages described above or Improvement requirement of prior art, the invention provides the wind-force in a kind of complex background environment
Machine blade surface image extraction method, the characteristics of it is for blade local environment complexity, is pasted by blade surface and includes blade
The mark figure of information is positioned to leaf position, primarily determines that the position of Region Of Interest, is realized to running wind in wild environment
The extraction of power machine leaf image, its superficial failure species is analyzed for corresponding computer vision algorithms make and occurs position, is simplified
Analyzing and processing process, accelerate analyze speed.
To achieve the above object, the pneumatic equipment bladess surface image that the present invention is proposed in a kind of complex background environment is extracted
Method, comprises the following steps:
(1) blade surface is set to mark:Square black and white case marking is set respectively at the windward side of each blade of wind energy conversion system
Figure, the position of stickup is plane and parallel with blade windward side, and then utilization marks the posture of figure to replace Leaf orientation;
(2) leaf image is gathered:Wind energy conversion system operation image is acquired using camera, and the image collected is real
When transmit into industrial computer;
(3) blade information is recognized:The image information collected to camera carries out image procossing to obtain blade surface figure
Picture, and pneumatic equipment bladess information is obtained according to the mark figure on blade.
As it is further preferred that the step (3) specifically includes following steps:
(3.1) gradation of image is changed:The image that camera is collected carries out gradation of image conversion, by color-map representation into
Black and white;
(3.2) image binaryzation:Each pixel in image after gradation conversion is become into black or white, wherein black
Pixel value be 0, white pixel value is 255, i.e., when certain point pixel is less than or equal to given threshold, the pixel is set to
0, it is on the contrary then be set to 255;
(3.3) profile is monitored:The area of black or white pixel UNICOM in image in obtaining step (3.2) after binaryzation
Domain, finds the profile in UNICOM region, obtains multiple profiles;
(3.4) candidates are searched for:Polygonal segments are carried out to the multiple profiles obtained in step (3.3), by number of vertex
Candidate is left as 4 polygonal region;
(3.5) selection markers:The polygonal region of each candidate is converted into square area using affine transformation, then
Mesh generation is carried out to square area, n × n grid is divided into, if a hoop net lattice of periphery are black, then it is assumed that
The area image is schemed for mark;
(3.6) leaf image and blade information are obtained:
According to size, installation site and the blade dimensions of mark figure, affine transformation tentatively intercepts the area where leaf image
Domain, then extracts blade border using thresholding method and obtains leaf image, the leaf is obtained according to the mark figure on leaf image
The information of piece.
If as it is further preferred that tentatively interception leaf image where region exceed camera field range when,
Zoom then is carried out to camera, expands field range, is specifically controlled using following control logic:
1) after camera collection image, graphical analysis is carried out, next step is carried out if mark figure is included in image, if
No then adjustment camera angle and focal length start to gather information, until mark figure is appeared in camera acquired image;
2) blade surface analysis is carried out to the image comprising mark figure, if leaf image is all in the image range of collection
It is interior, then leaf image is intercepted, if not existing, zoom is carried out centered on marking figure to camera, multiple blades are found
Respective mark figure, then swinging camera makes its central region be center determined by multiple marks, and continuation zoom, which expands, to be regarded
Wild scope makes multiple blades all include in the picture.
As it is further preferred that the individual small positive squares of n × n for marking figure by the line segmentation parallel to side into decile
Region is constituted, and is made a circle outside mark figure and be designed to black, by (n-2) × (n- of the inside made a circle except setting
2) small square color is encoded to mark, and the coding is corresponding with blade information, it is preferred that 5≤n≤10.
As it is further preferred that the black bars or the white square length of side of the composition mark figure are designed as being more than 10 pictures
Element.
In general, possess following compared with prior art, mainly by the contemplated above technical scheme of the present invention
Technological merit:
1. the present invention compared with training the method for pneumatic equipment bladess grader with can substantially reduce workload, due to grader
Several thousand sheets subject image to be identified is needed as input picture, so sample finds difficult, while the training time is longer and precision
It is not high, and the present invention utilizes the known parameters such as label size, paste position, while marking the length of side in the image that camera is gathered
Shared image pixel quantity, carries out affine transformation, you can try to achieve the approximate location of blade, finally obtain leaf image, with seeking
Look for easily and fast, accurate advantage.
2. the present invention is due to that using the strong black and white color lump of color contrast, to photoperiod sensitivity reduction, can be greatly reduced
Due to the complicated caused leaf recognition precise decreasing of outdoor environment image irradiation, and enter rower relative to using colour-coded thing
Note, the present invention uses label information more to enrich and makes label simply, while make use of the geometry information of label.
3. the image processing algorithm complexity that the present invention is used is low, optimization blade extracts flow, is appropriate for monitoring in real time.
Brief description of the drawings
Fig. 1 is the schematic diagram of the leaf image extraction element of the embodiment of the present invention;
Fig. 2 is the blade surface mark figure of the embodiment of the present invention;
Fig. 3 is the flow chart of the leaf image extracting method of the embodiment of the present invention;
Fig. 4 is the blade border obtained of the embodiment of the present invention;
Fig. 5 is the leaf image after the interception of the embodiment of the present invention;
Fig. 6 is the camera control logic figure of the embodiment of the present invention.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, and
It is not used in the restriction present invention.As long as in addition, technical characteristic involved in each embodiment of invention described below
Not constituting conflict each other can just be mutually combined.
The present invention provides a kind of pneumatic equipment bladess surface image extracting method, and it is viscous to pneumatic equipment bladess surface correct position
Patch or spraying mark figure, then may include the ambient image of blade using picture pick-up device random shooting, monitoring blade surface
Mark whether to determine the region of leaf image in the picture, take off leaf image, marked according to the black and white for being pasted onto blade surface
Figure determines the position of leaf image, i.e., by marking the content of image to judge to monitor the numbering of blade.Whole process is related to
Setting, the extraction of leaf image and the determination of blade information of blade surface mark.
The extracting method is realized by leaf image extraction element, as shown in figure 1, leaf image extraction element includes:Wind
Machine blade surface mark figure, image acquisition units, what monitoring display unit of data storage analytic unit.
Image acquisition units, comprising responsible leaf image gather picture pick-up device (such as fan monitoring camera) and to taking the photograph
As the corresponding mechanical control equipment that head is exposed, zoom and the anglec of rotation are controlled, line number is entered with industrial computer while should also contain
According to the image pick-up card and data line of transmission.
Data storage analytic unit (such as industrial computer), including the hardware device that is stored to view data and to data
The software systems analyzed and processed, main energy supply is that the monitoring image that capture card is gathered is analyzed, and finds corresponding failure
Data are simultaneously transferred to display device.
Monitoring display unit (such as human-computer interaction device), realize with the interactive function of operating personnel, display monitoring picture and
Result.
Wherein, fan monitoring camera includes and automatically controls rotating function up and down and focusing function, the image of collection
The image collected can be passed to industrial computer by wired or wireless mode and be handled by information, industrial computer feedback control signal
Control the rotation and focusing of camera.Rotate and focusing signal determines that leaf position and field range are determined to mark, such as
The leaf position that fruit determines then needs to carry out camera rotation beyond camera coverage scope or focal length variations are come leaf packet
Containing into the visual field.
Specifically, leaf image extracting method of the present invention mainly comprises the following steps:
Step (1):As shown in Fig. 2 pasting at pneumatic equipment bladess windward side or spraying square black and white case marking figure, paste
Position should be plane and substantially parallel with blade windward side, so as to replace Leaf orientation substantially with the posture of mark figure.One
As mark choose length of blade center, it is ensured that be marked on whole blade at the same appear in probability in field of view compared with
Greatly.
Wherein, mark figure is made up of square area, in order to be made a distinction to not isolabeling internal information, Mei Gebiao
Remember that region, into n × n small positive square regions of decile, and is made a circle outside mark figure and designed by the line segmentation parallel to side
Into black, the small square colors of (n-2) × (n-2) by the inside made a circle except setting are black or white to mark progress
Coding, i.e., internal (n-2) × (n-2) net region is represented that 2 can at most be included by then marking per a line by n-2 (bit)(n -2)×(n-2)Individual change is possible, each change one blade id information of corresponding coding correspondence, and blade letter is set up in database
Table is ceased, more detailed blade information (predominantly blade numbering) is searched according to ID.
As shown in Fig. 2 pneumatic equipment bladess surface markers figure is the square indicia being made up of black and white square, by changing
The square position of the internal black and white of mark causes its id information comprising mark and directional information, has enough to meet mark
Black and white lattice area is unlikely to too small and requires 5≤n≤10 in abundant information and mark, the mark in such as Fig. 2, by 5 × 5 black and white
Lattice are constituted, and it is black to make a circle outside, and black represents 0 white and represents 1, then the information marked in Fig. 2 can be obtained:The first row
(00000), the second row (00110), the third line (01010), fourth line (01000), fifth line (00000), by increasing black and white
The quantity of square can also include more information.
Mark the size of figure relevant with the selection of camera, meet the mark figure picture material that algorithm can be gathered to camera
It is identified.The black bars or the white square length of side that constitute mark figure elementary cell are designed as being more than 10 in the present invention
Pixel, can meet requirement under this condition.
Step (2):Wind energy conversion system operation image is acquired using camera, camera putting position needs that three can be collected
The complete image of individual blade, can change the position of camera, the Image Real-time Transmission of camera collection according to vane propeller-changing situation
Into industrial computer.
Step (3):The image information that camera is gathered is carried out image procossing to obtain blade information.Gathered by camera
Information may can easily be obtained by image processing method comprising all or part of leaf image and complicated background image
The posture and internal information of image are marked, the present invention is known using priori known to mark figure size, paste position, blade dimensions etc.
Know, while image pixel quantity shared by the mark length of side marked in the image that camera is gathered, carries out affine transformation, you can ask
Blade approximate location, intercept the position imagery exploitation image processing method obtain included border, border in partly i.e.
For leaf image, as shown in Figure 4.
Specifically, comprising the following steps:
Step (3.1):Gradation of image is changed
Because the picture that camera is shot includes a variety of images such as leaf image, background image, it is coloured image, therefore
The image for gathering camera is needed to carry out gradation of image conversion first, by color-map representation into artwork master, and due on blade
Mark figure is made up of black and white block, and the black and white part contrast that image is converted into after gray-scale map in mark figure will not reduce.
Step (3.2):Image binaryzation
Each pixel in image after gradation conversion is become into black (pixel value is 0) or white (pixel value is 255),
The step realized by setting image threshold, first sets threshold value according to the actual requirements, such as 175, then when certain point pixel
Less than or equal to setting threshold value 175 when, then the pixel is set to 0 (i.e. black), on the contrary then be set to 255 (whites).
Step (3.3):Profile is monitored
By black in bianry image or the region of white pixel UNICOM, the profile in UNICOM region is found, multiple wheels are obtained
It is wide.
Step (3.4):Search for candidates
Polygonal segments are carried out to the multiple profiles obtained in step (3.3), the number of vertex of profile are judged, due to original mark
Note figure is square, and it should be the polygon that number of vertex is 4 in the picture, therefore number of vertex leaves for 4 polygonal region
It is used as candidate.
Step (3.5):Selection markers
In order to filter out final mark figure from the polygonal region of candidate, code identification need to be carried out to region inside,
The polygonal region of each candidate is converted to square area using affine transformation (it is prior art, be will not be described here)
(because original marking figure is square), then carries out mesh generation to region, and the grid for being for example divided into 5 × 5 (is divided into
Grid chart as mark figure grid number), the grid of grid inside 3 × 3 includes the information of label, if inside 3 × 3
Grid is surrounded by black border, then it is assumed that the area image is schemed for mark.Wherein, the grid of inside 3 × 3 includes 29=512
Kind of coded system, but need to deduct pure white 2 of black, and due to caused by image rotation different coding image turn into same
One coding 510/4=127, so can be obtained by 127 coding informations.Blade information (blade has been corresponded to for every kind of coding
Exclusive ID, represents the numbering of blade), it is known that in image whether there is blade by the information, has which blade etc. to believe
Breath.
Step (3.6):Obtain leaf image and blade information
According to the size of mark figure size, installation site and blade, affine transformation tentatively intercepts the area where leaf image
Domain, then extracts blade border and obtains leaf image, and the information of the blade is obtained according to the mark figure on leaf image.
Because the size of the installation site, the size of blade and the mark that are marked in blade is all known, it is possible to use
Information above in shooting image using marking image to derive the positional information of blade in the image area, it is specific as follows:
Assuming that mark is attached to the medium position of blade, i.e., at length of blade half, the size of mark is M × M meters, blade
Size is L meters × W meters (long × wide), because shooting angle will not can be marked perpendicular to mark in obtained image is shot
Remember the parallelogram A that image is one long a width of m pixels × n-pixel, a longer wide-edge and A both sides are done centered on the A centres of form
Parallel parallelogram A ', wherein, the length of A ' and m when parallel is m × L/M pixels, edge lengths parallel with n sides A '
For n × W/M pixels, the region overwhelming majority that the A ' now obtained is included all is leaf image, and background image has been only introduced very
Small part.
After the preliminary extraction image (i.e. A ') for obtaining blade region, it is all blade to extract most areas in region
Image, due to reduce external environment scenery and possibly into shooting image other blades disturb, then can utilize
The edge detection algorithms such as cannyEdge edge detectors extract blade border, and border inner region is exactly leaf image, and according to
Mark figure on leaf image can obtain the information of the blade.
If intercepting scope exceeds camera view scope, carrying out zoom to camera expands field range.In order to
Blade complete image can be quickly and accurately found, the present invention devises camera control logic, as shown in fig. 6, its control logic
For:
1) after camera collection image, graphical analysis is carried out, next step is carried out if mark figure is included in image, if
No then adjustment camera angle and focal length start to gather information, until mark figure appears in camera and gathered in image I.
2) blade surface analysis is carried out for the image I comprising mark figure, if analysis rear blade scope is all in image
In the range of then leaf image is intercepted, if it was not then carrying out zoom centered on marking figure to camera, find multiple
The respective mark figure of (such as three) blade, then swinging camera makes its central region be center determined by multiple mark figures
(such as the centre of form of three mark triangles that determine), continuing zoom expansion field range can make multiple (such as three) blades complete
Include in image I in portion.
As it will be easily appreciated by one skilled in the art that the foregoing is merely illustrative of the preferred embodiments of the present invention, it is not used to
The limitation present invention, any modifications, equivalent substitutions and improvements made within the spirit and principles of the invention etc., it all should include
Within protection scope of the present invention.
Claims (4)
1. the pneumatic equipment bladess surface image extracting method in a kind of complex background environment, it is characterised in that comprise the following steps:
(1) blade surface is set to mark:Square black and white case marking figure is set respectively at the windward side of each blade of wind energy conversion system, glued
The position of patch is plane and parallel with blade windward side, and then using marking the posture of figure to replace Leaf orientation, with according to mark
The acquisition of information leaf image region of figure;
(2) leaf image is gathered:Wind energy conversion system operation image is acquired using camera, and the image collected is passed in real time
Transport in industrial computer;
(3) blade information is recognized:The image information collected to camera carries out image procossing to obtain blade surface image, and
Pneumatic equipment bladess information is obtained according to the mark figure on blade, following sub-step is specifically included:
(3.1) gradation of image is changed:The image that camera is collected carries out gradation of image conversion, by color-map representation into black and white
Figure;
(3.2) image binaryzation:Each pixel in image after gradation conversion is become to the picture of black or white, wherein black
Element value is 0, and white pixel value is 255, i.e., when certain point pixel is less than or equal to given threshold, the pixel is set into 0, instead
Be then set to 255;
(3.3) profile is monitored:The region of black or white pixel UNICOM, seeks in image in obtaining step (3.2) after binaryzation
The profile in UNICOM region is looked for, multiple profiles are obtained;
(3.4) candidates are searched for:Polygonal segments are carried out to the multiple profiles obtained in step (3.3), are 4 by number of vertex
Polygonal region is left as candidate;
(3.5) selection markers:The polygonal region of each candidate is converted into square area, then square area carried out
Mesh generation, is divided into n × n grid, if a hoop net lattice of periphery are black, then it is assumed that the area image is mark
Figure;
(3.6) leaf image and blade information are obtained:According to size, installation site and the blade dimensions of mark figure, preliminary interception
Region where leaf image, then extracts blade border and obtains leaf image, be somebody's turn to do according to the mark figure on leaf image
The information of blade.
2. the pneumatic equipment bladess surface image extracting method in complex background environment as claimed in claim 1, it is characterised in that
If the region tentatively where interception leaf image exceeds the field range of camera, zoom is carried out to camera, makes the visual field
Scope expands, and is specifically controlled using following control logic:
1) after camera collection image, graphical analysis is carried out, next step is carried out if mark figure is included in image, if do not had
Then adjustment camera angle and focal length start to gather information, until mark figure is appeared in camera acquired image;
2) blade surface analysis is carried out to the image comprising mark figure, if leaf image is all in the image range of collection,
Leaf image is intercepted, if not existing, zoom is carried out centered on marking figure to camera, multiple blades are found respective
Mark figure, then swinging camera makes its central region be center determined by multiple marks, continues zoom and expands field range
Multiple blades are made all to include in the picture.
3. the pneumatic equipment bladess surface image extracting method in complex background environment as claimed in claim 2, it is characterised in that
The mark figure is made up of n × n small positive square regions of the line segmentation parallel to side into decile, and marks figure periphery one
Circle be designed to black, by the individual small square colors of (n-2) × (n-2) of the inside made a circle except setting to mark into
Row coding, the coding is corresponding with blade information, it is preferred that 5≤n≤10.
4. the pneumatic equipment bladess surface image extracting method in complex background environment as claimed in claim 3, it is characterised in that
The black bars or the white square length of side of the composition mark figure are designed as being more than 10 pixels.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611065764.4A CN106778723B (en) | 2016-11-28 | 2016-11-28 | A kind of pneumatic equipment bladess surface image extracting method in complex background environment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611065764.4A CN106778723B (en) | 2016-11-28 | 2016-11-28 | A kind of pneumatic equipment bladess surface image extracting method in complex background environment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106778723A CN106778723A (en) | 2017-05-31 |
CN106778723B true CN106778723B (en) | 2017-09-26 |
Family
ID=58902118
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611065764.4A Active CN106778723B (en) | 2016-11-28 | 2016-11-28 | A kind of pneumatic equipment bladess surface image extracting method in complex background environment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106778723B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108873044A (en) * | 2018-07-05 | 2018-11-23 | 北京航空航天大学 | A method of helicopter screw propeller is surveyed with respect to fuselage posture with GPS receiver |
CN109900719B (en) * | 2019-03-04 | 2020-08-04 | 华中科技大学 | Visual detection method for blade surface knife lines |
CN110748463A (en) * | 2019-11-18 | 2020-02-04 | 内蒙古工业大学 | Deicing device for blades of offshore wind turbine |
Family Cites Families (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2010007635A (en) * | 2008-06-30 | 2010-01-14 | Hatsuta Seisakusho Co Ltd | Fire detection system for wind turbine generator and fire detection method for wind turbine generator |
CN102231187B (en) * | 2011-07-12 | 2013-07-24 | 四川大学 | Computer vision detection technology-based method for detecting and identifying QR (Quick Response) code |
CN102254144A (en) * | 2011-07-12 | 2011-11-23 | 四川大学 | Robust method for extracting two-dimensional code area in image |
CN103984952B (en) * | 2014-04-18 | 2017-09-22 | 广东电网公司电力科学研究院 | The method diagnosed based on machine vision image to power system blade of wind-driven generator face crack failure |
CN103984956B (en) * | 2014-04-25 | 2017-07-18 | 广东电网公司电力科学研究院 | The method diagnosed based on machine vision image to power system blade of wind-driven generator surface pitting failure |
CN103982378B (en) * | 2014-04-25 | 2017-01-11 | 广东电网公司电力科学研究院 | Method for diagnosing surface icing faults of wind power generator blade of power system based on machine visual images |
CN104865269A (en) * | 2015-04-13 | 2015-08-26 | 华北理工大学 | Wind turbine blade fault diagnosis method |
CN105204742B (en) * | 2015-09-28 | 2019-07-09 | 小米科技有限责任公司 | Control method, device and the terminal of electronic equipment |
KR101581802B1 (en) * | 2015-09-30 | 2016-01-12 | 한상선 | Application system of goods augmented reality |
-
2016
- 2016-11-28 CN CN201611065764.4A patent/CN106778723B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN106778723A (en) | 2017-05-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111126325B (en) | Intelligent personnel security identification statistical method based on video | |
CN107729808A (en) | A kind of image intelligent acquisition system and method for power transmission line unmanned machine inspection | |
CN109949316A (en) | A kind of Weakly supervised example dividing method of grid equipment image based on RGB-T fusion | |
Wang et al. | Recognition and drop-off detection of insulator based on aerial image | |
Liu et al. | Insulator detection in aerial images based on faster regions with convolutional neural network | |
CN108648169A (en) | The method and device of high voltage power transmission tower defects of insulator automatic identification | |
CN106778723B (en) | A kind of pneumatic equipment bladess surface image extracting method in complex background environment | |
CN110610483B (en) | Crack image acquisition and detection method, computer equipment and readable storage medium | |
CN109271991A (en) | A kind of detection method of license plate based on deep learning | |
CN102878647B (en) | Automatic monitoring system and automatic monitoring method for operating state of air conditioner | |
CN109325935A (en) | A kind of transmission line faultlocating method based on unmanned plane image | |
CN107194396A (en) | Method for early warning is recognized based on the specific architecture against regulations in land resources video monitoring system | |
Wang et al. | Bottle detection in the wild using low-altitude unmanned aerial vehicles | |
CN111080691A (en) | Infrared hot spot detection method and device for photovoltaic module | |
CN113903081A (en) | Visual identification artificial intelligence alarm method and device for images of hydraulic power plant | |
CN104091175B (en) | A kind of insect automatic distinguishing method for image based on Kinect depth information acquiring technology | |
CN107509055A (en) | A kind of rotary panorama focus identification optronic tracker and its implementation | |
CN115359239A (en) | Wind power blade defect detection and positioning method and device, storage medium and electronic equipment | |
CN112465854A (en) | Unmanned aerial vehicle tracking method based on anchor-free detection algorithm | |
Zhang et al. | Automatic diagnosis system of transmission line abnormalities and defects based on UAV | |
CN110245592A (en) | A method of for promoting pedestrian's weight discrimination of monitoring scene | |
CN115240089A (en) | Vehicle detection method of aerial remote sensing image | |
CN114419444A (en) | Lightweight high-resolution bird group identification method based on deep learning network | |
CN113033386B (en) | High-resolution remote sensing image-based transmission line channel hidden danger identification method and system | |
CN113902792A (en) | Building height detection method and system based on improved RetinaNet network and electronic equipment |
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 |